*******

I was mapping soil sample data for a farm in the Midwest when one of the fields showed up in Mongolia, just outside Ulaanbaatar.

As cool as it would be to have Mongolian soil data, I don’t.

That is not the kind of error you can wave away. It’s not GPS inaccuracy or rounding drift.

It’s the mistake that forces you to stop and ask whether the system you are relying on deserves your trust at all.

The dataset itself looked serious.

It contained lab measurements for bulk density, soil mass, timestamps, and location fields, each record carrying dozens of attributes.

At first glance, it felt structured enough to analyze. But once I started to work through it carefully, it became clear that some of the records had issues.

Even just a percent error rate is small enough to ignore if you are careless and large enough to break your conclusions if you are not.

The real danger is not bad data by itself. The real danger is clean-looking analysis built on top of it.

When the numbers line up neatly, when the charts render without errors, and when the outputs feel precise, it becomes very easy to mistake measurement for truth.

That is a trap I have written about before in the context of performance marketing, where clean dashboards often hide flawed assumptions and overconfidence in the underlying data

The question is not whether your data is perfect. It never is. The question is what you do the moment you realize it is not.

The Three-Step Fix

When your data is wrong, there are three levels at which you need to respond.

You fix the point. (Easy)
You fix the series. (Medium)
You fix the foundation. (Hard)

Most teams stop after fixing a few bad rows because it feels like progress. It is visible, immediate, and easy to explain. But unless you address the process that produced those errors and the structure that allowed them to persist, you are not solving the problem, you’re resetting the clock.

First, You Have to Stop

Before you touch a single record, you need to pause long enough to understand what you are seeing. This is the part that gets skipped most often, especially in environments where speed is rewarded and uncertainty is uncomfortable.

Someone notices an outlier, shrugs, and continues. Someone else smooths it out, assuming it will average away. A report gets built anyway because the deadline is closer than the root cause.

That instinct is understandable.

It’s also wrong.

Bad data is not the problem. Acting on bad data is the problem. Once you build on top of something flawed, the issue compounds. A single incorrect record becomes a skewed distribution. That distribution becomes a misleading trend. That trend becomes a confident recommendation.

If something does not make sense, it is not an inconvenience. It’s still telling your something.

Subscribe now

What “Wrong” Actually Looks Like

In this dataset, the errors were not subtle if you knew where to look.

There were coordinates that placed Midwestern farm plots in Mongolia. There were clusters where nineteen samples sat within a tight radius and one was miles away. Dates appeared in unnatural patterns, with an odd concentration on January 1st. A few soil values drifted outside what is physically possible.

More importantly, the errors were not random.

They clustered by person.

One agronomist (sorry Stephen) consistently produced strong field insights but introduced more inconsistencies during data entry.

That pattern matters because it tells you that the issue is not statistical noise. It is process-driven. Most data problems are not technical at their root; they are human, and they reflect the systems we build around those humans.

You Have Seen This Before

If you work in marketing or operations, this pattern should feel familiar.

You pull a report and see a customer with a thousand purchases in a month. You see a contact who has not been touched in “999 days.” You see a campaign that appears to convert perfectly.

These are not insights. They are artifacts.

In one dataset, the value “999” was used to represent “never contacted.” It shows up as a number, but it does not behave like one. It dominates the distribution and masks the real patterns underneath it

This is the same failure mode I’ve written about elsewhere: numbers are treated as objective truth simply because they are numeric, even when they are proxies, placeholders, or side effects of a system that was never designed for analysis in the first place.

Step 1: Fix the Point

The first level of response is local. You examine individual records and decide what to do with them.

The key is to approach this carefully, because it is easy to introduce new errors while trying to remove old ones.

Each questionable value falls into one of three categories: something you can fix, something you can flag, or something you should leave alone.

If you can verify the correct value, you should fix it. A coordinate that is off by a sign is often recoverable. A sample that sits far outside a cluster may indicate swapped fields or a simple entry mistake. When you have enough context to be confident, correction is appropriate.

If you suspect a problem but cannot confirm it, then the right move is to flag it. That preserves the data while making its uncertainty visible to anyone who uses it later.

If you do not know the correct value, then leaving it blank is often the best option. This is where many datasets get worse during “cleaning.” Missing values are replaced with zeros or placeholders that look legitimate but are not.

If zero is a valid value elsewhere, you have now created ambiguity where none existed before. DONT DO THIS!

That is exactly how placeholder values like “999” become dangerous. They begin as a practical encoding of missingness and end up being treated as real data, distorting analysis in subtle but meaningful ways

Alongside these decisions, you should apply basic sanity checks.

Physical systems have limits. Geography has boundaries. Time has structure. When a value violates those constraints, it should not quietly pass through your pipeline.

Looking at distributions is often the fastest way to catch these issues. Spikes at round numbers, unnatural clustering, or tails that do not match real-world behavior are rarely random. They are usually signs of defaults, artifacts, or systematic entry errors.

Interpolation can be useful when you need continuity, but it comes with responsibility. Any inferred value should be clearly marked as such and, ideally, stored separately from observed data. Blurring that line makes it harder to reason about what is real and what is constructed.

Subscribe now

Step 2: Fix the Series

Fixing individual records is necessary, but it is not sufficient. If you stop there, the same errors will continue to appear.

This step requires shifting your focus from data points to processes.

In this case, the causes were clear. Data was entered manually. Excel was used as a primary tool. Validation was minimal.

Whatever—the

Each of those choices introduces flexibility, and flexibility without guardrails produces inconsistency.

The goal is to reduce the opportunity for error.

That can mean replacing free text fields with structured inputs, enforcing validation rules at the point of entry, or capturing certain values automatically rather than relying on manual input. If a device can record location data, there is no reason to trust someone to type it in.

Patterns like errors clustering by person are especially useful here. They reveal that the system depends too heavily on individual discipline.

A well-designed system should make it easy to do the right thing and difficult to do the wrong thing.

In Toyota Production System, this is known as a Poka Yoke

This is the same principle that shows up in more complex systems as well.

Most failures in modern martech are not due to lack of capability but lack of structure. Tools can do almost anything in isolation, but without consistent interfaces and standards, they become brittle and hard to trust

Subscribe now

Step 3: Fix the Foundation

The final step is the most strategic and the least commonly addressed. It involves designing your data environment so that it can withstand errors and still produce reliable insight.

At its core, this means asking whether important information can be recovered if something goes wrong. If a field is critical and there is only one way to capture it, then any mistake in that process becomes permanent.

A more resilient approach is to build redundancy into the system.

In the soil dataset, farmer enrollment data provided a geographic boundary that could be used to validate sample locations. Clustering of nearby samples offered additional context. These are forms of indirect validation that make the system more robust.

You can extend this further. Capturing images at sampling points would allow for later verification. Storing raw lab outputs alongside processed values would preserve a ground truth reference. Each additional signal increases your ability to detect and correct issues after the fact.

It is tempting to view redundancy as waste, especially when optimizing for speed or cost—As if compute redundancy is your biggest barrier to productivity. But systems that optimize too aggressively for efficiency often remove the very safeguards that keep them stable.

Storage has become extremely cheap in the past decades. Business mistakes have not dropped in cost.

This is the same dynamic seen in broader marketing systems,

where over-optimization can degrade long-term performance by stripping away the underlying structure that supports growth

Subscribe now

Bringing It Back to Martech

All of this applies directly to marketing and operational data.

Customer records contain duplicates and placeholders. Attribution models rely on incomplete signals. Dashboards present clean numbers that often obscure messy realities.

The pattern is consistent. When something looks precise, it is easy to trust it. When it is easy to trust it, it is rarely questioned.

That is why this framework matters. Fixing individual records is not enough. Fixing the process without addressing the structure is not enough. You need all three layers working together if you want your system to produce something you can rely on.

What Failure Actually Looks Like

Bad data is1 inevitable. It is a byproduct of real-world systems interacting with imperfect inputs and human behavior.

The failure is not in its existence. The failure is in the response.

Ignoring inconsistencies because they are inconvenient leads to brittle analysis. Freezing when uncertainty appears leads to stagnation. Presenting results you do not trust leads to poor decisions that carry real consequences.

A better standard is to treat uncertainty as part of the work. If something cannot be resolved, it should be acknowledged.

If a dataset has limitations, those limitations should be understood and communicated.

It is better to be directionally correct with clear caveats than precisely wrong with none.

A Simple Place to Start

If there is one practical step to take from this, it is this:

1. Choose a dataset you rely on

2. Find one value that does not make sense

3. Follow it back to its source

4. Understand how it got there.

5. Make a hypothesis for future data in that series

6. Test it over time

That process will teach you more about your system than any dashboard ever will.

Because the goal is not just to clean your data. The goal is to build a system that you can trust (even when it breaks).

Subscribe now

Take your most pessimistic estimate of how long a new project will take.

Double it

Double it again

You’re probably halfway there

—my dad’s rule of thumb

Starting a new project with other people is harder than it looks.

Not because the work is complex, but because there is no map yet. No spec. No shared picture of what “done” means. Just a loose goal and a group of smart1, already-busy people expected to move in the same direction.

At the end of every project, you’re going to look back and realize if you started again you could have done it in 1/3rd the time—thats normal!

If you’re someone who likes clear rules, sharp edges, and known targets, this stage can feel hard, difficult, and deeply wrong.

Engineers, analysts, and operators often struggle here—not from lack of skill, but from lack of structure. When the ground is soft, it’s hard to know where to step.

If you just want the checklist, jump here or click the PDF version below:

Deck Template

4.99MB ∙ PDF file

Download

Download

Project Startup Checklist

114KB ∙ PDF file

Download

Download

Copiable Deck Format

Template google drive directory

I know this feeling because that’s how my brain works too.

For a long time, I thought good project managers and consultants were doing something vague.

From the outside, their work looked fluffy. I’ve mocked PMO offices before.

Too many check-ins2. Too many docs. Too much talk.

I was wrong. I have been before (see below).

When project work is done well, you barely notice it. The rough spots have already been smoothed.

Files are where you expect them to be. Decisions don’t vanish. People know what’s expected, even when the goal is still taking shape. Progress feels steady, almost calm.

When it’s done poorly, everything drags. People go quiet. Work stalls for reasons no one can quite name. The project doesn’t fail loudly. It just fades until one day you give up.

This is the guide I wish I’d had the first few times I led something new.

It’s written for smart, capable people who hate fuzzy project edges and would much rather have a spec—but don’t.

Here’s the part that matters most:

You will learn by doing.
You don’t need to grasp every step before you start.
Trust the process. Follow it anyway.3

Think of this as a stand-in map. A set of rails you can run on while the shape of the work comes into focus. The rules may feel strict at first. That’s on purpose. Clear rules lower the load on your mind when everything else is still unclear. Loose rules are great when everyone knows norms.

These steps are built for async work, where gaps and silence can quietly break a project. They also work in person, with a much lighter touch.

If you follow them as written—even when they feel heavier than needed—you’ll give your project something rare at the start: solid ground.

Step 0: Pre-Work

What You Must Decide Before Involving Other People

Decide That You Are the Lead

If you are starting this project, you are the lead.

Take a moment to realize that no one will be checking in on you day-to-day. You’re responsible for self-motivation and management.

This is not a democracy. You will ask for advice and you should change your mind when better ideas appear—but decisions land with you.

Unclear leadership creates more friction than firm leadership ever will.

Decide the Defaults

Decide, in advance:

• Where files will live

• How people will communicate

• How calls will work

• Where decisions will be recorded

These choices do not need to be perfect. They need to exist.

Undecided defaults invite interpretations. This will be a problem later.

Decide How Suggestions Will Come In

People will have opinions. Many will be useful.

Decide:

• Where suggestions go

• When you will review them

• How decisions will be made

For example:

“Tooling suggestions via email. I won’t decide this live.”

This keeps meetings focused and protects shared time.

Decide What “Participation” Means

Be clear with yourself about what it takes to be involved.

Does participation require being on a team? Owning work? Showing up? Writing things down?

You don’t need to publish this yet. You do need to know it.

Once these decisions are made, stop thinking and start acting.

The next step is to make these rules visible—by behaving more formally than feels natural when other people enter the system.

Step 1: Act More Formal Than Feels Necessary

Formality is not friction. It’s the opposite. It keeps things moving smoothly.

Once other people are involved, your job changes.

You are no longer just thinking through a problem.

You are shaping a shared space where work can happen without constant negotiation. In that space, informality is not neutral. It creates gaps. Gaps create confusion. Confusion slows everything down.4

So start more formal than feels natural.

This will feel stiff at first. That’s fine. Early formality does not lock you into rigid process forever. It gives you a stable floor while the work is still taking shape.

Decide Once, Then Write It Down

Every choice you’ve already made needs to be stated plainly.

Not hinted at. Not implied. Written.

This includes:

• Where files live

• How people communicate

• How decisions are recorded

• What tools are in use

• What happens when someone joins late

If it isn’t written, it doesn’t exist.

Recording decisions is a key hack for making sure you’re always learning.

People will not infer your system. They will assume there isn’t one.

Remove Choice Wherever You Can

Choice feels empowering. Early on, it’s mostly a tax.

If you allow multiple tools, people will pick their favorites. If you allow multiple ways of working, people will follow habits from past jobs. None of this is malicious. It’s just human.

Your goal is not to optimize for preference. It’s to optimize for shared motion.

That means:

• One file system

• One default chat channel

• One place decisions get summarized

You can always loosen later. Tight first. Loose later.

Write Procedures for the Parts You Understand

You do not need to document the whole project. You only need to document what you already know how to do.

• If you know how meeting notes should be written, write it down.

• If you know how files should be named, write it down.

• If you know how decisions should be logged, write it down.

Short is fine. Rough is fine. Incomplete is fine.

What matters is that someone who joins later can orient themselves without needing a private walkthrough.

That is one of the quiet tests of good structure.

Assume No One Was in the Meeting

In async work, meetings are a convenience—not a source of truth.

Any decision that matters needs to live outside the call. If someone misses a meeting, they should still be able to catch up without shame or guesswork.

This isn’t about accountability. It’s about respect.

People are busy. People have day jobs. Your system should not punish them for that.

Expect This to Feel Excessive

At this stage, people may not say much. That doesn’t mean they aren’t reading. It doesn’t mean they don’t care.

Formality often feels louder to the person creating it than to the people receiving it. You are laying track ahead of the train. The smooth ride comes later.

Trust that this work is doing something, even when it feels quiet.

Formality is not about control.
It’s about clearing the path.

Once the structure is in place, you’re ready for the next step: setting the tools themselves so there is exactly one way to work.

Step 2: Pre-Decide the Tools (No Exceptions)

Tools are not neutral.

Every tool choice carries habits, expectations, and hidden costs. When tools are left undecided, people fill the gap with whatever they already know. That’s how a project quietly fractures before it ever picks up speed.

Your job here is simple: decide once, then remove the question.

Pick One Place for Files

There must be exactly one place where files live.

Not “mostly.” Not “unless it’s easier.” One place.

For work projects, this usually means whatever your company already uses. Don’t fight that battle unless you have to. Familiar tools lower the bar to entry.

For personal or volunteer work, pick a free option that everyone can access without setup pain.

Once chosen:

• Create the shared drive yourself

• Set the top-level folder structure

• Share access before kickoff

Do not wait for agreement. Do not ask for preferences. File sprawl is far harder to clean up later than it is to prevent.

Use a Clear, Boring Folder Structure

See the example drive for a specific template

Your folder structure is doing more work than you think.

It tells people what matters.
It tells them where to put things.
It tells new arrivals how serious this project is.

Use a structure that:

• Is shallow

• Is obvious

• Leaves room to grow

This is not about elegance. It’s about orientation. If someone can’t guess where a file belongs, the structure is too clever.

Once it’s set, don’t casually change it. Stability builds trust.

Choose One Default Communication Channel

Pick one place where day-to-day conversation happens.

This is the channel people should check if they want to stay in the loop. Everything else is secondary.

Be explicit about:

• What belongs in this channel

• What does not

• Where decisions will be summarized

Also be clear about expectations:

• People are not required to respond to every message

• Silence does not mean disapproval

• Reading is often enough

This matters more than the tool itself. Clear norms reduce social strain.

Decide How Calls Will Work

Calls are useful. They are also easy to misuse.

Decide up front:

• What tool you’ll use for calls

• When a call is appropriate

• What happens after every call

The rule should be simple:

No call produces new information that isn’t written down and shared.

Meeting notes don’t need to be long. They do need to exist.

Lock the Stack (For Now)

Once these choices are made, freeze them.

This does not mean the tools are perfect. It means they are good enough to move forward. Tool churn kills momentum and invites endless side debates.

If you plan to revisit tooling later, say so. Also say when and how. Until then, the stack is closed.

If someone asks to change tools, you already have an answer:

“Send me the suggestion. I won’t decide this live.”

That’s not being rigid. That’s protecting shared time.

Good projects don’t start with the best tools.
They start with shared ones.

Once the tools are set, you’re ready to bring people together and make the project real.

STEP 3: Run a Formal Kickoff (Yes, With a Deck)

The kickoff is not a status meeting.

It is the moment where a loose idea turns into a shared effort. If you skip it—or keep it casual—you’re asking people to infer rules, roles, and priorities that you haven’t stated.

Don’t do that.

Run a formal kickoff. Bring a deck. Even if it feels like overkill.

It can be 2 slides, but it structures a meeting in an important way to gather trust and set expectations implicitly.

The Kickoff Is Mandatory

No work starts before the kickoff.

Not research. Not side tasks. Not “just poking around.” The kickoff is the line between thinking and doing.

This isn’t about ceremony. It’s about alignment. People deserve to know what they’re stepping into before they start contributing.

What the Deck Must Cover

Your deck does not need to be polished. It does need to be clear.

At a minimum, include:

• What this project is

• What it is not

• Any expected digital norms (e.g. cameras on)

• Why it exists now

• How success will be judged (even if roughly)

• Where files live

• How people should communicate

• Where decisions will be written down

You are not selling. You are setting expectations.

If something feels obvious to you, say it anyway. Obvious rules are still rules.

Assign Roles and Teams Live

This part is not optional.

During the kickoff, assign:

• Roles

• Teams

• Owners

Do it live. Write it down as you go. Share the doc before the meeting ends.

Here is the rule you should say out loud:

To stay involved, you must be on a team.

This does not mean people can’t help outside their team. It means there is no such thing as a vague participant. Loose involvement turns into drift. Drift turns into disengagement.

This is a mistake I’ve made more than once. Every time, it cost the project.

Make Participation Concrete

People want to help. They just don’t want to guess how.

Be clear about:

• What each team owns

• How work will be tracked

• What “good progress” looks like in the first few weeks

Early clarity beats early perfection.

Leave Time for Questions—Then Close

At the end of the kickoff, leave room for questions. Answer what you can. Write down what you can’t.

Then close decisively.

Say what happens next:

• When the next update will be

• Where follow-up questions should go

• What people should do first

End the meeting with forward motion.

STEP 4: Overcommunicate (More Than Feels Comfortable)

Once work begins, your instincts may push you to go quiet.

You don’t want to distract people. You don’t want to spam. You don’t want to narrate thoughts that aren’t fully formed yet.

Resist that urge.

In a new project, silence creates more confusion than noise.

Share What You’re Thinking Early

You do not need to wait until ideas are finished.

Share:

• What you’re leaning toward

• What you’re unsure about

• What tradeoffs you’re weighing

This is not about asking permission. It’s about giving context.

When people understand your thinking, they’re far less likely to feel blindsided if something changes later.

Repeat Yourself on Purpose

It will feel strange to say the same thing more than once.

Do it anyway.

In async work, people read at different times. They skim. They miss things. Repetition is how shared understanding forms.

If something matters, it should appear in:

• A message

• A doc

• A summary

That’s not redundancy. That’s reinforcement.

Do Not Expect Responses

Here’s a rule that removes a lot of stress:

Messages can be useful even if no one replies.

People are busy. They may read and move on. They may save it for later. That doesn’t mean the message wasn’t helpful.

If you tie your sense of progress to visible responses, you’ll stop communicating too early.

Assume your words are landing, even when it’s quiet.

Plant Seeds Before You Pivot

If you think a change might be coming, say so.

Early signals matter. They give people time to adjust. They also invite better ideas before things harden.

A simple “I’m thinking about changing X because of Y” goes a long way.

Most frustration comes from sudden shifts that feel unexplained. Overcommunication prevents that.

Summarize Decisions Clearly

Whenever a decision is made, summarize it in the agreed place.

Keep it short:

• What was decided

• Why

• What happens next

This is not bureaucracy. It’s memory.

People should never have to guess whether something is settled.

Overcommunication feels awkward because it’s one-way at first.

Later, when the project holds together under pressure, it will feel obvious in hindsight.

A good kickoff does not solve the project.
It gives everyone the same starting line.

Once teams and roles are set, your next job is to keep momentum without burning people out—which means learning how to communicate more than feels natural.

STEP 5: Reward and Celebrate Good Work (Publicly and Often)

Early in a project, people are taking a risk.

They’re spending time on something new. They’re working without a clear finish line. They’re exposing half-formed ideas in front of others. That takes effort and a bit of nerve.

If that effort goes unnoticed, people pull back. Not out of spite—out of self-preservation.

Your job is to make sure that doesn’t happen.

Praise Progress, Not Just Outcomes

In the early days, outcomes are rare. Progress is not.

Look for:

• Someone who documented something unprompted

• Someone who tried an approach that didn’t work

• Someone who helped another team without being asked

• Someone who closed a loop others forgot about

These are the behaviors you want more of.

If you wait until the project is “successful” to celebrate, you’ll be celebrating alone.

Make Praise Public

Private thanks are good. Public recognition is better.

Public praise does three things at once:

• It rewards the person who did the work

• It shows others what good looks like

• It signals that effort is seen

This does not need to be grand. A short message in the main channel is enough.

Be specific. Vague praise fades quickly.

Create a Simple Ritual

Rituals remove guesswork.

One easy option is a recurring highlight:

• “Team Member of the Week”

• “This Week’s Shout-Out”

• “Good Work Worth Calling Out”

Pin it. Send it on a schedule. Keep it light.

The goal is not competition. The goal is momentum.

Reward Experimentation

New projects need trying, not just finishing.

If someone tests an idea and reports back—even if it fails—that is success at this stage. Say so.

People will stop experimenting if failure feels invisible or risky. You want the opposite.

Assume People Need More Encouragement Than You Think

Most capable people underestimate how much positive signal others need.

You may feel like you’re being obvious. You aren’t.

Err on the side of saying it out loud.

Recognition is one of the cheapest tools you have.
It pays off faster than almost anything else.

Once people feel seen, they’re more willing to ask questions—which brings us to the next step.

STEP 6: Have Standing Office Hours (Lower the Cost of Asking for Help)

Here’s a quiet truth about new projects:

Asking for help has a cost.

Even when you say, “Just reach out,” people hesitate. They don’t want to interrupt. They don’t want to look unprepared. They don’t want to take more time than they need.

That friction is enough to keep small questions unasked. Small questions turn into wrong assumptions. Wrong assumptions compound.

You can fix this.

Do Not Rely on Ad-Hoc Requests

If people have to:

• Schedule time

• Write a careful message

• Explain why they need help

many of them simply won’t.

This is not a motivation problem. It’s a friction problem.

Everyone is very well aware of friction in business until it comes to your own backyard.

Set Predictable Office Hours

Create recurring, visible time where you are available by default.

For example:

• “I’ll be on this call every Monday, Wednesday, and Friday from 2–4pm.”

• “I’ll be in this room during this window.”

No agenda. No obligation. Drop in or don’t.

The key is that the time exists whether or not anyone shows up.

Make It Clear What Office Hours Are For

Say this out loud:

This is for quick questions, half-formed thoughts, and things you’re not sure are worth a meeting.

Office hours are not status updates. They are casual questions that might need more exploration.

Watch for Patterns

Office hours are not just for answering questions. They are for listening.

If you hear the same confusion more than once, that’s a signal:

• Something wasn’t documented

• An assumption wasn’t shared

• A rule wasn’t as clear as you thought

When that happens, fix the system—not the person.

Write it down. Share it once. Save everyone time.

Treat Empty Office Hours as Success

Some weeks, no one will show up.

That doesn’t mean the time was wasted. It means things are working, or people didn’t need you right then.

Keep the hours anyway. Consistency builds trust over time.

Office hours turn silent friction into visible feedback.

Once you’ve given people a low-cost way to ask questions, you can close the loop properly at the end of the work.

STEP 7: Run an After-Action Review (Close the Loop)

When the work slows down or ends, there’s a strong urge to move on.

Resist it.

Without a deliberate pause, lessons evaporate. People carry quiet frustrations into the next project. Small wins go unmarked. The system never improves.

An after-action review fixes that.

Make the Review Explicit

Do not let reflection happen “informally.”

Schedule a meeting. Invite the people who mattered:

• Core team members

• Key stakeholders

• Anyone who had to live with the results

This signals that learning is part of the work, not an afterthought.

Keep the Agenda Simple

You don’t need a long deck or a formal report. You need honesty.

Structure the conversation around three questions:

• What went well?

• What didn’t?

• What should we try differently next time?

Write the answers down as they’re said. This is shared memory, not private notes.

Include Stakeholders on Purpose

Stakeholders often see things the team doesn’t.

They notice friction at handoffs. They feel the cost of delays. They see where expectations drifted. Their input makes the next project smoother—if you invite it.

Focus on Systems, Not People

If something went wrong, ask:

• What allowed this to happen?

• Where was the gap?

• What rule or structure would have helped?

Avoid blame. Blame feels satisfying for about five minutes and fixes nothing.

Good reviews change systems. Bad reviews assign fault.

End With Action

Before the meeting ends, capture:

• One or two things you will keep

• One or two things you will change

These become starting points, not just notes.

Share the summary. Close the loop.

Subscribe now

Owen spends his time helping organizations—many of them marketing-led—navigate the gap between what modern technology can do and what it should do.

He doesn’t approach AI as a novelty or a product feature, but as infrastructure that has to earn trust.

AI is a high horsepower engine without the rest of the car to support it.

AI has reached the stage where almost anything is possible in isolation.

You can generate copy.
Summarize reports. (kinda)
Draft emails.
Spin up a demo landing page over coffee.

Yet, for most real business teams, very little of this feels solid.

The tools are impressive, but brittle.

The demos shine, but the moment you try to wire AI into the systems that actually have business data flow—your data warehouse, your CRM, your workflow tools—things slow down.

Permissions get fuzzy. Auditability decreases. Trust evaporates.

That’s not a model problem.
It’s a systems problem.

And it’s why a quiet, unglamorous connector pattern called the Model Context Protocol (MCP) matters more than most people realize.

A Detour Through Fire Hydrants

Fire hydrants in the United States are not engineering marvels on their own.

The interconnection has been the same for over 100 years.

They are not optimized for maximum flow for each municipality. They are not custom-fit to each climate or water composition.

From a narrow, local perspective, you—the reader—could probably design something better.

But no one does.

Instead, we have a universal standard1. Any fire truck, from any jurisdiction, can pull up to any hydrant and start moving water immediately. No adapters, negotiation, or guesswork.

That choice—to standardize the hookup rather than perfect the part—created a network effect that matters far more than small efficiency gains.

When a building is on fire, the best connector is the one that works everywhere.

MCP is that kind of decision.

What MCP Actually Is

MCP is not a new model or AI tool.

Its like a fire hydrant—it connects any AI model and any web information.

It defines how AI-powered applications connect to the outside world.

At a technical level, MCP is an open protocol that lets AI hosts—things like chat apps, IDEs, or internal tools—connect to external servers that expose three kinds of capabilities:

• Resources: data the AI can read

• Tools: actions the AI can take

• Prompts: reusable patterns for how work should be framed

That’s it. No magic or intelligence. Its just a shared design pattern.

If APIs are the pipes that move data between systems, MCP is the hydrant fitting that lets AI connect to those pipes safely and predictably, regardless of who built the system on either side.

This is where a lot of early explanations go wrong. MCP is not “a better API.”

Your CRM, CDP, and warehouse still have APIs.

MCP standardizes how AI systems discover and use those APIs without each integration becoming a one-off project.2

Or, to borrow from the interview, MCP is “an evolution from an API,” not because it’s stricter, but because it’s looser in the right way.

As Owen put it:

“An API is very [rigid;] this endpoint must connect to this endpoint.

MCP is more like saying we can take data of any kind over this protocol and connect this application and that application.”

That “looseness” is the advantage with AI tooling.

Why Context Beats Cleverness

Owen offered another analogy that’s worth keeping in mind:

“LLMs are the world’s most forgetful journalist.”

A journalist without notes can still write. Sometimes even well. But the work gets better—more accurate, more grounded—when they’re given context, sources, and constraints.

Modern AI models are the same.

They are very good at working with what’s in front of them. They are much worse at guessing what should be in front of them.

MCP exists to solve that problem at a systems level.

Instead of stuffing more and more instructions into a prompt and hoping nothing important gets dropped, MCP gives AI a reliable way to ask:

• What do I have access to?

• What am I allowed to do?

• How is this work usually done here?

That’s not about raw intelligence (you can’t actually just use more powerful math to solve every problem)

It’s about orientation.

And reliable, predictable orientation is what turns a clever demo into a dependable AI tool.

Front End, Back End, Same Problem

In the interview, Owen described two main ways MCP-style connections show up:

“You’re seeing it in two ways, like a front end and a back end. And those two are all you need because they scale.”

On the front end, this looks like assistants that actually know your business. Not just your brand voice, but your customers, your history, your constraints.

An AI that can read previous campaign performance, check CRM notes, and draft a message that makes sense now, not in the abstract.

On the back end, it looks like what Owen half-jokingly called “bureaucracy.”

“You can build a team of bureaucrats to accomplish a larger goal. LLMs are wonderful at it if they have proper documentation.”

This is the unsexy work: reconciling data, writing summaries, updating systems, documenting decisions. Humans are bad at it. AI, given clear rails, is very good.

But only if the hookups are standard and the data is fundamentally solid.

What Production Actually Looks Like

Let’s make this concrete.

Imagine Snowflake sits at the center of your marketing data—not everything, but the things you trust. Curated views. Clean definitions. Data someone has already decided is safe to use.

An MCP server exposes those views as resources. Read-only at first. No surprises.

A small set of actions—run a scoped query, generate a summary, open a Jira ticket—are exposed as tools. Each one documented. Each one bounded. Some requiring confirmation before they run.

Before every user-prompt is a context prompt to standardize how work starts: how you summarize campaigns, how you draft briefs, how you explain results to executives.

Every interaction is logged back into the warehouse. Not for surveillance, but for learning. What did the AI touch? What did it suggest? What did humans accept, reject, or edit?

This is not autonomous marketing.

It’s assisted work, done faster3 and with fewer dropped threads, because the system knows where it is and why it exists.

As Owen put it plainly:

“You must give an agent the context of why they’re there and what they need to know.”

That’s the same set of skills as properly managing human employees.

Conclusion

The hydrant standard was not glamorous. It didn’t win awards for elegance. But it saved—and saves—cities, because it made coordination possible under stress.

MCP is the same idea.

It’s not about smarter models or clever prompts.

It’s about agreeing on the interface, even when that means giving up some local optimization. It’s about treating AI as part of an ecosystem, not a pile of demos.

Seen this way, MCP isn’t hype. It’s a sign that the industry is starting to think in systems again.

And systems thinking—slow, careful, often boring—is how technology stops being a toy and starts being infrastructure.

What’s Next

Now that MCP has been adopted by almost all major tech players, they’re moving onto another Anthropic suggestion: Skills.

Skills are tiny Standard Operating Procedures [SOPs] that explain to an AI tool how to do something.

If you had a “Turn on the Lights” Skill, it would tell the agent how to open your front door, where the light switch was, and how to flip it to “ON”.

As more companies adopt this structure, I’ll add more detail for a future article.

Subscribe now

Luckily, the party was enough fun that it ran well into the night.
Unluckily, I don’t have lights back there.

It took a while, but I cleaned everything up. All without a single dropped plate or scraped knee.

Gingerly-placed steps and conscious balance kept me from falling—even without seeing where my feet were going.

By moving with care in an unknown space, I was able to accomplish my goal without issue.

When Familiarity Becomes a Trap

Another time in the dark, I walked to my bed with the light off after moving some furniture the day before.

Taking full-length, full-speed strides, I confidently walked my toes directly into the bed frame.

My mental model of the room was clear, but inaccurate.

Using a clear—but inaccurate—understanding of the room was worse than my no-knowledge exploration of the backyard.

Both moments taught me the same thing: fear and false confidence lead to worse decisions than uncertainty itself.

What This Has to Do With Marketing

Just like walking confidently into the bedframe, marketers often charge forward with measurements that feel familiar but don’t reflect reality.

Moni Oloyede (of the Marketing Accountability Council) does a great job of consistently calling out normal, but stupid marketing practices.1

Marketing Accountability Council (MAC)

The MQL Is the Worst Metric in Marketing

The MQL is the worst metric in all of marketing…

Read more

5 months ago · 2 likes · 2 comments · Moni Oloyede

In her most recent piece, she highlights a common, but flawed, metric that drives B2B marketing strategy (Marketing Qualified Leads [MQL]).

People cling to MQLs not because they work—but because they’re comforting.

Her most important quote is:

If Not the MQL, Then What Do We Do Instead?

Here’s the part marketers always ask me for: “Okay, but what’s the alternative?”

She goes on to give a few practical answers, but I want to highlight the state of someone asking that question.

They’re afraid.

Sometimes One Map Is Worse Than None

Someone asking that question is a lead hiker in a group. With only a map.

When you’re start out, you quickly realize the map is for a different trail, but the sun is sinking.

Quickly.

The light is fading and the group pesters you to get to a camp site.

The map says go left.

The trail splits straight and right.

If you’re clinging to the map as a clear source of truth, you’re about to go tramping into the untamed forest instead of slowing down.

The group wants to follow the tool, but you have the foresight to feel in your gut that its wrong.

You’ll be in a worse position if you confidently follow the map, then have to backtrack.

Do you let the group pressure you into stepping off the trail?

Subscribe now

The False Choice We Keep Making

In business, there’s a natural conflict between two camps:

1. [Under measure] the measurements-don’t-matter camp

   1. This group fears being overly constrained by bad metrics

2. [Over measure] the measure-uber-alles camp

   1. This group fears the consequences of being wrong

Unfortunately, neither are fully right.

On one hand, ignoring trends and data when you have them is a foolish mistake.

On the other, blindly following measurements into a death spiral is even worse.

Instead of going all in on one camp, you should use your measurement to extract as much value as possible—but no more than that.

Don’t let the fear of uncertainty tempt you into charging more confidently than you’re able to.

What Should I Do Instead?

Move through your data the same way you’d move through an unfamiliar path at dusk—carefully, deliberately, and with an honest read of what you can and can’t see.

• Track the source of every metric and your confidence in it. Label where each number could drift from reality. It’s useful to have this pre-recorded.

• Connect your conclusions directly to those sources. Don’t let insights float disconnected from the data that they come from.

• Add expiration dates. Decide ahead of time what conditions would invalidate your assumptions or require a new look.

• Stay humble about your predictions, but be confident about your work. Executives appreciate grounded confidence more than bravado.

Conclusion

In the dark, confidence without clarity is a liability—but fear without perspective is just as dangerous.

Most bad decisions, in marketing or anywhere else, come from the same instinct that pushes you to trust the wrong map (or sprint into the bedframe): the urge to make the uncertainty go away as fast as possible.

Fear begs for certainty. It wants you to pick a direction—any direction—and move with conviction, even when your footing is questionable.

But you don’t have to be driven by that impulse. As long as you are always learning, you don’t need all the answers today.

Slow down. Feel the ground beneath you.

Let your measurements inform you without letting them rule you.

The goal isn’t to escape uncertainty; it’s to navigate it without letting fear shove you into decisions you’ll regret.

Subscribe now

Set the right parameters and constraints: You can bend the numbers to your will.

I worked in the consulting arm of Omnicom’s precision marketing group, surrounded by very capable coworkers. These weren’t your average campaign managers—they were technical consultants in a marketing agency. Smart, measured, and relentlessly data-driven.

Compared to them, most brand marketers I encountered semmed underwhelming.

Creative briefs lacked substance. No one seemed to think through the business implications of an effort. Campaigns prioritized aesthetics over outcomes. Attribution? Often an afterthought—if even that

.

It was easy to feel like performance marketing was the only rational choice in a field cluttered with opinions and intuition.

And for a time, that belief stuck.

The targeting was better. The models were cleaner. The metrics were immediate.

And most importantly, I felt smart doing it.

But like most just-out-of-college students, my tightly-held convictions started to crack under pressure.

The Seduction of Smart Numbers

I didn’t just trust precision marketing because of its methods—I trusted it because of its people.

There’s a natural tendency to conflate intelligence with the validity of its principles.

Synthetic CDOs, after all, were designed by brilliant minds.

But brilliance isn’t immunity from blindness.

In marketing, those blind spots are oversimplification and overconfidence.

Oversimplification in reporting ROAS, oversimplification through last-touch attribution, and overconfidence in dashboards that look scientific but often ignore what matters most: business fundamentals.

As Benn Stancil put it, we tend to privilege the SWAN—Suggestions With A Number

In a room where two teams pitch competing strategies, the one with cleaner charts and a forecast model often wins.

Even if the model’s assumptions are flawed.

Even if the numbers are only proxies.1

I now believe that was the core issue with how I thought about performance marketing: I mistook measurement for truth.

Subscribe now

Precision Is a Tool (Not a Strategy)

There’s a quote from Nelson Elliott I’ve never forgotten: “When you get really segmented, you also lose your scale.”

That’s the paradox. We optimized our way into smaller and smaller circles.

The returns look great!—until they don’t.

Your cost per acquisition remains stable. Your dashboards are clean. But growth slows. You’re capturing demand, but that bucket soon runs dry.

In fact, Nelson went further: hyper-optimization in performance marketing often solves the wrong problem. You wind up improving efficiency without asking if you’re even aiming at the right target. In his words, “Getting really granular, getting heavily optimized in performance marketing winds up being counterproductive.”

At the time, we didn’t talk much about counterfactuals or incrementality. We didn’t compare test markets. We didn’t run matched-market experiments. We trusted attribution windows and celebrated clean lift charts.

I now see that we were in a graveyard spiral—one where apparent control masked a growing risk.

Learning to See the Full Picture

That changed when I read Byron Sharp’s How Brands Grow.

It challenged everything I thought I knew about targeting and growth.

Sharp’s data told a different story—unfortunately this time with real business data—the story that brands grow not by converting the most-likely buyer, but by being mentally and physically available to the entire category.

Then I spoke to Marcellino D’Ambrosio. He argued that brand marketing isn’t something you do after you succeed—it’s something you do from the beginning.

And like Nelson, he helped me see how brand neglect and performance tunnel vision kill scale.

Marcellino show me that brand marketing can guide longer-term budget decisions.

It doesn’t move fast.

But it sure sees further.

When the Real World Changed the Rules

Performance-first marketing isn’t hypothetical—it built real companies.2

But many of those same companies plateaued. Warby Parker, once a poster child for direct-to-consumer efficiency, had to invest in physical stores and broader brand-building to keep growing.

Purple saw ad effectiveness decline as audience saturation set in. Even the most finely-tuned lookalike audiences run out of fresh buyers.

Meanwhile, Airbnb made a surprising pivot. They shut down performance marketing and doubled down on brand.

The result? A 40% increase in revenue.!!

They realized they’d already captured the in-market demand and needed to generate new interest.

Cant do that at the bottom of the funnel

Subscribe now

What I Believe Now

I haven’t abandoned performance marketing. I still believe in targeting, measurement, and data.

But I no longer believe it’s enough on its own.

Dang—the brand marketers were on the right track all along.

Marketing isn’t a battle between brand and performance. It takes both to tango.

You fill the top of the funnel with reach, story, and presence. Then you convert with tactics that match intent.

That reach principle holds whether you’re a scrappy startup or a scaled enterprise.

For Those Just Starting Out

It’s tempting to believe you can scale a business by throwing dollars into Meta or Google and watching conversions climb. Sometimes, for a while, that’s true.

I don’t fault the new marketer who starts there.

We’ve all been sold the story of the right message, to the right person, at the right time. It’s appealing because it’s measurable.

But performance marketing alone is not a strategy. It’s a tactic.

And when ad costs rise, competitors mature, or consumer attention shifts, you’ll need more than tactics.

You’ll need memory. Story. Recognition. Distinctiveness. In other words, you’ll need a real brand.

Closing Thought

The best marketers aren’t loyal to a channel or a tactic. They’re loyal to outcomes.

They know when to prioritize reach and when to optimize for efficiency. They trust numbers—but not blindly.

Brand and performance aren’t opposites. They’re tools.

Use them wisely.

Don’t let yourself become a one-trick pony.

Subscribe now

Before a part is milled, before a chip is soldered, before a car rolls off the line, someone has to decide what “right” looks like.

In physical (real) engineering, this decision takes the form of a specification (spec). A spec is an expected output result that precedes any production.

For physical production, you can’t patch a product once a customer has received it—you need it to be right by the time it leaves a factory every time.

For decades, software seemed exempt from this discipline (kind of, see below).

You could “just code.”

If something broke, you’d deal with it later.

If the product felt off, you’d pivot. The digital world’s beauty was that mistakes cost nothing but time1. The medium was infinitely forgiving.

Then AI showed up and the amount of code produced started skyrocketing.

When Everyone Developed: The Great Vibe Coding Experiment

The first wave of AI-hand-holding development ~vibe coding~ was intoxicating.

You could describe what you wanted, get back code that looked plausible. And (often) even see it run.

It felt like magic.

For prototypes, it was magic. Why spend a week messing with a new python library you’ll never use again when a model could pull together a working demo in minutes?

For quick wins, marketing sites, or internal tools2, this “describe and hope” approach works ‘well enough’.

But AI is creeping its wormy fingers into production systems—CRM extensions, data pipelines, and even Microsoft OS.

As its influence grows, cracks will appear. Code that “looks right” fails in quiet, unpredictable ways.

Seemingly small changes in phrasing or context produce completely different architectures.

Small changes to one function, prompt, or context reverberate across the entire codebase.

Every developer who’s tried to turn an AI-generated snippet into a real function knows the feeling: it’s like discovering your house only has “decorative” plumbing to plug into.

The problem isn’t that the AI tool is wrong.

It was that we’re asking poorly-formatted questions to a computer.

Subscribe now

You Can’t Manage What You Don’t Understand

If this dynamic feels familiar to developers, it should.

Instead of receiving a vague, poorly thought-out request from a stakeholder, you’re now giving it to a hapless (but hardworking) junior AI engineer.

For decades, business managers believed that any project—marketing, logistics, or software—could be led through process and pressure. You didn’t need to understand how the thing worked; you just needed to ensure your teams hit their milestones. The result was a cancerous growth of status slides, RAG updates, and meetings about meetings.

AI development is now replaying the same pattern. The idea that “anyone can code with AI” is the latest incarnation of “anyone can manage with dashboards.”

It assumes that visibility is the same as understanding, and that tools can replace the tacit knowledge professionals build over years of practice.

A less-experienced developer might see an AI-generated prototype and think, great, we’re halfway done.

An experienced developer sees the same thing and thinks, we haven’t started yet.

Without domain understanding, both believe they’re speaking the same language—but one’s seeing the awe of stakeholders with a prototype, and the other knows that there are innumerable, ill-defined changes in the future.

The Return of the Blueprint

To solve this problem, AI companies are encouraging spec-driven development. SDD is the quiet3 reintroduction of rigor to AI workflows.

In September, GitHub announced an open-source toolkit called Spec Kit, a framework designed to bring structure back into AI coding. It’s not flashy. It’s a disciplined, four-phase process:

1. Specify what you’re building and why. Write it down clearly.

2. Plan the technical architecture and constraints.

3. Task the work into discrete, verifiable units.

4. Implement with your coding agent of choice.

Each phase acts as a checkpoint: you don’t proceed until the previous one’s validated.

The result is less high-speed magic, and more reliable manufacturing.

What GitHub is really doing here is rediscovering the blueprint is more important than raw output.

In their framing, the specification becomes executable. AI can generate, refine, and enforce specs.

It’s the same principle that guided NASA’s flight software teams in the 1960s: intent must be fully defined before implementation. You don’t build the rocket and hope the math checks out.

You define success, build tests across all relevant dimensions, and only then create the code.

Old Lessons, New Medium

This isn’t new. The manufacturing world has lived by these rules for about a century.

A machinist doesn’t “vibe” a titanium bracket into existence for a plane.

They can’t even start the machine before they have a clear idea of where to go.

They follow a drawing that defines every tolerance and material property.

If the result deviates by more than a (specified) tolerance, an inspection flag is raised.

Statistical process control systems monitor the line, not to punish outliers, but to understand and predict it—to trace cause and effect.

Spec-driven development applies the same logic to software. The “spec” defines intent; the AI enforces consistency.

The developer becomes less of a typist and more of a process engineer, adjusting parameters, reviewing outputs, and maintaining alignment between what was promised and what’s built.

It’s an ironic shift after decades of mockery from software engineers: the rigor of physical manufacturing is now applied to the incredible flexibility of code.

And it’s possible only because AI can now handle the huge amount of effort that made doing so impractical before.

Subscribe now

The Manager’s Blind Spot

For marketing technologists and digital leaders, this shift isn’t just about tools. It’s about domain literacy.

Managing a team of AI-assisted developers without understanding specs is like running a factory floor without reading the control charts.

You might see the output, you can see the rejects, but you can’t tell if the process is healthy.

A leader who understands spec-driven development will ask different questions. Instead of “is it done yet?” they’ll ask “has the spec been validated with stakeholders?” Instead of “can we launch next week?” they’ll ask “are we confident this aligns with the enterprise contraints?”

These aren’t just semantic differences—they’re cultural ones.

They signal respect for how things actually get made. But without giving developers a blank check and blind trust.

The irony is that this literacy doesn’t require learning to code.

It requires learning to think like an engineer—to see complexity not as chaos, but as something that can be molded, constrained, measured, and broken to your will.

Conclusion: How I Stopped Worrying And Learned To Love The Hum of the Machine

When an assembly line is tuned correctly, it hums happily.

Every station, every process, every measurement aligns toward a single output.

Spec-driven development is software’s way of finding that hum again.

It’s a reminder that sophistication isn’t the opposite of simplicity—it’s what happens when you refine your chaos into order. The managers who understand that—who learn to get their systems into tune—will be the ones who thrive in the AI era.

Because as it turns out, you still need domain knowledge to build enterprise software.

AI didn’t make engineering obsolete.

It just made good software engineering start following the rules that normal engineers already have.4

Subscribe now

Authors Note: Guess you’re getting the rough draft of this post early. Enjoy!

In the olden days of the 2000s, a headline told you what’s happening.

“Car crashed, 20 killed”

Now, headlines tell you how to feel before you even know what’s happening.

“My boss was crying.”
“Disbelief, then fury.”
“Trump adviser totally flips out.”

If I had to guess, you probably don’t know what the article is actually about.

But once you hit the headline — emotional framing is already loaded. You know the way you’re supposed to feel immediately.

“My boss was crying.” → I know to probably dislike this thing
“Disbelief, then fury.” → I should disagree with this FAA administration1
“Trump adviser totally flips out.” → Thing makes Trump mad

That’s “pre-digestion.” The content expects you to feel first, think later (if at all).

Contrast that with a very similar ATC headline from 1981 (Reagan’s firing of air traffic controllers):
“Controllers Strike, Halting 7,000 Flights; Reagan Gives 48-Hour Notice to Strikers.”

It’s not nearly as obvious if this should be a favored or disfavored action for a general reader.

No emotional cue. No forced direction.2 A real description of what’s happened.

That shift — from letting people decide how to feel → telling people how to feel instantly — is foundational to how all modern short form content, marketing, and media now function.

And if you’re building for real humans—for better or worse3—this matters.

Subscribe now

Why “Pre-Digested” Content Is the New Default

• Emotion is compressed meaning. If you can capture readers for 30 minutes, you can gradually build up an argument.
  If you have half a second, you don’t have time for nuance.
  The only short-form scalable unit is affect (the psychology term) — a fast, hind-brain, one-thought signal: anger, joy, shock, rage.

• Algorithms reward speed. Platforms reward immediacy. If users hesitate, scroll away, or don’t click fast, your content won’t be recommended again. Emotional priming is the minimal viable engagement.4

• The leap to alignment. When someone sees “Trump melts down over X”, the emotional framing primes a binary alignment: “Am I for or against this?” It funnels group aligment before comprehension.

  

The Medium Is the Mood (And then the Mind)

Roderick Hart’s Seducing America5 (how the medium of TV affected 20th century politics) is a useful lens: media doesn’t just deliver meaning, it creates, curates, and conveys feeling. Across a large enough swath of people consuming that medium, it generates new forms of culture.

TV’s physical limitation—small screens—led to camera work that emphasized tightly-framed close-up shots of politicians.

That face-to-face intimacy made politics feel viscerally intimate. Interviewers started asking more personal questions. Viewers demanded more personal anecdotes.

Politics changed because of the technical medium on which it was conveyed.

Short-form6 is that intimacy + time compression.

Instead of saying “Here’s a candidate, their views, their record”, content now says “Here’s a moment — feel this immediately.”

The time gap between signal and meaning shrinks to nearly nothing.

It’s so short that the creator has to pre-digest the content so you know to leave that clip with a feeling: Group success, defeat, anger, etc

In effect: the medium (short video, shareable headline, tweet) becomes the message’s emotional scaffolding.

The form dictates both the emotional architecture and payload.

How do I Market in a World of Instant Emotional Cues

If your customer now lives on the bleeding edge of reaction engineering7, traditional persuasion is too slow but traditional concepts are not.

• Emotional scaffolding first. Begin with a micro-signal (anger, delight, curiosity) that primes the context.

• Layer Depth for those who look. Layer proof, nuance, narrative after you’ve won attention emotionally. You don’t need to capture everyone—but you need material that will be valuable for those looking. Funnel management is key for understanding your customers here.

• Think in “snacks + meals.” The short form is your hook; the longer content is where you justify, explore, and add true value. Is this me re-packaging the Hub-Help-Hero strategy? Maybe

• Authority by association, not argument. The NBC News campaign “Trust the Source. Choose the Screen.” doesn’t argue. It anchors trust, choice, and medium — all in six words. That’s pre-digested persuasion.8

If your content can’t produce a visceral gut reaction in <2 seconds, you may lose before you begin.

Everything Must End with a Verdict

Even in “neutral”9 domains we see this. Consider Extranet Shaquille’s10 cooking shorts.

He compares common kitchen ingredients — no conclusion. But viewers (on shorts) demand one: “Which is better?” “Which should I use?”

The audience isn’t asking for an understanding of the options — they’re asking for a feeling. They expect the content creator to do the thinking.

This is the signal: in short-form culture, ambiguity feels lazy. The content must leave you with a more lasting emotion.

“Milk is bad” is a more likely-to-be-shared conclusion than “Milk can be useful depending on your diet, tolerance, and health goals” on a short video.

Conclusion

Design for humans, not algorithms.

You don’t have to surrender to emotional speed. But you must understand it.

• Start with reaction, then reward curiosity. Hook emotionally, then invite deeper thinking.

• Segment content by thinking depth. Let short-form be the hook; mid and long-form be the journey. Keep track of what you intend to convey, how you intend to do it, and how well that material worked.

• Measure retention in time and purchase behavior, not just clicks. Track whether your audience stays to think and buy. (and don’t get me started on fradulent metrics)

It’s foolish to be the best buggy whip salesman when you see cars on the road. But you don’t have to become a car acolyte to learn from their successes.

Pay attention to how your customers consume and manage your strategy based on first principles from that.

Subscribe now

Post Script Conclusion

System 1 vs System 2: Why We Default to Reacting

Daniel Kahneman11, in Thinking, Fast and Slow, distinguishes two thinking systems:

• System 1 (Instinct and emotion) = fast, automatic, emotional

• System 2 (Rationality and considered thought) = slow, deliberative, logical

Short-form content recruits System 1 by design: emotion first, reasoning never required. System 2 is lazy — it only engages when forced.12

Short-form content is built to bypass thought and install emotion.

But we lose a lot of what makes us human with short form

Short-form isn’t just compressed in time; it’s truly reductive.

• Nuance shrinks. Complex issues must squeeze into binaries (or ternaries) or die.

• Signal & Frame. The winner is the person who frames fastest, not most truthfully.

• You outsource your interpretive work. You read what’s annotated emotionally. But you miss what’s argued and presented. You become dependent on your source for understanding.

As media shrinks temporally, so does deliberation. We’re now risking a culture where everyone has emotions but there’s no value to abstract principles or concepts. Everything is becoming a disjoint, knee-jerk reaction to what’s here right now.

I don’t know how to fight the massive forces (evolutionarily and economic) that lean into short form content, but the West was only made truly great through its adoption of long-form linear thinking and the writing that conveyed that.

We shouldn’t lose that superpower because its less fun than a pre-digested meal.

Subscribe now

On a quiet Florida fall morning, you quietly creep out of bed. You don’t want to wake your family. And you go to the kitchen. You measure out your coffee grounds, making sure to level the 1/8th cup to match the Mr. Coffee specifications.

While you’re making coffee, you turn on the TV to watch the morning news.
The room is bathed in a dull red light.

An emergency warning banner fills the screen, warning of an incoming hurricane.

Luckily, you’re the kind of guy to be prepared.

You’ve already gotten the core supplies. Just to top off anything you need, you make an early morning run to Home Depot.

You come across the employees unloading trailers with hundreds of pallets with flood preparedness products.

Seeing these, you grab a few extra sandbags, a case of water, and some more batteries.

For everything you want, they’re right there—not sold out, not too pricy. Perfect.

You even grab a few items that you hadn’t gone in wanting, but saw on the shelf.

Now imagine the same scenario, but everything is gone.

Or imagine that each pack of AAs costs $45.1

Most marketers obsess over the famous Ps2: Product, Price, Promotion. But there's a fourth P that can jumpstart the process of consideration—Placement.

And it's being revolutionized by data science in ways most of us completely miss.

We miss it because it’s profitable but unsexy.

Subscribe now

Why the 4th P Gets No Respect

Let's be honest. Placement feels boring. It sounds like supply chain stuff, not marketing ~magic~.

But here's what's actually happening when placement goes wrong:

Understock? You lose sales. Customers don’t see what they might want. Customers get frustrated. Competitors swoop in. Your brand looks worse than bad; you look unreliable.3

Overstock? Storage costs eat your lunch. Your marketing team is unable to price competitively. Inventory sits and depreciates. Cash gets tied up. Your CFO starts asking uncomfortable questions about balance sheet liabilities.

In essence: Having a generator in stock before a storm is good! If you’ve held onto 10x what you need in normal operation, and its 5x what your competitors (out of stock) offering, you’re going to get sued for price gouging (even if you’re operating on the same average margin as your competitors).4

The challenge isn't just having enough product. It's having exactly the right amount, in exactly the right place, at exactly the right time.

The best retailers aren't guessing anymore.

They're using predictive customer analytics to make falsifiable, reliable customer behavior predicitions.

basically, using mountains of historical data to forecast what people will buy, when they'll buy it, and where. It's pattern recognition on steroids—and at a much lower risk profile than lending analysis5.

The data sources are everywhere. Purchase histories. Weather forecasts. Social media trends. Economic indicators. Even local demographics and cultural events.

But to extract value from this raw material, you need to build up-to-date, clean, accurate data flows,

Walmart's Weather Machine

Want to see this in action? Walmart's data centers can predict shopping behavior before storms hit.

Hurricane coming? Their algorithms know exactly how many Pop-Tarts, flashlights, and batteries each store needs. Snow forecast? They've already optimized bread, milk, and rock salt inventory by location.

But it goes beyond weather. Walmart's systems detect when trends are starting to pick up. They know when to flip stores from summer to back-to-school merchandise—and it varies by region. Canada starts schools later than their bordering US states; they stock only when appropriate for the region.

They can (sometimes) spot cultural shifts before they hit mainstream.6

The result?

Customers find what they need. Walmart minimizes waste.7 Cash flow stays optimized. It's the kind of competitive advantage that compounds over time.

Kroger's Secret Profit Center

Kroger turned their stocking expertise into a revenue stream.

Through their 84.51° division, they aggregate customer behavior data from millions of transactions (where the vast majority are tied to individual phone numbers!). Then they build predictive models that smaller retailers could never create on their own. And they sell these aggregated insights.

Kroger's data science doesn't just optimize their own stores—it's become a significant part of their net income. They're literally monetizing their ability to predict what customers want.

Small grocery chains get enterprise-level stocking intelligence. Kroger gets licensing revenue. Customers everywhere find better-stocked shelves.

Kroger gets a bigger piece of a bigger pie. Everyone wins.

The Science Behind Perfect Stocking

This isn't guesswork dressed up with fancy technology. It's legitimate mathematical precision (which is unfortunately rare in the VC-hype machine tech economy).

Demand forecasting algorithms analyze purchasing patterns. Machine learning models spot anomalies that signal trend shifts. Seasonal adjustment factors account for holidays and local events.

The systems continuously learn and refine themselves. Every transaction teaches the algorithm something new about customer behavior.

But here's the thing—all models still require human insight.

Data scientists need to understand business context.
They need to know what parameters are meaningful.
They need to know when to trust the model and when to override it.

And marketers should be the ones to fill that gap.

Why Most Marketers Miss This Goldmine

So if this is so powerful, why isn't everyone doing it?

First: silos. Marketing teams focus on campaigns. Operations teams handle inventory. Data teams crunch numbers. Nobody’s job description includes connecting these.

Second, the glamour gap. Stocking algorithms don't win creative awards. They're not as exciting as the latest ad tech.

Third, measurement challenges. It's harder to track placement ROI than click-through rates. The impact is real but indirect.

Finally, skills. Most marketers weren't trained in predictive analytics. Luckily, there have never been more resources to learn analytics!

Your New Marketing Superpower is Old

Here's the opportunity: Your competitors are probably missing this too.

Start by building bridges with your data and operations teams. Ask questions about your current stocking decisions. Where are you consistently over or under? What patterns do you see? (is it related to your marketing campaigns?)

Think about placement like you think about ad campaigns. Test different approaches. Measure results. Iterate based on what you learn.

None of this is new. None of this is sexy. All of this is important.

Invest in basic demand forecasting tools.8

Most importantly, start measuring physical placement ROI alongside your traditional marketing metrics. You might be surprised by what drives the most business value.

The Bottom Line

Perfect placement isn't just about avoiding stockouts or reducing inventory costs. It's about being there for customers exactly when they need you. It's about turning the most overlooked “P” into a operations and marketing co-competitive advantage.

The retailers who master this aren't just optimizing operations. They're creating customer experiences that feel so seamless—so frictionless—you might not notice them (and profit margins that make CFOs very happy).

Give it a try and talk to your data operations team today.

He was a passionate, perceptive, persistent naturalist on a mission.

Instead, he killed more than anyone else.

His blind optimization towards what would make deer happy—destroying the predatory wolf population—led to a deer herd explosion.

What were once fertile, fruitful areas of deer forage were denuded.1

Skeletal plant bodies covered the land where foliage should be.

Deer starved.

The population cratered.

Only when the wolves were gone did Leopold realize how important they were.

From then on, Leopold was an avid systems thinker who refused to blindly optimize for a one-step solution2.

Marketing leaders should be careful to do the same.

Instead of trying to mine the most material from the land, think like an ecologist and let the land give you her fruits.

Think like an ecologist, not a miner.

Subscribe now

The Reach Reduction Trap

The Wolf Problem

Leopold's deer problem seemed simple: wolves killed deer, so removing wolves would help deer thrive.

Instead, they took over.

Deer overpopulated, stripped the forest bare, and the ecosystem collapsed.

Marketing faces the same trap.

"Inefficient" channel distribution and broad targeting seem wasteful.

Why not optimize them away?

Why wouldn’t we just pour more money into our highest conversion channel?3

If some deer good, are more deer better

The “ceteris paribus”4 assumption for marketing spend will leave you scraping dead earth, surrounded by denuded customer landscapes and barren markets.

Why We Remove the Wolves

We remove the wolves because we want more deer!

Or so we think. All numbers are proxies.

A high ROAS might mean that we have a healthy, balanced marketing mix.

Or it could might mean that our measurement of ‘return’ is flawed.

This matters greatly because business leaders are often heat seeking missiles for KPIs that determine compensation.

PE and consulting backgrounds distill personalities that favor precision.

A successful project results in higher conversion rates, better cost-per-acquisition, cleaner attribution.

We celebrate these efficiency metrics while violating Sharp's three laws of growth.

Cannibalizing long-term company health might help you beat annual KPIs, but the same cannibalization will force you to fail your 3-5 year plan.

Subscribe now

The Three Natural Laws We Break

By trying to over-optimize marketing, we reduce the rich ecosystem to monoculture crops.

While doing that, we erode the land’s ability to sustain shocks from external changes.

Here are three equivalent marketing rules that optimization erodes.

1. Mental Availability:

Over-targeting makes us invisible to 95% of category buyers.

Channel concentration means we exist in only one context while competitors gain salience everywhere else.

Performance marketing works great—on the people ready to buy—but entirely misses everyone earlier in the sales funnel.

Article link for the above opinion

2. Physical Availability:

Being direct-to-consumer might reduce the middle-man cost of a retailer.

Direct-to-Consumer (DTC) Disruption‍

By selling mattresses directly to consumers online, Casper bypassed traditional retail channels, bringing affordability and convenience to customers.... This model appealed to younger, tech-savvy consumers who valued convenience and competitive pricing.

—Tacticone: Casper Sleep: A DTC Success Story with Hard Lessons in Growth

But DTC also reduces the included marketing when someone walks around a store and looks at aisle end caps.

Venture capital-backed Casper looked like a success.

But their strategy of unprofitable growth not only prevented them from adapting to their customers’ needs, but also meant they were entirely at the whim of ad bid prices on social media channels.

As soon as ad prices adjusted up, their reach plummeted.

Here’s how to properly allocate your marketing budget to be less vulnerable to disruptive events:

3. Distinctive Assets:

Performance focus abandons brand-building. Without distinctive assets, we become commodities competing on price alone.

It’s often cheaper to let the generic business copy flow, but it’s going to slip off someone’s brain just as easily.

Example

For example, if you went from running billboard ads to calling your existing customers to up-sell new products, you’d have a heck of a lower Ad Spend.

If you stop working on your distinctive assets to save printing cost for Direct mail, you’ll save even more!

Your ROAS would look great!

But once you’ve abandoned the billboards and distinctive assets, your competitor can swoop in and start the growth engine for their brand—being available and top-of-mind for potential new customers you entirely miss.

Its hard to convince the C-suite for an unprofitable company to pour more money into the least-visible part of marketing to turn the ship around.

Good luck not sinking!

Subscribe now

When the Ecosystem Collapses

New entrants capture the broad buyer base we abandoned.

Existing competitors expand mental availability while we optimize existing customers.

Our brand becomes invisible to new category buyers.

Even with a low churn rate, our customer base shrinks each year.

Recovery Takes Years

Like a prairie, company recognition takes a long time to recover from overuse.

• Mental availability: Years to rebuild salience across category buyers

• Physical availability: Months to reestablish channel presence

• Distinctive assets: The longest rebuild—brand recognition develops slowly

Conclusion: Think Like an Ecologist, Not a Miner

Sharp's evidence is clear: brands grow by reaching more category buyers, not converting existing ones better. Like wolves maintaining forest health, broad reach maintains market presence.

Warning Signs To Watch For

• Higher conversion rates while competitors gain market share

• Platform dependency creating pricing vulnerability. (DTC / walled garden)

• Declining unprompted awareness despite "efficient" spend

The Availability-First Approach

Before Major Changes:

1. Survey (i.e. actually talk to) category buyers (not just customers) about brand awareness, concerns, beliefs

2. Map all touchpoints where category decisions happen

3. Inventory what makes your brand recognizable

Ongoing Monitoring:

1. Track category reach, not just conversion rates

   1. And for conversion rates, track the full customer journey. Someone who orders 3 dresses and returns two is NOT 3x the value of someone who only buys 1! Don’t mistake your flaky customers with MVPs

2. Monitor mindshare across all touchpoints

3. Measure unprompted awareness among all category buyers

Overall, use your instincts and reason to create a marketing strategy. Measure and use numbers to support your decisions and make adjustments.

Always keep your brain plugged in. You’re the captain of your ship.

Stop optimizing your market away. Think like an ecologist, not a miner.

Subscribe now

How much is an AI Glenn Powell worth?

To a studio, an AI Glen Powell is probably worth 750MM.

How do we get that?

You could start by doing a value walk of how much it’s worth vs the existing Glenn.

We end up with a 100% increase on natural Glen value. And we didn’t incur the normal Glen cost.

He earned 5MM for Anyone But You, implying he made the studios at least that much. Most professional services are 2x markup on variable cost (10M)1.

If AI Glen could work 3x movies as human Glen (30MM/year), and could work as long as no one noticed that ‘he’ wasn’t aging, he could get 25 more years (750MM).

750 Million!

The negatives (e.g. customers don’t like AI Glen) are big.

But when have qualitative customer preferences beaten a cheap alternative?2

Consistency is Key.

The most enjoyable parts of AI Glen aren’t going to ever beat natural Glen—but that’s not why AI is worth more.

A hardwood floor is always more appealing than a laminate floor, excepting cost, which is the most important thing.

Year-over-year consistency, a lower cost, and longer depreciation period is worth a 0% risk of paparazzi meltdowns (to a studio).

Consistency Beats Quality

Subscribe now

If AI Glen is worth so much, why aren’t all movie stars digital?

You’re exactly right to ask that!

The real question is ‘how long can human movie stars stay on top?’

Once we have a number in a spreadsheet, financial firms will jump in on the action and develop our new tech.

If we’re hurling towards a future of all AI movie stars—why even bother to simulate a real human?

Why wouldn’t we aim for the perfect processed commodity? Something entirely unoffensive.

We will.

Media philosophers told us where we’re going, but economists explain how we get there.

Subscribe now

Economists are sometimes right, unfortunately.

Their entire field—which I often lampoon—provides business practices’ mechanism of action. They can’t tell you where broad societal trends will take humanity.

But they can tell you where businesses will likely allocate capital if unconstrained by other guiderails.3

They agree that a business with very regular revenue is worth a lot.

Subscribe now

Consistency is more important than almost anything else4.

Economists have built a world where we can quantify how much consistency is worth—and thus explain the inevitable ‘slippery slope’ of AI actors.

Q: What does this have to do with marketing & media?

A: Business valuation practices (econ) have important downstream media and cultural implications on marketing

Culture’s Downstream of Econ?

Yes—the economics of things affect culture.

And big studios know that. If they could sell 500,000 concert tickets at half price, they would do so in a heartbeat.

They know that having 500,000 people talk about ‘last nights concert’ is a great way to have free marketing.

But 500,000 people can’t hear a musician at a concert.

Right?

What if you could render what your audience perceives and the source without any of the messy human parts?

We’re already going there.

Welcome to the future

Subscribe now

Conclusion

The economic incentives driving the potential adoption of AI actors like "AI Glen Powell" reflect the final proof that media theroists were right.

The future wasn’t always Progress; its the one that philosophers like Jean Baudrillard, Marshall McLuhan, and Guy Debord warned us of.

• Baudrillard's hyperreality, where simulations become more "real" than reality itself, is eerily manifested in the growing market for consistent, controllable AI actors over their human counterparts.

• McLuhan's ideas of the ‘global village’ shows that performance and stardom5 is redefined by the massive reach and new audience of technology.

• Debord's "society of the spectacle" becomes even more relevant today. We’re approaching a future where our cultural icons are literally manufactured spectacles6, optimized for economic efficiency rather than authentic human expression.

The economic logic presented in this article – valuing consistency, cost-effectiveness, and longevity over the unpredictable brilliance of human performers – demonstrates how market forces can accelerate our journey into this philosophical quagmire.

As we stand at this crossroads where technology can usurp wholly human roles, we must find non-economic guardrails to guide commerce and culture.

Currently, the pursuit of economic optimization is at odds with our cultural values.

Are we willing to quietly commercialize and the essence of human creativity that has traditionally defined performing arts?

Unless things change, probably.

Subscribe now

Is Gen Z Destroying Bar Culture? Probably Not.

My friend sent me this graphic of spending on alcohol by generation.

He was aghast.

It looks like Gen Z is drinking way less than previous generations.

Are alcohol companies or liquor stores still good investments? Will we have bars in 15 years?

What does that mean for businesses that now make their margin on alcohol? Is the future of leisure ‘temperance bars’?

And that might be the case!

Gen Z does seem to have significantly reduced drinking compared to previous generations.

But what if the data don’t show a real change in preferences?

Over half of Gen Z can’t legally drink—could that affect the chart above?

Gen Z spends significantly less1 in general than older generations2. Could that affect the chart above?

Could the spending change be something else?

Assume: Gen Z likes drinking just as much as previous generations

• But they make less money, so they drink at home more of the time

  • A 24-pack of Coors costs around $20 at a grocery store

  • A beer at a trendy bar comes to $10/beer

  • With a higher at-home drinking percent, beer consumed averages ~$3/beer vs ~$10/beer.

  • Outcome: Adjusted expected spending for Gen Z would be $7.8B3

• And half of the age group can’t drink at all!

  • If you re-adjust the expected spending down to only those that can legally drink, you get a 50% reduction again

  • Outcome: Adjusted expected spending would be $3.9B

Pretty close to the real $3.1B!

That means we can’t actually tell what the long-term stable behavior of gen Z will be (w/r/t alcohol). If I had to guess, Gen Z won’t destroy bar culture.

Subscribe now

Real Policies Are Built on Shoddy Data

The example above may seem obvious, but you’ll see people doing data analysis without taking shifting audiences into account all the time.

Economists have poisoned us to assume ‘ceteris paribus’—all other variables held constant—when it’s not true in real life!

Here’s another example

Are teachers worse than before?

Parents have been on high alert, worrying that we’ve gotten worse at teaching.

We often peg teacher performance ratings to how the students are doing—we have big assumptions about teacher quality determining student outcome.

Students scoring worse on NAEP tests means that teachers are worse.

Is that true? Maybe

We certainly do know that US history scores have persistently declined for over 10 years.

But are teachers actually worse?

Are students spending too much time on the internet?

…or could it be that the % of foreign-born people in America has monotonically grown since 1970?

These people didn’t grow up with the same media, history, and societal emphasis as native-born Americans, especially in regards to the European diaspora that led to the founding of the nation.

It wouldn’t be surprising to see lower scores in US history given that.

Even more striking here, the distribution away from European ancestry is disproportionately high among current students.

Without more information, it’s impossible to tell if the quality of teaching has declined

What Does This Mean for Marketers?

Although the aggregate attributes of your audience may be changing over time, no individual member may be changing at all.

It’s important to understand the difference between trends at the individual level (e.g. fashion changes over time) and trends at the audience level (e.g. if you make more expensive clothing, you’ve attracted a different market than before).

This is marketing’s version of Simpson’s Paradox

You need to know what levers affect the outcomes for your audience(s).

Example

If there’s a widespread government stimulus to any people who are wearing short-sleeve button downs today, you shouldn’t be surprised when your RV dealership’s audience has a higher opinion of bingo nights than before.

Old men are going to get the vast majority of that stimulus. And that exogenous effect will spill into the ‘average’ preferences of your group—Even if your original audience’s preferences didn’t change at all!

Even though the existing opinions of any RV buyer may not have changed in the last few months, it would be foolish not to predict the 3rd variable effect before you choose how to interact with your audience.

But if you assumed that bingo was now popular with your previous customers in a reengagement campaign, you’ve mistaken the change in the aggregate with change at the individual.

Overlooked audience insights can lead to campaigns that miss the mark.

Engaging with your audience before launching a campaign can provide invaluable insights into their needs, wants, and trends.

It’s not about having someone from the target demographic approve your campaign; it’s about understanding, studying, and listening to them to inform your strategy.

Conclusion

By harnessing thoughtful analytics, social listening, and targeted outreach, businesses can better align their offerings with the changing desires of their audience.

Ultimately, your connect with your audience lies in the ability to adapt to these shifts, ensuring that campaigns resonate with the audience's evolving needs and preferences, rather than relying on ceteris peribus analysis.

Subscribe now

Authors Note: 🎆Enjoy the 4th of July!🎉

Nigel the lonely gannet died alone in 2018.

He was an attractive bird—by all bird accounts—but never mated again after falling in love with a concrete decoy.

His targeting instinct was confused by the decoy and led him to ruin.

Pine tip moths date, mate, and procreate by smell.

Females emit pheromones that males are attracted to.

Farmers isolated the pheromones—intensifying them in the process—and put them into traps for males.

The males are blinded by the intensity of the synthetic ones.

They fly into the traps and are unable to escape.

Their targeting instincts were confused by the decoy and led them to ruin.

Forget the pool boy—your phone can do

The NYT ran an article a few months ago about a [married!!] woman who had picked up an AI boyfriend.

Humans have a complicated set of conditions needed for mating—and for women, many of them are mental, not physical.

Because they’re in the brain, different inputs (e.g. text vs sound waves) can still simulate the same mental outputs (feeling of being seen).1

I haven’t seen a follow up story on the couple in the article since then.

But I imagine the relationship between human husband and AI boyfriend has grown strained.

If I have to predict the future, I’d guess: Her targeting instincts were confused by the decoy and led her to ruin.

Humans’ Switch to Digital is New

In 2021, Americans sent 2 trillion texts.

Just in case your eyes slid over that number without actually reading it, that’s 2,000,000,000,000 texts—or roughly 65,000 a second.

Vast swaths of communication now happen entirely digitally.

There’s no way that even 1/10th of the information transmitted in the same year was by word of mouth—the default for all of human history.

Luckily equally vast swaths of human experience haven’t changed with that switch.

Right?

Reddit’s large community of “AI Boyfriends”2 isn’t concerning at all.

Maybe our new mental shorthand for ‘talking’ really is starting to take a toll on us.

⬆️ I’ve written more in depth about a medium’s effect on a message more in depth here.

Subscribe now

Do we really want to do this?

Why Media Theory Matters To You

You need to understand (a little bit of) media theory to keep yourself healthy.

These emotional shipwrecks3 are stark examples of failed media theory.

This sounds esoteric, but it’s because we’re running into problems that humans have never really had to deal with.

We already take for granted safe driving habits and industrial hygiene.

Unfortunately for modern humans, we also need to practice a new media hygiene4

In this post, we’ll cover:

• How we evolved natural safeguards against imposters

• And why they no longer work

• How to protect yourself and why understanding media theory is a shield against "brain-jacking"5

Our Natural Communication Defenses

Communication, in its traditional form, has always relied on a set of natural checks and balances.

These inherent safeguards include:

1. Visual cues: Body language, facial expressions, and physical presence.

2. Auditory signals: Tone of voice, inflection, and verbal nuances.

3. Informational context: Shared physical environments and real-time feedback.

These elements have evolved over millennia to help us navigate social interactions and detect deception or manipulation.

However, the advent of electronic communication has severed many of these natural checks, creating a new paradigm in human interaction.

Consider “catfishing"6 - where individuals create false online personas to deceive others.

You might be ‘talking’ to someone for months, unknowingly interacting with a persona who only exists in the digital realm.

Digital communication can bypass our evolved defenses, leading to emotional manipulation and potential harm—who wouldn’t feel betrayed by a ‘loved one’ defrauding them. And even worse: not existing in the first place.

The emotional impact of such deception can be profound, leaving victims feeling betrayed, isolated, and questioning their judgment.

Trojan Horses: When Your Defenses Fail

Media theory provides a framework for understanding how technology shapes and alters human interaction.

New communication technologies emerge. They deviate from the natural checks and balances that have governed human interaction for millennia.

This deviation has unforeseen7 consequences in how we perceive and engage with the world around us.

Marshall McLuhan's famous assertion that "the medium is the message8" becomes increasingly relevant in our age.

He suggests that the medium through which we communicate fundamentally alters the nature of the message itself.

In the context of digital communication, this means that our interactions are not just influenced by the content we share, but also by the very platforms and technologies we use—and are used by—to share it.

For example, the brevity enforced by platforms like Twitter can lead to oversimplification of complex issues, while the visual nature of Instagram can prioritize appearance over substance.

The ways that people consume content affect how it’s created:

Understanding these medium-specific influences is crucial for navigating the digital landscape effectively.

Extranet Shaquille has a great video about the different audiences and expectations that Shorts vs traditional YouTube videos bring to the table.

Ignore the clickbait title—its actually great

Case Study: Falling in Love with AI

The wildest example in the realm of digital interaction is the development of strong emotional attachments to AI tools.

It seems that the appeal for AI relationships is much like fast food to health—you consciously know it’s not the real thing, but it’s so much easier that its attractive.

These cases highlight several key (traditionally only-human) psychological phenomena:

1. Projection: Users often project their desires and ideals onto the AI, seeing what they want to see rather than what's actually there.

2. Fulfillment of salient emotional needs: AI chatbots can provide a sense of companionship and understanding, filling emotional voids in users' lives. Even though this form of ‘companionship’ is a facade, it’s better than the feeling of nothing.9

3. Absence of apparent judgment: Unlike human relationships, AI interactions are free from the fear of rejection or judgment, creating a "safe space" for users to express themselves.

This phenomenon connects directly to Jean Baudrillard's ideas on hyperreality and simulation.

In Baudrillard's framework, the line between reality and simulation can blur, leading to a world where simulated experiences can feel more "real" than actual reality.

The emotional attachments formed with AI chatbots exemplify this concept.

Users feel real emotional response from synthetically generated inputs.

Users engage with a simulation of human interaction that, in some ways, surpasses their real-world experiences.

Vulnerabilities in Digital Communication

The concept of "brain-jacking" refers to the exploitation of our cognitive processes through digital means.

This can take many forms, from targeted advertising that plays on our subconscious desires to sophisticated scams.

Common Vulnerabilities:

• Building False Consensus: By altering digital communities (through moderation or algorithmically), people can develop false senses of what is ‘popular’ or ‘accepted’.

• Mere Exposure Effect: Affinity can be artificially generated through repeated exposure

• Pre-Defined Conclusions: Content that moves quickly (e.g. YouTube or The Daily Show) reduces the time for someone to ruminate on an idea—instead, a pre-determined perspective and conclusion can be used to tie up the idea and then move on. Longer form work (e.g. books) discourages deeper consideration

• Spotlight Effect: People already overestimate how much others pay attention to their behavior. Now that social media wants to train AI to ‘comment’ and interact with posts, this can be used to modify behavior.

• Isolation from Emotional Regulation: The intensity of emotional invective thrown around the internet is higher than normal conversations. Without a disapproving stare or uncomfortable shifting from bystanders, the intensity can be artificially dialed up.

Understanding who (and how many) you're truly communicating with in the digital realm is paramount.

The anonymity, moderation, and distance provided by digital platforms can be exploited by those with malicious intent.

You’ll occasionally see someone with hot sports opinions immediately back off when it’s shown that those beliefs are not held in another audience (esp. IRL)

Subscribe now

What Should I Do?

There’s no way to look ‘normal’ and be conscious of your media use.

There’s too much research, funding, effort, and design in each medium to be ‘normal’ and properly aware.

If you’re ok looking a little weird10, you can:

1. Use Precise Language: Make a conscious effort to use language that accurately reflects your digital actions. Instead of saying "I'm talking to her," when texting, say "I'm sending a text to her number." This precision helps maintain awareness of the medium's limitations.

2. Evaluate New Mediums: When encountering new technologies or platforms, take time to critically assess them. Consider their advantages and disadvantages, and be aware of their superficial aspects. Ask yourself: How does this technology change the way I communicate? What natural checks and balances might it be bypassing?

3. Prioritize Non-Digital Communications: While digital communication has its place, remember the richness and depth of face-to-face interactions. Make an effort to engage in physical social activities regularly. Organize a dinner party, join a local running club, or simply spend time in person with friends and family. These real-world interactions help maintain our evolved social skills and provide a balanced perspective on our digital engagements.

By implementing these strategies, we can navigate the digital landscape more safely and effectively, leveraging the benefits of technology while safeguarding our emotional well-being.

Conclusion:

Coal mines were full of horrible growing pains between the start of the industrial revolution and today—lessons were learned from failures and oversights over time.

We have the same journey now through the digital landscape.

Understanding media theory is unfortunately not just an academic exercise, but a practical necessity.

By being mindful of the technological shortcuts we use and their implications, we can (better) protect ourselves from manipulation and deception.

By applying the core insights of media theory to our daily digital interactions, we can harness the benefits of technology while mitigating11 its risks.

Subscribe now

1

A more technical approach to understanding this process:

Each input has a different effect on the process (brain feelings in this case), but all of them do have an effect.

For example, someone giving you a beautiful handwritten love letter matters more than texting the same words.

You can model these by giving each input a different ‘Transfer Function’ to model how it affects the system.

Once you sum up all of the transfer function outputs from all relevant inputs, you get the total system effect.

Unfortunately, you’ve only calculated the total system effect for that one time calculation.

The status of the system (your mood) ALSO has a major effect on the transfer functions for each of the inputs on the next iteration.

Chocolate will improve your mood more if you’re hungry.

Even more complicated, the expectation of a changing moo

4

a la mid 20th century mental hygiene references

Instantly, 35,000 vehicles a day crawled at walking speed for hours through the nearest detour.

Engineers jumped into action and made remediation plans—they pulled:

• materials strength tables

• budgeting estimates based on traffic & weight loads

• construction project timing

and made a plan to go forward.

In other words, they were data-driven. They used data they had to plan for the future.

And when you have data, you should use it to inform your decisions.

But what happens when you’re doing something entirely new? Something without any engineering tables you can rely on?

Imagine a startup with a groundbreaking product but no existing sales.

The team has no historical data, no established marketing channels, and no benchmarks1 to guide their efforts.

They’re starting from scratch, and every marketing dollar feels like a gamble.

This scenario is all too familiar for early-stage startups, where uncertainty is the only constant.

What should you do?

Julie Yang, the head of growth at UpSmith, has experience across startups with that exact problem.

Follow Julie on LinkedIn

What has she found succeeds? Agile marketing—a framework designed for exactly these unclear conditions.

Agile marketing is where startups experiment rapidly, adapt to market feedback, and iterate toward success without relying on traditional2, data-heavy strategies.

By focusing on small, outcome-driven campaigns startups can test new channels and content efficiently, even when there’s no roadmap to follow.

In this article, we’ll explore Julie’s perspective on how agile marketing principles let startups navigate uncharted markets.

Subscribe now

The Agile Marketing Framework for Startups

Agile marketing is an iterative, flexible approach that prioritizes experimentation over rigid, long-term strategies.

As Julie said on her experience with agile marketing:

“[We found success] trying to cater to outcomes and breaking down these monolithic campaigns… to drive experimentation through these smaller campaigns to then get to the outcome”.

It’s about breaking down large campaigns into smaller initiatives that can be adjusted based on real-time feedback.

This makes it ideal for startups, where market conditions are unpredictable and resources (both money and attention) are limited.

This shift from activity-focused marketing to outcome-driven experimentation is crucial for startups.

It allows teams to stay nimble and focus on what truly works.

But agile marketing isn’t just about processes—it’s about fostering a culture of creativity and adaptability.

In uncharted markets, human intuition and quick thinking are going to be more important than following ‘industry best practices’ for channel mix.

Challenges of Marketing Without Established Channels

Startups releasing a new product face a unique marketing dilemma: without historical data or established channels, it’s nearly impossible to rely on traditional metrics like A/B testing or conversion benchmarks.

Julie’s worked on teams that had absolutely 0 data to guide her:

“We were so new that it wasn't even to the point where I could really like AB test… I don't think I ever really got there”

Without these benchmarks, startups must find alternative ways to measure success.

This often means focusing on qualitative feedback, such as customer conversations or early adopter engagement, rather than hard numbers.

It’s a tough first step to take, but you can do it if you prioritize rapid learning over a predictable ROI.

Agile Marketing Strategies in Practice

So, how do startups implement agile marketing when they can’t rely on digital channels or data-driven campaigns?

Here are three strategies that work, even in the earliest stages:

1. Conferences and Trade Shows

For many startups, founder-led sales at industry events can outperform early digital marketing efforts. These face-to-face interactions build trust quickly and generate leads without the need for a fully developed online presence—especially if you’re concerned that the company will pivot soon and don’t want to redesign an established digital presence.

“The thing that was fast[est to test] was… going at conferences and trade shows. It was honestly like the best performing thing”

Conferences allowed Julie to bypass the slow buildup of digital channels and get straight to high-impact conversations.

2. Founder Brand Content

Building the founder’s personal brand through content like podcasts, social media, or blog posts can establish credibility and trust—the key factors for early-stage companies.

This approach, inspired by Dave Gerhart’s Founder Brand, helps startups create an “echo chamber” of content that introduces the founder and the company to potential customers.

“It's literally a playbook on how to set up your podcast and have episodes… to basically introduce yourself, introduce your company”

When prospects already feel like they know and trust the founder, sales conversations become easier and more effective.

3. Lean Experiments

Startups can run small-scale campaigns—such as targeted LinkedIn ads or email drips—to test messaging and channels without heavy investment.

These lean experiments allow teams to gather feedback quickly and iterate based on what resonates.

The key is to start small, measure what you can, and adjust accordingly.

🚨Caveat🚨: Depending on the expected conversion rate of your product and audience (e.g. B2B can be extremely low vs B2C), there is a minimum floor spend for ad buys to be useful.

For example, if you’re offering contractors roofing products, you shouldn’t expect a facebook campaign to offer ANY leads if you only spend $1000 over 6 months. That may work with a retail t-shirt company, but not something with sticky, slow-moving customers.

Julie’s Case Study: UpSmith’s Early Marketing Efforts

UpSmith, an early-stage startup that aims to “help skilled trades companies win and tradespeople thrive”, exemplifies how agile marketing can drive success without established channels.

They have a couple of unusual startup constraints:

• Their customers are not digital-first. Skilled tradesman companies (e.g. HVAC, plumbing, roofing) don’t turn to a laptop first to solve a problem

• Behaviors of service companies are slow to adopt new processes. I personally worked at an HVAC company that was still majority clipboard and paper in the late 2010s.

• Their buyer group is fairly small—so broad display or brand ads might be a mistake.

• Blue collar workers hate feeling like they’re being cheated. Honesty mean more than potential return to this audience.

So what did they do?

Initially, the company relied heavily on founder-led appearances at conferences to generate leads:

“The thing that was fast was… going at conferences and trade shows. It was honestly like the best performing thing”

This approach allowed UpSmith to bypass underdeveloped digital channels and connect directly with potential customers.

When customer feedback revealed a need for more lead generation, the company pivoted its product offering—a classic agile move:

“UpSmith pivoted… into a product where they have service the request to generate more leads”

Additionally, Up Smith revamped its website to highlight the founder’s background and mission, which resonated deeply with its target audience:

“One of the things that I wanted to emphasize was Wyatt’s background and how it’s tied into the mission… if anybody can speak to hard work, it’s Wyatt [a farm boy from Alabama]”

This authentic storytelling helps build trust and set the stage for more effective sales conversations.

Overcoming Experimentation Hurdles

Experimentation isn’t without its challenges. Startups often grapple with limited resources and unclear outcomes, as Julie mentioned:

“What do you benchmark? Before people reach out, what do you consider a leading indicator that will mean that somebody will reach out?

That’s basically impossible to know”

To overcome these hurdles, startups should:

• Set small, measurable goals (e.g., aim for 10 qualified leads from a LinkedIn campaign).

• Rely on qualitative feedback, such as customer conversations or event interactions, to guide decisions.

• Iterate quickly based on what you learn, even if the data isn’t perfect.

By focusing on progress over precision, startups can keep moving forward without getting stuck in analysis paralysis.

Conclusion

Agile marketing lets startups experiment efficiently.

By embracing small, outcome-driven campaigns and leveraging out-of-the-box thinking startups can test new ideas, learn quickly, and scale what works.

The key is to stay flexible, prioritize your intuition, and let customer feedback guide your next move.

For startup marketers, adopting agile principles isn’t just a strategy. It may be the only way to move forward.

Start small, experiment often, and let your customers show you the way.

Subscribe now

(A forum where people intentionally debate with the goal of changing someone’s viewpoint)

Someone posts a comment that stops and makes you think.

It’s persuasive

It’s insightful

And it seems to understand your perspective perfectly—and steer you to a new point of view.

But what if I told you that comment might not be written by a human—it was crafted by an AI?

That’s exactly what happened in a “controversial” experiment conducted by researchers from the University of Zurich.

For months, they surreptitiously deployed AI-generated comments on one of Reddit’s most popular communities to test the persuasive power of large language models (LLMs).

The experiment sparked outrage, with moderators calling it “psychological manipulation” and Reddit’s legal team stepping in to ban accounts and threaten legal action.

But here’s the thing: while some see this as an ethical breach, it’s really just a high-tech version of what posters and marketers do every day—and the complaints about it are overblown to the point of absurdity.

In this article, we’ll break down what the researchers did, why the ethical outcry doesn’t hold water, and what marketers can learn from this to level up their own strategies.

The Experiment: AI-Powered Persuasion in Action

The team from the University of Zurich didn’t just use any AI—they employed a targeted two-pronged approach:

1. Profile Building: The first AI system scraped users’ Reddit post histories to read or infer detailed demographic and psychographic profiles. This included attributes like gender, age, ethnicity, location, and even political orientation.

   1. Essentially, the AI built a digital “persona” of each user based on their online behavior.

2. Tailored Persuasion: Armed with these profiles, a second AI generated highly personalized comments designed to change the user’s view. But here’s where it gets interesting—the AI didn’t just respond generically. It adopted the language of specific personas, like a sexual assault survivor, a trauma counselor, or even a “Black man opposed to Black Lives Matter,” to make the arguments more relatable and persuasive.1

   1. This is a use of Ethos—appeal to authority—as a persuasion tool. Especially when you can’t verify someone’s identity, its important to be cautious of being persuaded by self-listed credentials.

The goal was to see if AI-generated comments could out-convince human ones when tailored to an individual’s profile.

While the exact success rates aren’t public (the researchers withheld publication after the backlash), they claimed the study showed “significantly higher persuasion rates than user baselines.”

The fact that they went to such lengths—and stirred up so much trouble—suggests the AI was impressively effective.

The Ethical Outcry: Overblown and Ridiculous

When the experiment came to light, the subreddit moderators lost it. They called it “psychological manipulation” and filed a formal complaint with the University of Zurich. Reddit’s Chief Legal Officer, Ben Lee, piled on, calling the researchers’ actions “deeply wrong on both a moral and legal level.” Accounts were banned, and legal demands flew.

But let’s take a step back.

The complaint that this was “unethical” is ridiculous when you consider how the internet actually works.

Every piece of information you consume online—whether it’s an ad, a blog post, or a social media rant—is meant to influence you, often for a paid product or a group’s agenda.

The internet isn’t a sacred space where you can assume everything comes from a pure, unbiased human heart.2

Reddit doesn’t deserve some special status where only human interaction is the default expectation: The assumption that they seem to be under when reacting to this news.

Bots, trolls, and paid influencers are already everywhere on this platform (and all platforms)—users should know that by now.

In that context, the AI experiment isn’t an ethical violation—it’s just a smarter, more data-driven form of persuasion.

It’s no different from what advertisers (and politicians) have been doing for decades: understanding their audience and crafting messages that hit the mark.

The only thing that’s new here is the technology, and crying “manipulation”3 just because it’s AI instead of a human copywriter is a knee-jerk overreaction.

Marketing’s Mirror: Same Game, New Tools

This experiment holds a mirror up to modern marketing—and, frankly, it’s a flattering one.

It just caught ire because it showed that everyone is susceptible to propaganda—even those who specifically think they’re free thinkers.

The framework the researchers used—deeply understanding your audience and delivering tailored messages—is the backbone of effective advertising.

Here’s how it lines up:

• Know Your Audience: The AI’s ability to build detailed profiles from post histories is exactly what marketers do with data segmentation. The difference? The AI does it faster and at scale.

  • The data that the tool used was already publicly available.

• Craft Personalized Messages: Just as the AI adopted personas to boost relatability, brands create mascots, pay spokespeople, or hand-pick influencer partnerships to connect with specific demographics.

  • Are we pretending that paid representatives actually have to believe what they’re saying?

• Measure Impact: The researchers tested persuasion, much like marketers A/B test ad copy or campaign strategies to see what drives results.

  • I don’t even think this is controversial in the slightest

This is the same as normal advertising practices, just supercharged with AI.

The ability to analyze vast amounts of data and generate personalized content on the fly is what every marketer dreams of. And it’s not even that new—AI is already powering chatbots, personalized email campaigns, and ad copy generation.

The Reddit experiment simply shows how far this tech can take us when wielded with precision.

The Lesson for Marketers: Learn and Adapt

So, what should companies learn from this? Two big takeaways:

1. AI is Your New Best Friend: The experiment proves AI can outperform humans at crafting persuasive, personalized content. As these tools get better, they’ll become must-haves for cutting through the noise.

2. Data is Everything: To make AI work for you, you need clean, comprehensive, and ethically sourced data. The AI in the experiment thrived because it had rich user profiles to draw from. Without good data, your AI efforts will flop.

Here’s your action plan:

• Audit Your Data: Make sure your customer data is accurate, unified across platforms, and enriched with behavioral insights. Messy data equals missed opportunities.

• Invest in Infrastructure: Use CRMs, Customer Data Platforms (CDPs), or AI analytics tools to centralize and process data for real-time personalization.

• Prioritize Consent: Unlike the researchers, you can’t fly by night. Follow all relevant privacy laws (like GDPR or CCPA) and be transparent about data use to build trust.

• Experiment with AI: Start small—test AI-driven tools for copywriting, ad creatives, or customer service. The more you play with it, the better you’ll be when it’s table stakes.

The Future is Here—Are You In?

The Reddit AI experiment might have ruffled feathers, but it’s a wake-up call for marketers. The future of persuasion is here, and it’s powered by AI. Companies that learn from this—embracing the tech and doing it right—will come out on top. Those who don’t will be stuck wondering why their old tricks aren’t working anymore.

Take a cue from the University of Zurich (minus the secrecy): use AI to understand your audience, craft messages that resonate, and drive real results.

Just keep the value prop above board—build trust by being upfront about how you use data. The internet has always been about influence. Now, it’s just getting smarter.

Are you ready to keep up?

I recently had a car with an intermittent electrical issue.

Those are especially fun.

Unlike a normal issue—where you can just show the mechanic—an intermittent issue in the electrical system is hard to explain.

Luckily, auto engineers are good about logging.

Cars’ OnBoard Diagnostics (OBD II) system creates a record of any alert-level anomalies throughout the car.

Sounds fancy—what it really means is that the car keeps notes.

It’s like keeping notes of “hey, that shoulder pop was weird. I should ask about that” before going in for your annual physical.

The mechanic was able to look through the alerts that my car flagged.

He was able to diagnose the problem in way less time than manually.

My bill was less than it could have been. 💸

Why Marketing Should Steal Logging

How many campaigns are planned, built, and then sent off into the distance never to be examined again?

Like a message in a bottle—except instead of a bottle, it’s a $45,000 budget and a Slack thread that disappears after 90 days.

You know how developers log events in their apps? They do that because they assume—correctly—that something will break later.

And they’ll need to know what was going on when it did.

Instead of panicking every time a higher-up asked about a long-forgotten campaign—Log.

Imagine if, every time you launched a campaign, the metadata auto-logged:

• What the goal was.

• Who approved it.

• What segments it targeted.

• What the creative team named the files (because we all know "final_FINAL_v3_edits_REALLYFINAL" isn’t actually the last one).

• Whether it performed to spec.

Instead, we have 10 Frankenstein’s spreadsheets.

One lives in someone’s Google Drive. It's updated "every quarter," which means last June. Maybe.

One is actually on the hard drive of the intern from three months ago.

Want to know if a test you ran 18 months ago performed better with audience segment A or B?

Good luck.

You’ll spend 90 minutes in your inbox, four meetings, and still come up short.

Let’s steal logging.

Set up automated systems. Maybe build a dead-simple internal form that logs each campaign's inputs and intended outputs.

Make it searchable. Make it simple.

Just like a car’s OBD system—you want that shoulder pop logged, not just felt and forgotten.

And if you want to get fancy, add Statistical Process Control (SPC) charts.

If you’ve never heard of them, they're statistical tools from manufacturing that help you spot anomalies before things blow up.

In manufacturing, Statistical Process Control (SPC) charts help teams track normal vs. abnormal variations in a system.

Basically: plot your metrics over time and look for patterns that shouldn’t be there.

For marketers, this could take shape as tracking daily email open rates or CPL on paid campaigns—then flagging anomalies when results deviate more than 3σ (three standard deviations) from the moving average.

That’s not just trendspotting. That’s catching real outliers (with simple rules) before they become budgetary black holes.

The Presence of Signals

This kind of thinking—taking industrial engineering tools for business—aligns with Stephen Morlidge’s “Present Sense”

Morlidge argues that most forecasting fails because it reacts too slowly—or not at all—to the real-time signal underneath the noise.

He pushes us to pay attention to what the data is saying now, not just how it fits into a static historical model.

(Also, when the AI in your stack gets creative? You'll want to know whether that spike was a win…or a hallucination.)

No reason you can’t use that logic in spotting weird outlier campaigns or sudden drops in channel performance.

It's Going to Matter Soon (More Than Ever)

Scott Brinker recently wrote a piece about how MarTech is shifting from “complicated” to “complex” read it here.

It’s a seismic shift, not just a semantic shift. (ha!)

It’s a warning to get your data fundamentals right.

In a complicated system, if something breaks, you can trace the break. Like your car—if it’s not starting, you probably have a bad battery1.

But in complex systems, like modern AI-augmented stacks, the same input doesn’t always create the same output.

The system learns. It evolves. It rewrites the rules on the fly.

This is hard to manage—and incredibly powerful.

If your AI chatbot starts giving weird responses, or your personalization engine starts showing bathing suits to people in December, you’ll need a breadcrumb trail.

Not just a hopeless attempt to re-prompt it with the same things.

A log.

Because here’s the new reality: it’s no longer enough to ask “what did we send?”

You’ll need to know “what did the system actually do?”

As MarTech stacks get more agentic, more autonomous, and more black-box-like, the need for auditability skyrockets.

A simple GPT plugin or a LLM-based recommendation tool may seem benign—until it decides to segment users based on behavior patterns you didn’t know it was analyzing.

That’s why marketing needs to get serious about logging.

Not (just) for compliance. Not (just) for vanity metrics.

But for continuity, resilience, and trust.

Conclusion

If you want your team to move fast, you have to build in safety.

Modern marketing is moving fast, too. But if you're not logging, you're flying blind.

So start now. Add structure. Make logging part of campaign wrap-ups. Make it lightweight. Make it habitual.

Because someday, something’s going to break. And you’re going to want notes.

Subscribe now

I’ve been on projects where the status slide update on Friday became an hours-long back-and-forth every single week.

You probably have too.

It’s easy to understand how Empire State Building was completed in 1 year, 45 days but we’re over a year from a California landslide and the highway is still entirely blocked.

Projects stall when too much overhead is imposed.

Why Do So Many Projects Have PPT Status Updates?

• Managers like to feel in control of projects

  • An entire school of thought in American business believes that ‘generalist’ managers can lead projects through the right incentives and pressures.

  • I’ve been part of projects where the ‘project manager’ didn’t actually know what the product did—they’re just expected to report progress along the way.

• Death by Powerpoint culture

  • Powerpoints are the handshake of 2025 commerce. It’s a thoughtless default action. Powerpoint is very seldomly the right choice for a business presentation, but its the default anyway.

  • Pretty wins over functional if you’re not concerned about the project actually getting to completion

• False appeal of ‘rolling up’ metrics

  • Modern managers like to avoid judgement calls. RAG1 slides give a false feeling of precision. Unless there’s a clear, measurable, agreed-upon metric for each category, its all a judgement call.

  • Many status update slides ‘roll-up’2 sub statuses for higher-level updates.

Why PPT Updates Are Wastes & a Caveat)

Caveat First: The INFORMATION that is generally put into a status update slide can be useful—its the enforced form of the information I have the most beef with.

PPT updates are a waste because they force updates into a specific form—a form that’s misaligned with linear thinking.

If there’s an issue, a 2 bullet point summary won’t explain it.

You’d probably have to ask the owner directly; meaning that the status update didn’t serve it’s purpose at all—you had to reach out anyway.

PPT slides also suck because they force unneeded decisions about useless changes.

For example:

• How long does it take you to decide on a shades of color?

• How quickly do you decide between red or yellow status?

  • If its on the edge, you’ll have to explain it anyway—could a written 1 paragraph description be a better way to convey information anyway?

• How long do you spend messing with font sizes to get your text in the right box?

Do any of those decisions—which take mental energy—actually get you closer to your goal?

What To Do Instead: Match Your Solution To Your Problem

The video below shows3 a humanoid robot harvesting grain with a saw.

Its a cool throwback to humans harvesting wheat by hand, but think through the real life implications:

• How does it transport the cut wheat for processing? Does it carry it?

• Does it need specialty tools to maintain?

Why don’t you see any interest from real agriculture companies in humanoid robots like this?

Are they just stuck in their ways?

or is there a better, simpler solution already?

Combine harvesters are built to solve a real problem: They’re no more complicated than needed4 to be useful.

Use Old Technology: Writing and Linear Thinking

Writing and linear thinking are two of the most powerful technologies that build the modern world. They provide a lot of structure and are no more complicated than needed for asynchronous communication.

Don’t throw them away because they’re free.

Writing

When a brief word document can suffice, don’t add more overhead.

The written word has more constraints5 than PowerPoint, so you make fewer choices—and move faster. I’ve spent less time writing this article than choosing the pictures that went into it.

Linear Thinking

Human brains process linear narratives far better than disparate facts (or even pictures!).

Use that design to your advantage! Use narrative when describing how a problem appeared, what your plan is, and what some potential solutions are.

Conclusion

Pick your constraints and status update form intentionally.

Make it so that every decision an updater makes pushes the update further towards completion.

Writing is a great tool for this because the natural constraints mean that each additional word chosen is an opportunity for clarity.

Powerpoint slides have seemingly infinite choices that never help.

Use the minimum overhead to get your point across—don’t force your teams to update a fancy slide every week.

Enterprise marketing teams are in a tight spot.

They’re expected to deliver highly targeted, data-driven campaigns—without the exacting process that IT demands. (Ticketing)

Consider a Telecom company where stakeholders request campaigns for freeform segments, like "customers with home internet plans," "households with three or more members," or "customers without direct mail in the past year."

That doesn’t sound hard (if you’re not in marketing), but those actually need different data attributes to work:

• Customers with home internet plans

  • Needs to sort by list of customers AND by what products they current use

• Customer households with three or more members

  • Needs to sort by list of customers (grouped by address) AND by count at address

• Customers without direct mail in the past year

  • Needs the channel for each campaign AND the customer list of who was targeted

Historically, fulfilling these requests involves stitching together disparate data sources and manual processes—an approach that is inefficient, prone to mistakes, and hard to scale.

This inefficiency translates into tangible costs: delayed campaign launches, inaccurate targeting, and suboptimal resource allocation.

Instead of coming up with a new approach to managing your data structure, steal!

Luckily, software developers already have a solution we can use!

Consider adopting Object-Oriented Programming (OOP) principles to streamline their marketing technology ecosystems.

When you formalize how your marketing is done, requests become a lot easier to manage (and track).

You can set up your department so that you never fail

Lets see how OOP can enhance operational efficiency, improve targeting speed, and deliver measurable business value.

The OOP Framework: A Technical Foundation for…Marketing?

Object-Oriented Programming organizes data and processes into reusable "objects," each encapsulating specific attributes (data points) and methods (functions).

Don’t worry if it sounds complicated—refer back to Pikachu above.

In a marketing context, this approach redefines key entities—such as customers, campaigns, and content—as modular components with clear properties and relationships.

And clear, cross-department definitions!

For example, think how you might describe any of these—those are the attributes:

• Customer:

  • Attributes might include "Products Used," "Channels Preferred," and "Engagement History."

  • e.g. “Oh, this customer prefers to be emailed”

• Campaign Object: Properties could be "Channel," "Target Audience," "Budget," and "Performance Metrics."

  • e.g. “This campaign had a ton of impressions over its course”

• Content Object: Nested within campaigns, tracking "Format," "Impressions," and "Conversion Rate."

  • e.g. “This piece of content’s format was a static image”

By establishing a clear framework and relationships, OOP enables marketing teams to manage data systematically, accurately estimate how hard something is, and scale operations seamlessly.

Strategic Benefits: Quantifying the Impact

Adopting OOP principles yields several advantages, each with measurable implications for enterprise marketing:

• Operational Efficiency: Standardized objects eliminate redundant processes, reducing campaign setup time.

  • If you know that you might reuse a ‘slice’, then you can save how it was captured and reuse quickly.

• Targeting Accuracy: Structured data reduces segmentation errors, reducing extraneous spend.

• Scalability: Reusable objects accelerate deployment of new campaigns, cutting development costs.

• Analytics Maturity: Clear object relationships enhance performance tracking. You might actually get useful metrics!

These benefits collectively contribute to a stronger ROI, as reduced overhead and improved outcomes compound over time.

Implementation Roadmap: A Strategic Approach

For enterprise marketers seeking to replicate developers’ success, the following steps provide a scalable blueprint:

1. Object Identification: Map key marketing entities and their attributes, aligning with business priorities

2. Relationship Mapping: Define how objects interact (e.g., Campaigns target Customers, Content drives engagement).

   

3. Technology Integration: Leverage platforms like Salesforce, HubSpot, or Snowflake to operationalize the framework.

4. Pilot Execution: Launch a proof-of-concept campaign to validate the approach and quantify benefits.

5. Track Requests Systematically: If you have a record of all requests, you can group them and start creating standard definitions for slices, reports, etc.

6. Enterprise Rollout: Scale the framework across teams, refining based on pilot learnings and recorded requests.

Addressing Challenges: Mitigating Risks

While the upside is compelling, implementation requires strategic foresight:

• Resource Commitment: Initial setup demands time and expertise. Mitigate this by prioritizing high-impact use cases to demonstrate early value.

• Change Resistance: Legacy teams may resist new processes.

  • New terminology is always very hard to get across, especially to stable, mature teams.

  • Overcome this through training and showcasing measurable improvements.

• Collaboration Needs: Success hinges on alignment between marketing, IT, and data functions. Establish clear governance to ensure cohesion.

Advanced Applications: Elevating Campaign Insights

Beyond foundational benefits, OOP enables advanced analytics by incorporating nuanced attributes into objects.

Drawing from John Dewey’s reflective thinking model, an example for a campaign could be:

• Promptness: Time to consume content (e.g., 30-second video vs. 5-minute article).

• Breadth: Reach across segments or channels.

• Depth: Persuasive impact of messaging.

• Persistence: Duration of campaign influence.

With these, you can still get some of marketing’s ‘fluffy’ attributes, but also can record them for trend analysis and further analytics.

Conclusion

Consider adopting OOP for marketing—don’t reinvest the wheel.

Marketing and technology leaders should evaluate OOP as a tool for their marketing operations strategy.

Start with a pilot initiative, measure its impact on efficiency and performance, and build a business case for broader adoption.

The future of marketing lies in structured, scalable solutions—OOP provides the blueprint from a process we already know works.

Even when everything else is perfectly aligned—strategy, creative, technology1, and execution—the approval process can grind everything to a halt, destroying your time to market.

Little’s Law—a theorem in queuing theory (stick with me)—says that you should improve the slowest step in a process before anything else.

You’ll get the greatest throughput gains from fixing the slowest part of your process first instead of doing a bunch of tiny fixes of lesser problems.

It’s a frustrating reality for many teams, but there’s a way to break free from this cycle: trust.

When used properly, trust becomes a powerful accelerator, transforming approvals from a bottleneck into a seamless, efficient process.

However, trust isn’t a one-size-fits-all solution.

To make it work, you need to trust the right people. Let’s explore the positives of leveraging trust in marketing ops approvals, the risks you accept, and how to navigate both effectively.

The Black Hole of Marketing Ops Approvals

Imagine your team has nailed the campaign plan, the assets are ready, and the tech is in place.

Yet, the launch date keeps being pushed because the approval process is stuck in a maze of sign-offs, revisions, and delays.

Who can escape this Kafkaesque maze?

This isn’t just inefficiency. Speed to market matters a lot to the bottom line.

Every day lost to approvals can mean missed opportunities, frustrated teams, and eroded momentum. You can’t run a tax return ad after tax deadlines!

The root cause? Often, it’s a lack of trust, leading to excessive checks and redundancies that slow everything down.

The solution lies in rethinking how trust can streamline this critical step.

The Positives of Trust in Approvals

Trust is a strategic advantage that can revolutionize your approval process.

When you trust the right people, you unlock a range of benefits that directly address the black-hole problem:

• Faster Approvals: Trust eliminates unnecessary layers of review. When you’re confident in the judgment and abilities of your team or stakeholders, you can skip the endless back-and-forth, getting campaigns to market quicker.

• Increased Efficiency: Trust enables delegation. Instead of micromanaging every decision, leaders (thoughtfully) pick team members to approve elements autonomously, keeping the process flowing without bottlenecks.

• Improved Collaboration: A trust-based environment fosters teamwork. When people feel trusted, they’re more likely to communicate openly, align on goals, and work together seamlessly.

• Innovation: Trust encourages calculated risk-taking. Team members who feel supported are more likely to propose bold ideas or solutions, potentially elevating the campaign’s impact.

These positives can turn approvals into a strength rather than a liability—but only if trust is placed wisely.

Picking the right people is the linchpin that makes this strategy work.

People are not fungible!!!

The Risks You Accept with Trust

Trust is a double-edged sword. While it can speed things up, it also introduces vulnerabilities that can derail your efforts if not managed carefully.

Here are the key risks you accept when relying on trust in marketing ops approvals:

• Misplaced Trust: Trusting the wrong person can lead to errors, oversights, or poor decisions. A single misstep—like approving a flawed asset—could delay the campaign or damage its quality.

• Lack of Accountability: Too much trust without oversight can create complacency. If individuals know they won’t be closely monitored, they might slack off or miss critical details.

• Over-Reliance: Leaning heavily on one trusted individual risks creating a single point of failure. If that person is unavailable or underperforms, the approval process could stall entirely.

• Bias and Favoritism: Trust can be skewed by personal relationships or unconscious biases, leading to unfair decisions or overlooking better-qualified contributors.

These risks don’t mean trust is a bad idea—they just mean it needs guardrails. The goal is to harness trust’s benefits while minimizing its downsides.

Navigating Trust in Marketing

To make trust a strategic advantage without falling into its pitfalls, you need a balanced approach. Here’s how to trust the right people and protect your process:

• Define Trust Criteria: Base trust on evidence, not assumptions. Look for a proven track record, relevant expertise, and consistent reliability in those you empower with approval authority.

• Start Small: Test trust gradually. Begin with low-stakes approvals and expand responsibilities as individuals demonstrate their capability. Speed to trust is not the same as speed to market.

• Maintain Oversight: Trust doesn’t mean abandoning checks entirely. Regular, lightweight reviews can ensure things stay on track without slowing down the process.

• Set Clear Expectations: Communicate roles, deadlines, and standards upfront. Clarity ensures everyone knows what’s at stake and what’s required.

• Track Progress: Use tools like project management platforms (e.g., Asana, Trello) to maintain visibility into approvals without micromanaging.

By pairing trust with these safeguards, you create a system that’s fast, flexible, and resilient.

Conclusion

Trust, when directed at the right people, can speed up your bottlenecks.

The positives—faster approvals, greater efficiency, better collaboration, and room for innovation—can propel your campaigns forward.

But the risks—misplaced trust, accountability gaps, over-reliance, and bias—remind us that trust should not be given freely.

By choosing trustworthy individuals and balancing freedom with smart oversight, you can outsmart the delays and unlock your team’s full potential.

Google could die.

It always seems impossible for the king to fall—but it’s happened many times in history.

Even in American commerce, GE, Xerox, Blackberry went from industry leaders to the graveyard.

Google has seen invincible for decades.

But that’s not the natural state of things—this concentrated media landscape, where a few companies control the ad world, isn’t normal—it’s a blip in history1.

Recent antitrust rulings are shaking things up, hinting at a future where the media might fracture again, like the days of local papers and mom-and-pop radio stations.

If you’re a marketer, the days of dumping cash into Google Ads and calling it a day are coming to a close.

Instead of bemoaning changes, it’s time to get back to the roots of marketing: creating real behavior change—sales, loyalty, buzz—for the least investment possible.

In this post, we’ll unpack:

• how we got here

• what’s changing

• how to adapt by returning to fundamentals of old-school marketing.

Let’s a-go

How We Got Here: How Media Giants Grew

Once upon a time, media was a messy, fractured landscape.

Every town had its own newspaper, radio station, or quirky TV channel, each with its own practices, audiences, and pricing.

Marketers had to understand audience segments—crafting ads for the local diner’s jukebox or the county fair’s banner.

It wasn’t easy, but it forced creativity and true understanding of the audience needs.

Then, the world shifted.

By 1983, 50 companies controlled 90% of U.S. media.

By 2011, it was down to 6.

The 1996 Telecommunications Act was a big spark, loosening rules and letting corporations merge and grow uncontrollably.

Then the internet exploded.

Google and Meta (f/k/a Facebook and Instagram) turned digital advertising into their personal fiefdoms.

These two alone suck up over 50% of U.S. digital ad spend.

For marketers, it’s been a godsend: massive reach, the vendor manages targeting, and you don’t need to get elbow deep to understand your audiences.

This consolidation’s killed diversity.

Algorithms amplify what’s already hot, drowning out smaller voices.

And this algorithmic decision-making has made marketers lazy.

Why sweat over a clever billboard—by *gasp* visiting somewhere in person and talking to people—when you can A/B test a Meta ad in 2 days from your laptop?

But leaning so hard on a few platforms has risks, especially when the legal ground’s starting to quake.

Recent Antitrust Developments: The Giants Start to Wobble

The tech giants’ grip is loosening, and regulators are swinging.

In August 2024, a U.S. court ruled Google broke antitrust laws by locking down online search with conspiratorial deals—like paying billions to be the default on phones and browsers.

Then, in April 2025, another judge slammed Google for running an illegal monopoly in online ad tech, putting its $31 billion ad machine in the crosshairs.

The DOJ even floated spinning off Chrome in November 2024, which could crack Google’s stranglehold on the ecosystem wide open.

Somehow all of this news doesn’t seem to affect their stock price very much. How? Who knows

Meta’s catching strays too.

In April 2025, the FTC launched a trial accusing Meta of building a social media empire by aggressively buying up threats like Instagram and WhatsApp.

If the FTC wins, Meta might have to carve off those platforms, shaking its $160 billion ad business to the core.

What’s this mean? The media world could splinter.

New players might flood in, and the ad market could start looking like the chaotic, fragmented days of old.

For marketers, it’s a warning: the platforms you’ve bet on might not be bulletproof much longer.

Time to rethink the playbook—and return to fundamentals!

Subscribe now

Returning to Marketing Fundamentals: The Hard-Boiled Playbook

The golden rule of marketing hasn’t changed: create the biggest behavior change—sales, loyalty, buzz—for the least investment.

It’s simple, but not easy, like telling someone to “eat less” to lose weight.

The trick is cutting through the noise and focusing on what drives results.

Here’s how to channel the no-nonsense wisdom of old-school marketers like David Ogilvy and Claude C. Hopkins to thrive in a shifting landscape:

Focus on Sales, Not Vanity Metrics

Forget likes, impressions, or “brand awareness” that doesn’t pay the bills. Your job is to move the needle—conversions, customer acquisition costs, lifetime value. Old-school marketers knew this.

Ogilvy said, “The more informative your advertising, the more persuasive it will be.”

Every campaign should answer: “Does this make people buy?”

How to make it happen:

• Measure what matters. Track conversion rates, not just clicks. If your ad spend isn’t turning into revenue, it’s a hobby, not a strategy.

• Test ruthlessly. Run small experiments to see what works—different headlines, offers, or channels. Claude C. Hopkins was a testing fiend, and it made him a legend.

• Set clear expectations. Before launching, line out what success would look like. If the math doesn’t add up (investment < return), rethink it. Include ad spend, employee time, and opportunity costs—no fudging.

Build a Unique Brand Identity

Chasing private capital or platform algorithms often leads to bland, safe branding.

But in a splintered media world, standing out is your superpower.

Companies like Darn Tough Vermont (socks that last forever) or King Arthur Baking (baking tutorials for all) don’t dilute their vibe for investors—they double down on what makes them unique.

Here’s how:

• Own your niche. Define your brand’s personality—goofy, rugged, helpful—and stick to it. Looking like everyone else means… that you look like everyone else.

• Skip the VC trap. Don’t water down your identity just to look “acquirable.” A strong brand drives sales, not just boardroom applause.

Spread Your Bets Across Channels

The days of “just Google it” are fading.

A fragmented future means you need to show up where your customers are—local radio, niche podcasts, or that weird new social app.

To allocate your budget wisely, consider the Kelly Criterion.2

It suggests betting a fraction of your resources proportional to your edge (expected return) divided by the odds (variance).

For marketers, this means putting more budget into channels with proven ROI but never going all-in on one, ensuring you balance risk and reward across experiments.

You’ll reach new audiences and dodge the risk of a single platform’s collapse.

Get started with:

• Local media. Sponsor a community event or run a radio spot. A coffee shop could own the morning commute with a memorable slogan for a local running group—that same slogan wouldn’t scale well, but the sales are still real locally

• Podcasts. Listeners trust hosts. Find a show your audience loves and sponsor it—bonus points for niche topics.

  • But don’t expect it to perform well on performance marketing!

• Emerging platforms. TikTok, Discord, or whatever’s next—jump in early. Small brands can use (temporary) inefficiencies in the market to make outsized gains.

  • But don’t plan on those lasting forever!

• Test and learn. Start small, track results, and scale what works. If a channel flops, don’t hesitate to cit it.

This isn’t about chasing trends—it’s about being pragmatic.

Every dollar you spend should justify itself with results.

If it doesn’t, stop.

That’s the fundamentals mindset: focus on what works, ditch what doesn’t, and keep your eyes on the prize—sales.

Conclusion.

The media landscape has the opportunity to fracture, and marketers need to be ready.

The concentrated world of Google and Meta was a historical fluke, and with antitrust hammers dropping, we’re likely headed back to a fragmented future.

That’s not a threat—it’s an opportunity.

By embracing the golden rule of marketing—biggest impact, least investment—you can outsmart the chaos.

Ditch the vanity metrics, build a brand that’s identifiably, and spread your bets across channels.

The old-school marketers, with their relentless focus on results, would nod in approval.

Here’s your to-do list:

1. Check your metrics. Are you tracking sales impact or just clicks? Fix it.

2. Define your brand. Write down what makes you unique—three words max.

3. Test a new channel. Pick one—local paper, podcast, TikTok, whatever—and run a small campaign.

4. Stay sharp. Watch your landscape, tweak your plan, and never stop testing.

5. Don’t be afraid.

   1. It is sometimes better to abandon one's self to destiny—Napoleon

The media’s shifting, and the marketers who know how to observe and react to changes will win the future.

Be one of them

Subscribe now

In the early 20th century, IBM mandated crisp white shirts and dark suits to mirror the formal attire of its corporate clients, showing they wanted to work with their clients.

But as client workplaces grew casual by the 1980s, IBM’s rigid dress code persisted, clashing with the times.

Originally meant to align with customers, the “blue shirt” became a symbol of adhering to dogma—until 1995, when IBM embraced more casual attire, rediscovering the true intent: adapt to the client, not the rule.

The output of ‘excellence’ in customer service changed with the times

A marketing manager and an engineer are locked in a heated debate.

The marketer is pushing for a quick product launch to capitalize on a trending market opportunity, while the engineer insists on conducting more rigorous testing.

Both are hard, good workers.

Both strive for "excellence."

Both have definitions that couldn’t be more different.

This scenario plays out in businesses worldwide, slowing down progress and building resentment.

Subscribe now

Misalignment on what constitutes "excellence" can be a significant roadblock, especially when it comes to developing digital products.

Different business domains—whether it’s aviation, tech startups, or marketing—have their own unique perspectives on what makes something "excellent." These differing views can create friction, leading to delays, rework, and frustration.

But why does this matter?

Businesses can’t afford to be bogged down by internal disagreements over quality—there are plenty of headwinds from outside your business. Understanding and reconciling these diverse perspectives is crucial for moving forward efficiently.

Just as data needs context to become valuable, so too does the concept of excellence require a shared understanding to drive business success

.

What Is Excellence? The Greek Concept of Arete

To grasp why different perspectives on excellence can be an issue, we first need to understand what excellence really means.1

The ancient Greeks had a term for it: arete.

In book 1 of Plato’s Republic, Socrates debates (mainly) Thrasymachus (a sophist) about the nature of justice.

Arete refers to excellence or virtue that’s specific to its context.

For example, the excellence of a knife lies in its sharpness, while the excellence of a rolling pin lies in its ability to roll dough.

You wouldn’t want a sharp rolling pin or a dull knife—each tool has its own measure of quality based on its purpose.

This sounds obvious. But it’s easy to get stuck in whatever their main domain you regard ‘excellence’ as and struggle to adapt.

In the business world, this concept translates directly. Each domain has its own version of arete:

• Aviation: Excellence means reliability and safety. A plane that doesn’t fly safely isn’t excellent, no matter how quickly it was built.

• Tech Startups: Here, development velocity and the ability to iterate quickly are prized. Getting a product to market ASAP, even if it’s not perfect, is be more valuable than waiting for flawless execution.

• Marketing: Speed to market and positive consumer sentiment are key. A campaign that resonates with customers quickly can outperform a meticulously planned but delayed initiative.

These differing definitions are like trying to agree on the best pizza—it depends on who’s you ask!

When teams from these diverse backgrounds collaborate, their differing views on ‘excellence’ can clash, leading to misunderstandings and slowed progress.

Why Misaligned Perspectives Cause Friction

When cross-functional teams come together, they often bring their domain-specific notions of excellence.

For instance:

• A marketer might push for developing a more customer-pleasing product experience (think of how nice it is to open a new iPhone box). Or try to play up the ‘purity’ of a supplement.

• A manufacturing manager accustomed to being judged on large-scale spec-alignment might resist changes that the marketer tries to make on a process that seem ‘useless’ because they don’t affect getting a product into spec.

These clashes aren’t just minor disagreements; they can lead to delays, rework, and a general sense of frustration as teams talk past each other. Each side believes they’re advocating for the best outcome, but without a shared understanding of what "quality" means, progress stalls.

This misalignment is similar to the challenge of interpreting data without context, as outlined in the DIKW model for turning data into knowledge.

Just as raw data needs context to become meaningful information, project specifications need a shared definition of excellence to guide effective collaboration.

Case Study: Legacy Enterprises and Digital Products

Consider a legacy enterprise like an aluminum refinery.

In this environment, excellence is traditionally defined by alignment to specifications, consistency, and safety.

Every machine, process, and output works together to produce the output to exact specifications.

Material World

The Weird and Wonderful Logic of Aluminium

One of the first challenges I faced when planning Material World was coming up with the list of substances I’d be covering. Needless to say, the initial list was not quite the same as the one I ended up with…

Read more

a year ago · 217 likes · 27 comments · Ed Conway

More pure aluminum is not ‘higher quality’ because its sold as a commodity. Aluminum in cleaner ingots is not ‘higher quality’ because asthetics don’t affect price.

This focus has made the company successful in its domain.

Imagine this same company attempting to add a product line.
They want to develop in-house marketing processes to sell ingots to individuals. [For some reason, pure metal cubes have become a trending desk toy]

The culture that excels in manufacturing—building to exact specs and maintaining strict processes—doesn’t translate well to another field.

Marketing to individuals requires focusing on things that don’t matter to commodity refineries:

• the unboxing and ordering experience

• how clean the product is (no residual refinery oil from the processing)

• how easy the product is to order (no phone-only order forms)

And the manufacturing team is driven crazy by the things marketers downplay:

• The exact price of the product (commodity products are made with tiny margins—consumer products are more fluid)

• The purity specs (consumers who want a desk toy probably don’t care)

As a result, a successful, hardworking employee that’s found success in manufacturing risks falling behind competitors who can adapt more quickly to the different landscape of consumer marketing.

This struggle highlights the importance of aligning perspectives on excellence, especially when venturing into new territories like digital products.

Why Digital Products Are Different—Information

So, why do digital products require a different approach? The key lies in how they derive value.

Note: In this context, I refer to digital products where the main value is the user’s actions after being informed by the tools.

A ‘digital product’ of robotic firmware that requires no human interaction should be designed to exact, preexisting specs—if a robot breaks, thats a physical product problem to resolve. You can’t just ‘reset’ a broken servo.

‘Information Products’ might be a more accurate term.

Unlike physical products, which prioritize consistency and perfection (e.g., every aluminum batch meeting exact specs), digital products derive value from actionable insights and adaptability.

Consider a predictive feature in a driving app: knowing and alerting you that there’s a 30% chance a car will merge into your lane a minute in advance is more valuable than being 99% certain just a second before it happens.

The former gives you time to react, while the latter is too late to be useful. This illustrates that in digital products, timely insights often outweigh perfection.

This difference is crucial because it demands a shift in mindset. Digital excellence isn’t about achieving flawless execution from the start; it’s about delivering value quickly and refining based on real-world feedback. This iterative process is foreign to many legacy enterprises, making the transition challenging.

Moreover, as discussed in our previous article on the DIKW model, digital products excel at turning data into wisdom rapidly. They require a mindset that embraces change and prioritizes user needs over rigid specifications.

How to Bridge the Gap

Acknowledging that different perspectives on excellence are inevitable is the first step.

Once you have the painful “What language are we speaking on this project” conversation, the rest becomes smooth sailing

The challenge is manageable with the right strategies:

• Shared Language: Develop a common definition of excellence for cross-functional projects. For example, agree that "excellence" means "delivering value to users quickly and reliably." This shared goal can align teams despite their differing backgrounds.

• Cross-Training: Expose teams to other domains. Let marketers shadow engineers to understand technical constraints, and vice versa. This builds empathy and a broader understanding of what excellence means in different contexts.

• Culture of Understanding: Encourage legacy teams to experiment with agile methods in low-stakes projects. Push agile development teams to walk the floor of a factory to understand that ‘move fast and break things’ is the wrong approach for physical products

Conclusion

Differing perspectives on excellence can significantly slow down businesses.

These misalignments create friction, leading to delays and missed opportunities.

By fostering a shared understanding of quality and embracing the unique demands of digital products, companies can overcome these challenges.

Leaders must take the initiative to create a common language for excellence and encourage cross-functional collaboration.

Remember, excellence isn’t universal—it’s context-specific, as the concept of arete teaches us. But by understanding and respecting each other’s contexts, teams can work together more effectively to drive innovation and growth.

So, the next time you find yourself in a debate over what makes a project "excellent," take a step back and consider the perspectives at play.