Don't Let Your Marketing Failures Stay Failures
Don't Let Your Marketing Failures Stay Failures
How a Can-Do attitude and some pre-planning can make it impossible to lose
The door thuds closed behind you. You wipe your clammy palms against your pant leg. The air conditioner hum breaks the heavy silence.
Separated from you by a mahogany desk, your boss looks up from a BI dashboard. Before he closes the computer, you get a glimpse of a dark red screen.
He glowers. You wish the desk were wider.
“Walk me through the results from last quarter’s campaigns” he says.
You sit up straighter—you’re not worried anymore.
“We ran 10 campaigns and learned a TON about our how our customers behave” you say
He raises an eyebrow
“Only 2 of the campaigns raised revenue” he says
“Yes, and1 we preregistered our hypotheses2. We built a better understanding of the levers we can pull to change our business.
We’re excited to run next quarters’ campaigns now” you say
Before he can respond, you whip open your laptop and show a spreadsheet of all the experiments that your team has tried—with the expected results vs real results.
“Here’s the data,” you continue.
“For each campaign, we tracked engagement, conversions, and customer feedback.
We identified two key email communication patterns and refined our timing strategies.
The two revenue-successful campaigns have shown us what resonates with our audience.”
Your boss leans forward, squinting at the numbers.
“And the others?”
“Provided crucial learning experiences.
For example, Campaign 3 taught us that our assumptions about generational preferences were off. Campaign 7 showed us that timing plays a bigger role than we thought.”
He nods slowly, still scanning the data.
“So, you’re saying we’re better prepared now?” he says
“Absolutely. We fine-tuned our approach and are already seeing better preliminary results from the new strategies we’re testing this quarter.
Just because the revenue didn’t immediately change doesn’t mean we didn’t succeed”
He leans back, a hint of a smile playing at the corner of his mouth.
“Alright, I like that strategy. Let’s see what you can do this quarter.”
You feel a surge of confidence. This is just the beginning.
Imagine having to win an office staredown with Don Draper.
The only way to win is to cultivate a ‘no-lose’ attitude.
That doesn’t mean developing an overwrought façade of arrogance, but rather treating your customer data as results from a real-life experiment.3
Luckily, most bosses are less intimidating than John Hamm, but the principle is the same.
What does this look like in real life?
In the simplest case, imagine that your marketing team broke Ohio into two population-equal sections.
They ran a controlled trial where region 1 had marketing interactions as normal.
They pre-registered the hypothesis that direct mail doesn’t account for many sales because very few customers used the on-card promo code.
For region 2, they decided to reduce direct mail marketing for top-of-funnel audiences to save money.
Region 1 had no meaningful change vs the previous quarter.
Region 2’s sales collapsed 60%.
Bad news, right?
Not at all!
You now know:
Customers are reading your direct mail
Customers aren’t correctly attributing mail as a sales channel4
Customers use direct mail to form a decision to buy your product
This is great! You’ve learned something about your customer and have opened the door to a lot of important follow-up questions:
Is direct mail the primary contributor for a sale or is it the repeated exposure across channels5
What other channels could be intertwined with direct mail to create this effect?
Would adding more spend into direct mail increase sales proportionally?
Is our media mix appropriately allocated with this new information?
Could we more closely match the copy, colors, and/or style of the mailers? They clearly resonate
Can we do customer interviews to figure out what groups we may need to re-segment?
With a little pre-planning, you can turn a ‘failure’ into a huge success.
Scientists don’t ‘lose’ when their hypothesis is wrong—you can be the same way.
If you plan out your ‘experiments’ thoughtfully, record your data, and analyze effectively, no experiment is a failure.
Don’t let your boss mis-interpret your hard work. Everything you learn about your customers is a step in the right direction.
Always “yes, and…” instead of “no, but…” if you can
Yes, this is actually the plural of ‘hypothesis’
This does NOT mean that you should call your customers ‘test subjects’
Depending on how you build your attribution model, it would be worth updating some parameters based on this result
