By Michael Kaminsky, Co-Founder of Recast
When marketing campaigns go wrong, the damage isn’t just monetary. Cringeworthy misfires can tank brand equity, dominate headlines, and become punchlines for years.
Many of us remember Pepsi’s protest-themed ad with Kendall Jenner, which was pulled within 24 hours and reportedly cost the company upwards of $5 million in production costs. There’s also Peloton’s 2019 holiday commercial, which caused so much backlash that the company’s market cap dropped by over a billion dollars in a matter of days.
Spectacular campaign failures can make marketers risk averse. But here’s what we often miss: most marketing failures don’t go viral – they’re small and disappear quietly. A campaign underperforms, a new channel disappoints, an A/B test falls flat. Teams bury these results and move on, particularly early in the year after a busy Q4.
That’s the real tragedy. Because while viral failures cost millions, forgotten failures arguably cost organizations more by robbing them of the insights that drive breakthrough growth. This makes Q1 (when organizational pressure is often lower and timelines are longer) the best time to build a marketing experimentation program that will help you win year-round.
Planning to Fail: A New Balance Case Study
Few global marketing powerhouses talk openly about failure the way New Balance does. In fact, Chris Davis, the company’s Brand President and CMO, has built his entire operating philosophy around it.
In a 2022 interview, Davis described New Balance’s budgeting framework as “50-30-20”:
- 50% goes to proven, evergreen campaigns
- 30% funds calculated marketing bets
- 20% is explicitly reserved for high-risk, high-upside efforts
This model incentivizes risk-taking and creates structural room for learning. Campaigns that begin in the experimental 20% bucket often evolve into calculated bets in next year’s 30% category, then graduate to the 50% tier of proven activities.
The results speak for themselves. New Balance posted $7.8 billion in revenue in 2024 – 20% growth and their fourth consecutive year at that pace. By embracing experiments that might fail, they’ve transformed from challenger brand to one of the industry’s fastest-growing players.
Misinterpreting Experiments Gone-Wrong
In my experience speaking with marketing teams across industries as diverse as fintech, B2B, and pharma, I’ve unfortunately seen different patterns play out.
New experiments are launched, and if they don’t exceed performance targets, their results go into a slide deck that never gets opened again. Consumer priorities shift, leadership gets impatient, and the team refocuses on the next “big swing” they want to take. Instead of using the original test failure to inform the next strategy and subsequent experiments, it gets erased entirely.
Of course, there are reasons for this. In addition to the financial and reputational costs of failure, teams are under pressure to produce quick wins. They may also lack a system for capturing and revisiting experiments so that they can inform future tests that ultimately go right. Most fundamentally, many organizations reward only success, creating cultures where admitting failure, even productive failure, becomes career-limiting rather than career-enhancing.
Whatever the cause, ignoring failed experiments often leads organizations to stop running them altogether. Worse, they may even repeat the same mistakes later on without realizing it.
Learning from Mistakes (By Writing Them Down)
To achieve results like New Balance, a simple but powerful organizational habit is critical: log every experiment you run, what you expected to happen, and what actually occurred.
I call this an “experiment ledger.” No matter where it’s housed, every good entry should include:
- The hypothesis behind the test
- The results, including point estimates and confidence intervals
- Any challenges or surprises during execution
- Explanations for why an experiment might have succeeded or failed
- Follow-on hypotheses for future testing
Here’s what a sample entry might look like:
Test: New TikTok prospecting campaign targeting women 35+
Hypothesis: Isolating delivery to a more targeted, affluent audience can drive customer acquisition under a $60 CPA vs. $75 CPA for other evergreen campaigns
Result: Estimate $90 CPA for the test audience with confidence interval of $78 – $102
Potential Learning: TikTok’s algorithmic targeting of broader audiences outperforms more filtered audiences created by channel managers
Next step: Continue using evergreen TikTok targeting and focus subsequent experiments on improving creative engagement and landing page conversion
Over time, this kind of ledger becomes a strategic asset. You’ll start to notice patterns across campaigns and channels, spot signals that a failed effort on one platform could succeed elsewhere, and ultimately stop making the same expensive mistakes twice.
But the power of an experiment ledger extends beyond data. It creates a cultural shift. When teams see that failed tests are documented and discussed rather than buried, it changes how they approach risk: running a bold test that doesn’t work becomes a contribution to shared knowledge rather than a career misstep. This visible commitment to learning promotes the kind of systematic experimentation that’s driving long-term growth for leaders like New Balance.
Failure as a Year-Round Growth Engine
Marketing experiments rarely produce clear wins, but that doesn’t preclude them from being valuable. A failed test eliminates a hypothesis, narrows your focus, and often sparks adjacent ideas worth exploring. The faster teams move through this cycle, the faster they learn.
Q1 is a great time to build this kind of experimentation program. For most businesses, the stakes are lower than they’ll be in Q3 or Q4 – timelines are longer, budgets haven’t been fully committed, and there’s organizational breathing room to try things that might not work. Use this window to start documenting tests, encourage reflection on what didn’t work, and train your organization to see failed experiments as contributions to shared knowledge rather than career setbacks.
The brands that learn fastest from their failures will be the ones leading the pack by year-end. Start building that capability now, while you have the space to get it right.
