Retest and Recombine Loop
Diagnose failed tests before discarding channels or messages
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 96%
The Retest and Recombine Loop treats marketing experiments as connected Lego bricks rather than indivisible successes or failures. When a campaign underperforms, the team decomposes it into message, audience, channel, offer, conversion path, and the product context that existed at the time. A channel may appear ineffective because the message was wrong; a strong message may fail because the channel or downstream conversion step was missing. Because an early company's value proposition and experience can change within weeks, an old result may also become obsolete. The team preserves learning, identifies the weakest component, creates a revised combination, and retests it. This keeps experimentation alive while avoiding the opposite mistake of blindly repeating an unchanged tactic.
Origin
Extracted from Sumaiya Balbale's advice on developing repeatable Series A and B marketing motions.
Core principles
- 01A past failure is conditional rather than permanent
- 02Rapid business evolution can change experiment outcomes
- 03Channels and messages are separate test components
- 04Marketing motions emerge by recombining connected parts
How to run it
- 1
Reconstruct the test
Document what was tested, for whom, through which channel, under which product and market conditions.
Pro tip Include the date and product state so later teams can judge relevance.
Watch out A label such as paid social failed contains too little information.
- 2
Split the components
Separate audience, message, creative, channel, offer, landing experience, and conversion process.
Pro tip Draw the sequence as a chain of handoffs.
Watch out Do not attribute the total result automatically to the most visible component.
- 3
Find the broken brick
Use quantitative and qualitative evidence to identify where attention, comprehension, intent, or conversion deteriorated.
Pro tip Combine funnel data with customer feedback.
Watch out Low final conversion does not prove the initial channel was bad.
- 4
Update for current reality
Account for changes in the product, value proposition, customer maturity, and company credibility since the original test.
Pro tip Retest when a material contextual variable has changed.
Watch out Do not revive a test merely because time has passed.
- 5
Recombine and rerun
Retain promising elements, replace the weak component, and execute a focused new experiment.
Pro tip Change as few variables as practical when diagnosing causality.
Watch out Changing everything prevents useful comparison.
- 6
Keep or retire with evidence
Scale a combination that works or retire it when repeated, well-constructed tests fail under relevant conditions.
Pro tip Record the reason for the final decision.
Watch out Never convert one ambiguous failure into permanent doctrine.
In the wild
A campaign generates strong engagement but few purchases. Instead of declaring the message ineffective, the team discovers that the landing experience omits the promised use case. It preserves the message, repairs the conversion path, and retests the combination.
→ The team learns which component failed and avoids discarding a useful message.
Common mistakes
Saying we tried that
A past test performed with a different product, message, or customer context may not predict current results.
Blaming the channel
The channel may have delivered the right audience while messaging or conversion mechanics caused the failure.
Changing every variable
A wholesale relaunch can improve results without revealing which mechanism was responsible.
Is it for you?
Best for
Rapidly evolving companies building their first repeatable sales and marketing motions.
Not ideal for
Teams lacking enough measurement or customer evidence to distinguish one test component from another.
From the transcript
“don't be quick to throw those things out.”
“So a channel may not work because the messaging was off, right?”
“it starts actually becoming this set of Lego bricks that you're playing around with.”
From the episode
The Ideal Marketing Strategy For Each Stage of Start-Up Growth with Sequoia's CMO Sumaiya Balbale
Sumaiya Balbale