Minimal Viable Test
Spend the fewest resources needed to gather evidence for a larger bet
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 99%
A Minimal Viable Test is not merely a cosmetic optimization. It is a deliberately small experiment designed to produce the maximum useful evidence for a consequential future bet while consuming the fewest resources possible. The team begins with the expensive decision it may make, identifies the assumptions on which that decision depends, and recreates the riskiest part through a limited product, campaign, partnership, or audience. Metrics such as signup conversion, usage, retention, and lifetime value reveal whether conviction should increase. The method protects teams from falling in love with an idea and scaling it before customer behavior validates the underlying thesis.
Origin
Extracted from Marketing Against The Grain during a discussion of how Quibi and CNN Plus might have reduced launch risk before committing hundreds of millions or billions of dollars.
Core principles
- 01A small test exists to inform a larger decision
- 02Maximize decision-relevant learning while minimizing resource expenditure
- 03Test the riskiest assumptions before scaling the full concept
- 04Measure conversion, retention, and value rather than superficial activity
- 05Evidence should strengthen or weaken conviction in the larger bet
How to run it
- 1
Define the Larger Bet
Write down the costly product, campaign, or expansion decision the team is considering.
Pro tip Include the capital, time, and reputational exposure at stake.
Watch out Do not start with a convenient test that is disconnected from a real decision.
- 2
Expose the Assumptions
List the beliefs that must hold for the larger bet to succeed, including audience behavior, format, channel, conversion, and retention.
Pro tip Ask which false assumption would invalidate the entire plan.
Watch out Teams often protect favored ideas by testing only low-risk details.
- 3
Build the Smallest Credible Test
Create a limited experience that gives customers a genuine opportunity to demonstrate the target behavior.
Pro tip Use partnerships or existing distribution instead of building full infrastructure.
Watch out A test that omits the core value proposition produces misleading evidence.
- 4
Measure Decision Metrics
Collect the conversion, consumption, retention, and value data required to evaluate the larger bet.
Pro tip Set thresholds for scale, revision, and cancellation before results arrive.
Watch out Do not substitute attention or compliments for behavioral evidence.
- 5
Update Conviction
Use the results to increase investment, modify the thesis, run another bounded test, or stop.
Pro tip Document what changed in the team's beliefs after the experiment.
Watch out Treating every result as validation defeats the purpose of testing.
In the wild
Instead of raising $1.7 billion and launching a mobile-only service, the team could have released a few short-form shows through promotional partners. It could then have measured signups, consumption, conversion rates, and lifetime value before funding the complete platform.
→ The test could have exposed weak mobile-only demand before the largest financial commitments were made.
A consultancy considering a software platform first delivers the proposed outcome manually to ten representative customers. It measures purchase conversion, repeat use, and willingness to pay before building automation.
→ The team obtains evidence about demand and economics without funding the entire platform.
Common mistakes
Testing Cosmetic Details
Changing a button color may optimize an existing funnel but usually does not validate the assumptions behind a major new bet.
Building Too Much
A test loses its risk-reduction advantage when the team constructs most of the final product before gathering evidence.
Falling in Love with the Idea
Emotional commitment can cause teams to reinterpret weak evidence as permission to proceed.
Is it for you?
Best for
It is best for expensive, uncertain launches where a smaller experiment can reproduce the most important customer decision or behavior.
Not ideal for
It is not ideal when the small test cannot represent the full product's essential value or when failure would create disproportionate safety or reputational harm.
From the transcript
“Minimal viable tests are the way that you iterate towards something large in terms of a risk, a bet that you want to make.”
“And the purpose of a minimal viable test is to try to get the maximum amount of data and research by expanding the least amount…”
“Minimal viable tests are a mechanism to get you data and research to make bigger bets with more conviction.”
From the episode
Argument Alert!: New Game Double Down, Table Stakes and Kill