Conditional-Layer Validation
Reduce a consumer product to ~4 things that must be true and validate them one at a time
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 4
- Confidence
- 95%
Model your product as a chain of conditional statements: if this is true, what must be true next for it to work? Each layer is a validation stage. The more layers, the riskier the product, so condense to about four must-be-true conditions. Then execute at 100% on the single stage you are validating and deliberately half-ass everything else, so you get a clean signal on one thing at a time.
Origin
Nikita Bier's zero-to-one development method, articulated from building TBH and Gas, where he validated the core flow, then intra-school spread, then school-hopping, then willingness to pay, each as a separate gated question.
Core principles
- 01Frame the product as layers of conditional 'if true, then next must be true' statements
- 02More conditional layers = higher product risk; condense to ~4 must-be-true conditions
- 03Execute at 100% only on the stage you're currently validating
- 04Deliberately half-ass everything else to avoid scope creep and confounded signals
- 05You are testing one thing at a time, so isolate it
How to run it
- 1
Write the chain of must-be-true conditions
List, in order, what must be true for the product to work: e.g. people use the core flow, it spreads within a peer group, it hops peer groups, people pay.
Pro tip Aim to condense to about four conditions; more layers means more compounding risk.
- 2
Pick the current stage and build it to 100%
For the condition you're validating now, make that specific experience flawless so a failure can't be blamed on execution.
Pro tip Bier made the polling experience perfect (questions, push notifications) before touching virality.
- 3
Half-ass everything downstream
Intentionally leave later-stage features rough so effort and attention stay on the one thing being tested.
Watch out Trying to solve everything at once creates scope creep and gives you no clean signal on any single layer.
- 4
Advance only on a clear yes
Move to the next conditional layer only when the current one is validated, then repeat the build-to-100% cycle.
In the wild
For Gas, Bier validated in sequence: would people send lots of messages, would it spread within a school, would it hop schools, would people pay. Each was a hard, separate problem in the new post-Twilio regulatory climate and took multiple relaunches to clear.
→ Gas reached 10M downloads and ~$11M in sales, growing bigger than TBH, before selling to Discord.
Common mistakes
Validating everything at once
Building the whole product before testing means you can't tell which layer failed, and scope creep kills momentum.
Too many conditional layers
Every additional must-be-true condition multiplies risk; a product needing many things to all be true is fragile.
Is it for you?
Best for
Zero-to-one consumer product teams sequencing what to build and test next
Not ideal for
Mature products optimizing a known, working funnel rather than discovering whether the concept works at all
From the transcript
“execute at 100% for the thing you're trying to validate at that specific stage of the product development cycle. And then the rest can kind…”
“if this is true then what next needs to be true for this thing to work out and these layers of conditional statements. And the…”
From the episode
How to consistently go viral: Nikita Bier’s playbook for winning at consumer apps (co-founder of TBH, Gas, advisor, investor)