Designing Validation Experiments for New Ventures
How to design lean experiments that produce real evidence about your venture's riskiest assumptions, with hypotheses and thresholds set in advance.
A validation experiment is a deliberate test of a specific assumption, designed to produce evidence quickly and cheaply. The goal is not to prove you are right — it is to find out whether you are, before you spend heavily. Good experiments start with a clear hypothesis, define what result would count as success, and use the lightest method that gives a trustworthy signal.
- You have identified your riskiest assumptions and need to test them
- You want evidence before committing budget or building
- You need to avoid analysis paralysis and start learning
Start with a falsifiable hypothesis
State exactly what you believe and what would prove it false.
- 'We believe [assumption]'
- 'We will know it is true if [measurable signal]'
- Set the threshold before you run the test
Choose the lightest credible test
Match the method to the assumption — interviews, landing pages, concierge tests, prototypes, pre-sales.
- Prefer cheap, fast tests first
- Escalate fidelity only as confidence grows
- Avoid building the full product to test demand
Measure real behavior, not opinions
Behavioral evidence beats stated intent. What people do is more reliable than what they say.
- Track actions: sign-ups, payments, usage
- Discount vanity metrics
- Capture enough data to be conclusive
Decide: persevere, pivot, or stop
Every experiment should end in a decision, tied back to the threshold you set.
- Compare the result to your threshold
- Record the implication for the thesis
- Choose the next experiment or the next stage
A validation experiment on one line
Assumption
Clinics will pre-pay for automated pre-authorization.
Experiment
Offer a paid pilot to 10 target clinics via a simple landing page.
Threshold
At least 3 of 10 put down a deposit within two weeks.
- Running experiments without a pre-set success threshold
- Building the product to test whether people want it
- Interpreting ambiguous results as whatever you hoped
How Cogliva helps
Venture Lab helps you turn risky assumptions into structured experiments — hypothesis, method, threshold, and evidence — and records the decision each one drives.
Keep exploring
Related guides
Product & method
Frequently asked questions
What makes a good experiment?
A clear hypothesis, a pre-set threshold, the lightest credible method, and a decision at the end.
Should I test one assumption at a time?
Usually yes. Testing several at once makes it hard to know which one the result speaks to.
What is a concierge test?
Manually delivering the value by hand to a few customers before automating anything — a cheap way to test demand and delivery.
Learn fast, spend little
Design experiments that give you real evidence about your riskiest assumptions.