Bayesian tools for product and growth decisions.
Decide whether a variant won, what an experiment is worth running, and whether churn is outrunning acquisition — with uncertainty you can show a stakeholder.
Product decisions rest on small numbers: a few hundred visitors per variant, one quarter of churn, a backlog of guesses. These tools put a range on each number and say what the range is conditional on. A stakeholder then sees how sure the answer is, not only which way it points. Everything runs in your browser, so you can paste real counts.
What do you need to find out?
Did variant B really beat A, and by how much?
The probability that B beats A and a credible interval for the lift, conditional on your prior. Bayesian A/B Test Calculator · How do you use prior information in an A/B test?
Is this experiment worth its cost before we run it?
The expected value of what the experiment would teach you, set against what it costs to run. Bayesian Innovation Options Simulator · How much is an experiment worth before you run it?
Where does steady acquisition against steady churn level off?
The customer count where sign-ups and losses balance, and how wide that ceiling is. Customer Churn Simulator · Why does customer growth level off when sign-ups stay steady?
What is a customer worth against what winning them cost?
Customer value as a range, read against the cost to acquire one. Customer Value Calculator · What can 1, 10, 100 and 1,000 customers tell you?
Earlier tests point one way. How much should they move this one?
A result that carries earlier tests in as a prior, with a check on how much the prior decided. A/B Test Calculator with Historical Data
What goes in the sprint: features, experiments or tech debt?
A sprint mix ranked by expected value, with the uncertainty in each estimate shown. Sprint Mix
Worked examples (synthetic data)
- Landing-page conversion, A against B
- A leaky consumer app over 18 months
- A subscription app's customer value
- A market test before a launch
Instruments
- Bayesian A/B Test Calculator
- A/B Test Calculator with Historical Data
- Customer Churn Simulator
- Customer Value Calculator
- Ceiling Calculator
- Sample Size Checker
- Bayesian Innovation Options Simulator
- Usage Count Estimator
- Call Coach
- Learning Value
- Learning Backlog
- Rollout Readout
- Sprint Mix
- WSJF Calculator
- Tech Debt Payback
- PERT Three-Point Estimate
- Sprint Forecast
- Beta-Binomial Calculator
- Expert Prior Builder
- Rating Estimator
- Time-to-Event Explorer
- Weighted Decision Matrix
- DAG Sketchpad
- Posting Time Estimator
- Conjunction Calculator
Articles
- How do you use prior information in an A/B test?
- How do you reuse earlier A/B tests as a prior?
- How much is an experiment worth before you run it?
- Why does customer growth level off when sign-ups stay steady?
- What can 1, 10, 100 and 1,000 customers tell you?
- How do you turn expert opinions into a prior?
- How do you estimate component life when some units are still running?