Make old tests reviewable
Import records from CSV, JSON, pasted spreadsheet rows, or supported platform exports, then keep the control, variant, metric, outcome, and caveats in one format.
Experiment memory
An A/B test repository that keeps the evidence with the result
GrowthLayer is a private library for the research, plan, result, and learning behind your experiments. It complements a testing platform; it does not split traffic or run variants for you.
What belongs in an experiment library?
A useful record makes the decision auditable later. It connects the customer problem and research signal to the proposed mechanism, the exact change, the outcome data, and the conclusion the data can reasonably support.
Review relevant history before planning
Use the library, status filters, and pipeline sorting so a new idea begins with recorded history rather than a blank brief.
Show the limits with the result
Record missing sample size, duration, diagnostic checks, or conflicting changes. The evidence tier keeps a reported outcome from becoming a stronger claim than it supports.
Use the library to improve the next decision
- 1. Import the record. Capture what was changed and what was measured.
- 2. Mark what is missing. Do not infer a causal result when the run details are unavailable.
- 3. Review before planning. Use the relevant context, not a headline lift in isolation.
- 4. Add the new learning. Keep wins, losses, and inconclusive tests available to the next owner.
A result is not automatically reusable proof
A historical outcome can be useful context even when it lacks allocation, sample size, duration, or diagnostic data. GrowthLayer keeps those gaps visible so the scope of the next decision stays proportionate to the evidence.
Use the experiment planning framework