Features built around one job: keep learning from every experiment.
GrowthLayer is not where your team runs A/B tests. It is the memory system that imports old results, validates the evidence, and keeps the learning in a consistent record for the next roadmap decision.
The memory system has four layers.
Import Layer
Get old experiments into the system fast
Bring historical tests in from spreadsheet rows, CSV, JSON, or supported platform exports without rebuilding every record by hand.
Explore import options- Paste CSV, spreadsheet rows, or a JSON array
- Upload Optimizely, VWO, AB Tasty, PostHog, Statsig, or GrowthBook exports
- Preview parsed experiments or map generic columns
- Keep import and evidence warnings available for review
Validation Layer
Make imported history trustworthy
A memory system only works if the data is credible. GrowthLayer checks the math and flags risky or incomplete conclusions before they become team knowledge.
See statistical tools- Lift, p-value, Bayesian probability, and confidence checks
- Sample ratio mismatch detection
- Warnings for missing raw data, short duration, or unsupported claims
- Calculator layer for planning and interpreting experiments
Memory Layer
Keep every test reviewable later
Keep the fields analysts need during planning together: what was tested, where, for which metric, with which outcome, and why.
View test library- Structured records for hypothesis, variants, results, and learnings
- Status filtering and priority sorting in the pipeline
- Private workspace with eligible team scope
- Manual record review through the library and pipeline
Insight Layer
Summarize recorded program outcomes
Use structured records to review coverage, recorded winners and losers, program velocity, and supplied impact fields without filling in missing evidence.
Explore insights- Coverage views across recorded site areas
- Outcome and evidence summaries
- Quarterly test volume and velocity reporting
- Documented learnings from completed experiments
Each capability covers a different moment in the experiment lifecycle.
Import gets history into the system. Structured records and status views make it reviewable. Coverage and reports summarize what has been recorded.
Organization & Search
Use structured fields, status filters, and sorting to review the test history.
Learn moreIteration Recommendations
Use past evidence to decide which winners deserve follow-up and which mistakes to avoid.
Learn moreReporting
Turn validated results into clear readouts for Slack, email, planning docs, and executives.
Learn moreIntegrations
Bring in historical data from CSV, JSON, pasted spreadsheet rows, and supported platform exports.
Learn moreStart with shipped workflows. Evaluate these additions separately.
The items below are future expansion paths, not capabilities represented as available in the current signed-in product.
Bulk history migration
Extend the current batch import and QA workflow for larger, assisted history migrations.
Duplicate and similar-test detection
Add a visible planning workflow for comparing new ideas with past hypotheses and outcomes.
Native source connectors
Move beyond the shipped export adapters to authenticated, ongoing source connections.
Program learning dashboards
Expand current coverage and quarterly reporting into more configurable leadership views.
Start by importing what your team already knows.
One old experiment is enough to show the product story. Ten to twenty old experiments is where the memory layer starts creating strategic value.