Archiving Results
Saving and documenting outcomes of completed experiments.
Why Archive is Needed
Archive is a repository of hypotheses with saved experiment results. It helps:
- Accumulate knowledge about what works in the product
- Avoid repeating failed tests
- Analyze trends and patterns in experiments
- Train new team members on real cases
Each hypothesis in the archive contains complete history: what was tested, what was the effect, what conclusions were made.
How Hypotheses Get into Archive
A hypothesis automatically moves to archive after saving experiment results.
Process:
- Complete experiment in product and upload data to Data Preparation section
- Analyze results: review tables and statistics
- Click "Save Results" button on analysis page
- In modal window select hypothesis or create new one
- Describe detailed experiment conclusions so non-statisticians can understand (can be added later when editing)
- Results will be automatically saved and hypothesis will move to Experiments Archive
Saving Results
After analyzing data on Data Preparation page, click "Save Results". A modal window will open with two tabs:
"Save to Hypothesis" Tab
Attach results to existing hypothesis:
- Start typing hypothesis name or external_id in search field
- Select hypothesis from dropdown list
- Click "Save"
Analysis results (tables, statistics) will automatically attach to hypothesis and it will move to archive.
"Create Experiment" Tab
Create new hypothesis directly from analysis results:
- Specify external_id (optional) — identifier from your task tracking system
- Specify hypothesis name
- Describe hypothesis — what was tested and why
- Select key metric from analyzed ones
- Click "Create and Save"
New hypothesis will be created with saved results and immediately go to archive.
Saving Results by Segments
If you analyze multiple segments separately (for example, separately for iOS, Android, and Web), save results of each segment to the same hypothesis. In the hypothesis card in archive, toggles will appear between saved segments, and you'll be able to view results for each separately.
After saving results, hypothesis will disappear from Hypotheses section and appear in Experiments Archive.
Viewing Archive
Open Experiments Archive section via sidebar menu. Here are stored all hypotheses with saved experiment results.
For each hypothesis displayed:
- Name and external_id (if specified)
- Hypothesis description
- Key metric
- Experiment start and end dates
- Last modification date
Click on card to open detailed information:
- Full hypothesis description
- Saved analysis results (tables, statistics)
Archive Search
Use search bar to quickly find needed experiments:
- Search by hypothesis name
- Search by external_id
- Search by description
Results update in real-time as you type.
Editing and Deleting
Hypotheses in archive can be edited (name, description) and deleted. When editing:
- Saved analysis results remain unchanged
- Can update hypothesis description or external_id
Important: Experiment start and end dates cannot be edited. When saving results, they are automatically set from actual experiment dates, overwriting dates from planning (if they were different).
Deletion: Organization administrators can delete hypotheses from archive. When deleting, all saved analysis results are lost.
Analyzing Experiment History
Archive is not just storage, but a tool for team learning.
Useful practices:
- Retrospectives: review archive quarterly and look for patterns — what types of changes work better?
- Onboarding: show new team members archive with comments so they quickly understand what works in the product
- Prioritization: before launching new hypothesis, check archive — perhaps similar experiment was already conducted