These cookies are required for the website to function properly. They ensure security, enable basic features like page navigation, and store user session data. You cannot disable these cookies.
- Retail Analytics
- Features
- Results and provenance
Results and provenance
Trace a measured value to its definition, source and operation history.
OverviewResult windows×
Retail Analytics · Results and provenance
New
Result windows
Metric results
| Id | Metric | Period start | Period end | Dimension key | Value | |
|---|---|---|---|---|---|---|
| IDD-565 | MD-7580 | 26 Oct | 02 Nov | DK-500 | 280 | |
| IDD-676 | MD-1040 | 08 Oct | 13 Oct | DK-989 | 170 | |
| IDD-433 | MD-7813 | 18 Oct | 26 Oct | DK-386 | 900 | |
| IDD-721 | MD-8012 | 27 Oct | 31 Oct | DK-251 | 390 | |
| IDD-620 | MD-1584 | 26 Oct | 29 Oct | DK-735 | 560 |
6 of 6 records
IDD-565Metric results · MR-25944
CalculatedMetricMD-7580
Period start26 Oct
Period end02 Nov
Dimension keyDK-500
Value280
What you can do
Read the result with its context
A metric result stores its period, dimension key, numerator, denominator, value, source row count, status and reason. It can link to a retail snapshot, metric definition and operation run. This lets the Retail Analyst answer a practical question: which source scope and approved definition explain the figure shown for a store period? A result marked suppressed or invalid should remain visible as an exception with its reason.
An operation run carries the operation key, frozen source revision, start and completion times, input hash, result reference and failure reason. Its numbered steps retain action names, input and output references, receipt state and any external receipt evidence. That detail is useful when a collection is interrupted. It distinguishes a completed result from a request whose outcome still needs reconciliation.
How the work is organized
- Result. The value and component counts stay with their period, dimension and status.
- Run. The operation key and source revision identify the exact execution context.
- Steps. Numbered actions and receipts explain what was attempted and what was confirmed.
What to agree during setup
Agree how the team chooses an eligible basket-fact scope for a result and which metric definition governs it. Confirm how a zero denominator is represented. A conversion rate should not appear as an ordinary calculated value when the denominator is zero. Agree how to reconcile a failed collection run, and keep any related result identity visible while that decision is made.
The reviewer should have a way to inspect the source row count and component values. A headline rate may move because the numerator changed, because the denominator changed or because the source window changed. Those are different issues and need different follow-up actions.
Check the handoff on an actual case
Open one reviewed store snapshot and the metric result for its period. Compare the linked metric definition with the recorded numerator, denominator and value. Check the source row count. If the result links to a run, inspect its frozen revision and numbered steps before relying on the recorded outcome. If the operation is marked for reconciliation, the Operations Lead investigates that record before the result is used.
Now compare a later period with a changed source window. The analyst should be able to explain the change from the stored period and source context, rather than editing the original result to make the trend look consistent. Keep exceptions visible in the Retail Overview follow-up queue.
What your team sees
- Result windows — Compare result values by metric, period and dimension.
- Result exceptions — Inspect suppressed or invalid outcomes and their reasons.
- Operation reconciliation — Find runs that failed or need a result check.
Follow the work through
What it keeps track of
Data model- Metric results
- Calculated metric results persisted once per approved definition and window.
- Metric definitions
- Approved metric definitions with grain, expressions and aggregation method.
- Retail snapshots
- Measured web properties with lifecycle status and observed outcome fields.
- Operation runs
- Workflow execution runs with frozen revision, input hash and result identity.
- Operation steps
- Numbered steps of an operation run with input/output and receipt state.
- Basket facts
- Consented observation-level session and conversion event facts.
Who uses it
Permissions- Retail Analyst
- , Independent Reviewer, Operations Lead
Set up Retail Analytics your way.
Start free, then use Proto to change fields, steps and approvals before your team goes live.