Documentation · Updated July 2026
Read Your Dashboard
A Metricsjar dashboard is a compact view of the rows your agent last published. It is not a live query into your source accounts. Check the freshness line before making a decision from a number.
Acquisition
The Acquisition tab combines search and site traffic:
- Target keywords tracks the phrases you have chosen to watch.
- Movers shows queries whose position changed enough to warrant a look.
- Pages groups Search Console performance around the page you can edit.
- Backlinks comes from DataForSEO when that source is connected.
- Traffic uses Search Console clicks and impressions, plus Google Analytics pageviews when available.
Search Console average position is an aggregate, not a single fixed rank. DataForSEO can independently check a target keyword from a specific location and device.
Product
The Product tab appears when app or product sources are configured. It can include downloads, active installs, app events, subscription revenue, trials, ratings, reviews, and retention cohorts.
Money figures need precise labels:
| Metric | Meaning |
|---|---|
| Revenue | Customer spend in the selected window |
| Proceeds | The amount after store commission and applicable adjustments |
| One-off | Non-subscription purchases included within revenue |
Do not subtract figures from different windows or sources and call the result a store fee. Metricsjar keeps these figures on one reconciled window.
Empty, zero, and pending are different
- Empty means the source or metric is not available.
- Zero means the source returned a measured value of zero.
- Pending means the upstream report has not settled yet.
- Stale means the last successful refresh is older than expected.
App Store Connect reports commonly lag by one or two days. Google Analytics privacy thresholds can also hide small cohorts. A blank recent day should not be silently converted into zero.
When a number looks wrong
Ask the agent to trace the metric back to its source:
Explain how this Metricsjar value was computed, including the source query,
date window, filters, and last refresh. Re-derive it from the source before
proposing a code change.
That distinction matters because many analytics APIs return HTTP 200 with a plausible value even when they ignored a filter. A confident number is not evidence that the query meant what you thought it meant.