Skip to content
Metricsjar

Article · Updated July 2026

What a Hosted Analytics Home Should Make Explicit

A recurring dashboard should reduce the work of checking your numbers without hiding how those numbers arrived. That requires more than putting charts on a hosted page. The product needs a clear boundary around access, freshness, and ownership.

Metricsjar starts with a private dashboard you create in the product. You then connect only the sources that dashboard needs and map each connection to the right project, dataset, app, or market.

Connection access stays explicit

Data health shows each supported connection, the access it needs, and the dashboard areas it fills.

Google uses an OAuth grant scoped to read-only BigQuery access. RevenueCat, App Store Connect, and DataForSEO accept provider credentials that you create and can revoke. Metricsjar encrypts those hosted credentials and stores the project, dataset, app, and market mappings separately for each dashboard.

Read how the own-keys model works for the boundary in detail.

A hosted result still needs verification

An API can ignore a filter and return 200. A date window can widen silently. “Revenue” can mean gross customer spend in one endpoint and proceeds in another. The durable pattern is:

  1. Query close to the source.
  2. Label the calculation precisely.
  3. Keep the source window and filters with the result.
  4. Re-derive surprising numbers before changing product or strategy.

The value of a hosted dashboard is not that every number becomes automatically correct. It is that the source, calculation, window, and last refresh remain visible enough to question.

Frequently asked questions

Does setup require a command-line tool?

No. Create the dashboard in Metricsjar, then connect and configure each source from Data health.

Does Metricsjar need my Google OAuth token?

Only if you connect Google. Metricsjar then requests read-only BigQuery access through Google OAuth.

Why keep a dashboard for a small set of metrics?

Because repeated questions are expensive in attention. A dashboard is best for stable numbers you check every morning, especially when they come from several source systems.

Keep reading