Features
Everything Metrica does
What does Metrica do? It reads your Stripe, Google Analytics 4, Search Console, Google Sheets or CSV data read-only, computes every change against last week, the 4-week average and your targets in code, and writes a weekly briefing in which every statement links to its metric, period and query. Everything below works today.
Sources
All read-only. Remote sources are re-read every 6 hours, and before every scheduled briefing.
- Stripe. A restricted key with read access (stored encrypted; full secret keys are refused) or Stripe Connect with the read-only permission. MRR, revenue collected, new paying customers, churned customers and failed payments, by day.
- Google Analytics 4. Read-only Google sign-in. Sessions and key events (both split by default channel group), active users, new users and engagement rate.
- Google Search Console. Clicks, impressions, click-through rate and average position from Google web search.
- Google Sheets. Any tab with a date column: each metric is a column (or a count of rows) by date, with an aggregation, an optional filter and an optional split.
- CSV upload. Up to 5 MB and 50,000 rows. Metrica suggests a metric per number column; upload a new version any time and everything is recomputed.
Metrics and targets
- Presets and your own. Stripe, GA4 and Search Console come with their metrics; CSV and sheet metrics are defined by columns: total, count, average, last value (for balances), highest or lowest day.
- Direction. Say whether higher or lower is better (failed payments, average position), so changes are marked better or worse correctly.
- Weekly targets. Set a target per metric; a miss of more than 10% is called out.
- Segments. Split a metric by a channel, plan or any column, so a briefing can say which part of a change moved.
Analysis, in code
- Week over week. Monday to Sunday in your time zone, against the week before.
- Against the 4-week average. So one odd week before doesn't make a normal week look dramatic.
- Unusual weeks. 2.5 standard deviations from the 8 weeks before.
- Thresholds you choose. Default: 15% week over week or 20% against the average; counts must also move by at least 3.
- Incomplete data is held back. If a source isn't complete for the week (Search Console publishes late), the metric is left out of comparisons and listed under attention instead of raising a false alarm.
The briefing
- Three sections. What changed, why it likely changed (only from the data), and what needs attention.
- Evidence on every statement. Metric, period, comparison, the computation in words, source, data freshness and the query behind it.
- AI wording with a number check. Written by Anthropic's Claude from the computed figures only; any sentence with a number not in its evidence is dropped. Turn AI wording off for template wording.
- Yours to change. Edit any statement (edits with a new number are flagged), reject it, regenerate from the latest data, mark it reviewed, export as Markdown, CSV or JSON.
Delivery
- Email on your schedule. Monday at 8 by default, in your workspace's time zone, to the team and up to 10 extra addresses; each email goes once per week.
- Slack. Paste a Slack incoming webhook and the briefing is posted to that channel.
- Share links. A read-only link to one briefing, without access to your workspace, revocable any time.
- Dashboard. Every metric's last week, 12-week charts with targets, and the source and freshness of every figure.
Team and data
- Roles. Viewers read; editors generate, edit and review briefings and set targets; admins connect sources and set schedules; the owner handles billing.
- Brands (Pro). Up to 25 brands, each with its own sources, metrics, briefings and schedules: an agency's clients, or several shops.
- Export and delete. Download the whole workspace as JSON; delete a source, a brand or the workspace; history is kept 1, 2 or 5 years, your choice.