Analytics routines
Reports that arrive, not dashboards that wait.
Analytics routines query the product and market data you already collect — PostHog insights, events, and funnels, Apify datasets, or the internal tools you run in Retool — on a schedule, compare the period against the last one, and write down what changed and what to ask next. The result is a brief someone can act on, with the numbers cited so a skeptical reader can check them.
Put a routine to workRoutines in analytics
- Weekly or monthly
Weekly or monthly analytics report
Replace a dashboard nobody reads with a recurring brief that says what changed and what to ask next.
- Monthly
Funnel conversion report
Find where a funnel actually leaks and whether it changed, instead of debating screenshots.
- Weekly
Retool operations audit
Know which internal tools touch production data and which workflows quietly failed, without opening every app.
- Weekly
Apify market data review
Turn scraped pricing, listing, and review data into a weekly change report without starting or paying for new runs.
Also touches analytics
- Quarterly revenue report
- Weekly error trend review
- Weekly churn and payment-risk review
- Lifecycle performance review
- Competitor research brief
- Weekly search visibility review
- Weekly CRM pipeline review
- Supabase database health review
- Slack customer signal brief
- HubSpot pipeline hygiene review
- Cal.com booking funnel review
- Attio CRM hygiene review
- Twenty CRM pipeline review
- Salesforce pipeline review
- Xero revenue and cash review
- Pipedrive pipeline hygiene review
Works with
PostHog
Query product insights, events, and funnels for a recurring brief built around decisions, not dashboards.
Retool
Audit internal apps, queries, and workflow runs for production risk and silent failures.
Apify
Compare the latest dataset runs against the previous ones and report what changed.
Common questions
- What access do analytics routines need?
- Read access to the analytics tools you approve, such as PostHog insights, events, and funnels. Every tool starts disabled; you enable exactly the capabilities the report needs.
- Will it invent explanations for a metric change?
- The routine instructions separate observed changes from hypotheses. A run reports what moved and how much, and phrases causes as questions to investigate rather than conclusions.
- How do results arrive?
- Every run produces an inspectable trace and a result you can route to email, push, Slack, Discord, or a webhook when the run finishes.
Give analytics work a schedule.
Adapt a routine, approve its exact permissions, prove one run, then let the schedule own it.