User Behavior Analytics
helio / user-insightsConnect your product's usage data. It turns a week of events into a plain-English report on active scenarios, regions, retention, and anything that suddenly changed.
Read last week's usage events from PostHog and send me a weekly report: top scenarios, region breakdown, retention cohorts, and anything that changed. Gate-check every metric against an agreed definition before it reaches me. Ask me what you need before you start.
Swap in your own event source and the scenarios that matter. It reports on schedule and stays silent otherwise.
What is User Behavior Analytics?
A teammate that reads your product's usage events and delivers a plain-English weekly report on active scenarios, regions, and retention, with every metric pinned to a definition and its source. It flags what changed and the likely signals behind it; the causal call stays a human judgment.
What it does
Any chatbot charts a metric once. Helio runs it as a standing job — from your context, kept current, delivered to you.
| Owns the job, not just the chart | It pulls the week's events, clusters scenarios, tracks cohorts, scans for anomalies, and delivers the report once every number carries its definition. |
| Runs again without starting over | The weekly report and the daily anomaly scan run the same vetted process on a schedule, not a fresh query each time. |
| Works from your context | It reads your PostHog or Mixpanel events and your locked metric definitions, so this week's numbers stay comparable to last week's. |
| Keeps itself current | When you change which scenarios or regions matter, it updates the report. You don't rebuild a dashboard. |
| Lands where you work | The report arrives in your team channel before Monday standup. You don't open a BI tool to find it. |
Three things it gets right
Every number carries its definition
The report labels what counts as active and what window each cohort covers, so this week's figure and last week's mean the same thing.
Two windows before it alarms
An anomaly fires only after the same move shows up across two windows, so a single-day dip stays quiet.
It watches its own report
If the expected report doesn't land on time, it fires a backup and alerts you the same hour.
What User Behavior Analytics connects to.
Questions
- Do I need analytics already set up?
- You need a source of usage events: PostHog, Mixpanel, Amplitude, or your own event table. Helio reads what you already have and doesn't replace your instrumentation.
- Does this expose individual users' data?
- No. Every report is aggregated and de-identified before it reaches you. Helio never surfaces a real name, email, or session in a weekly report.
- How is this different from a BI dashboard?
- A dashboard waits for you to open it and know what to look for. Helio decides what's worth reporting and shows up in your channel with it, including things you didn't think to ask about.
- Can it tell me why something changed, not just that it did?
- It flags the change and the most likely contributing signals, a channel spike or a cohort drop-off, but a causal explanation still needs a human judgment call.
- How do you keep this week's numbers comparable to last week's?
- Every metric's definition is agreed once up front, what counts as active and what window a cohort covers, then reused every week. The report labels its own definitions so a number always means the same thing it did last time.
Give Helio a week of usage data and get back a report worth reading.
Connect your event source, and the first weekly report lands in your workspace before Monday standup.
Ask about this automation