Product analytics
Insights, dashboards, the live event feed and your event schema.
Product analytics answers four questions about your events:
- How much, over time? Trends count events, people or companies, or add up a number like
amount. - Where do people drop off? Funnels measure conversion through a series of steps.
- Do they come back? Retention follows groups of people week by week.
- What's happening right now? Activity shows every event as it arrives.
Save any of these as an insight and put it on a dashboard. Product analytics is always on, and it includes people, cohorts and groups.
Insights
Trends, funnels and retention.
Dashboards
Build, customise and share boards of metrics and charts.
Activity
Activity in the sidebar lists every event, newest first. Use it to check that a new tracking call works, or to follow one user's session.
- With live on, the feed refreshes every few seconds. Turn it off to read without the list moving.
- Filter by event, date range, properties, person, group or cohort.
- Click a row to see all its properties, or click the person to open their profile.
- Load more goes further back in time.
Activity vs Changes
Activity shows the events your product sends. Changes on the Home page shows what people and agents changed in datahop: insights saved, flags switched, experiments launched.
Data
Data lists every event and property Datahop has seen, with its type, sample values and volume. You don't declare anything up front: a new event appears here a few seconds after it first arrives. There are four tabs: Events, Event properties, Person properties and Group properties.
Describe your events
Click Add description… next to an event or property, type, and press Enter. Everyone in the project sees descriptions, and so do agents, which read them before they write a query.
Compare what an agent sees without and with a description:
order_completed(2,856 events)order_completed(2,856 events): "Paid order, sent from the server after payment is confirmed.amountis incurrencyunits, after discounts, before tax."
With the second, the agent knows refunds and failed payments aren't included and that amounts aren't all in one currency. A good description says:
- when it fires (browser or server, before or after confirmation);
- the units and currency of numbers;
- known quirks ("before October this was also sent on retries");
- what not to use it for ("for revenue, use
order_completed, not this").