Falorb
GitHubGet started
Cross-project insights/insights

Questions that span more than one product

Running several products creates a category of question that per-site analytics cannot answer at all: which of my properties feed each other, and who uses more than one. The insights builder is where those are asked.

Install FalorbSee the whole product
The insights page: a builder selecting metric, dimension and chart type, with results and a people-across-products table.The insights page: a builder selecting metric, dimension and chart type, with results and a people-across-products table.
Metric × dimension, across the portfolio

Pick a metric, a dimension and a shape

Visitors, sessions, pageviews, events, bounce rate or session duration, broken down by any allowed dimension, drawn as the chart that suits it. The whole query lives in the URL, so an insight is shared by sending the address.

People across products

Everybody seen on two or more of your properties in the range, with which ones, their event volume and when they were last active. On a portfolio this is the single most useful list there is: your existing audience, and it is the cheapest launch channel you have.

A people-across-products table: each row a person, the number of properties they used, which ones, event count and last seen.A people-across-products table: each row a person, the number of properties they used, which ones, event count and last seen.

Built on the identity graph, not on guesswork

The cross-product join is only as good as the identity resolution underneath it, which is why that resolution is deterministic. Two anonymous visitors are never assumed to be the same person because they look similar; they are the same person because the same identify() id appeared on both properties, or because one of your own decorated links carried them from one domain to the other.

Questions

Cross-project insights: the usual questions

How is this different from the portfolio overview?

The overview compares properties as units: visitors, sessions and trend per site. Insights compares the data inside them on a dimension you choose, and resolves the same people across them.

Related

Next, most people read

Person profilesEvery event, every device, every referrer and every product, on one page per person, resolved across the portfolio.Portfolio overviewTotals for the whole portfolio, then every property ranked with its own sparkline and period-over-period delta.RetentionCohort grids by day, week or month, and a stickiness distribution of how often people return within a period.

Run it yourself this afternoon

Clone the repository, bring up the stack with Docker Compose, paste one script tag. The data never leaves your machines.

QuickstartView on GitHub