Snowplow home
For analytics teams

Trusted answers from your own warehouse

Snowplow gives analytics teams the tools to collect and own event-level behavioral data in their warehouse, so every metric traces back to events and every analyst starts from the same tables.

  • Tables ready to query

    Views, sessions and users are modeled in your warehouse, ready for the BI tool you already use.

  • Numbers that trace back

    Every metric starts from an event row you can inspect, so any figure can be checked.

  • Questions beyond the defaults

    Attribution and journey logic is SQL your team owns and can change.

From your events to tables analysts can query

Start from modeled sessions, not raw logs

The Unified Digital dbt package builds views, sessions and users tables with measures such as engaged time and scroll depth, and your BI tool reads them directly.

Read about Unified Digital

Follow one person across devices

Identities adds a Snowplow ID to each event and links anonymous activity to the signed-in user across sessions and devices.

Read about Identities

Keep invalid events out of your reports

Events that fail their schema go to a separate table with the reason, so a broken tag shows up as a count, not a wrong chart.

Read about failed events

Strava’s analyst and product teams self-serve the data they need, from derived tables in Snowflake built on Snowplow events.

We would not have achieved our current level of self-serve data without Snowplow. It has enabled us to democratize our data culture, significantly improving our analytics coverage and deepening our insights.
Daniel HuangData Engineer, Strava
Read the Strava story

Questions

Do analysts need engineers to query the data?

Engineers set up tracking once. Analysts then work from warehouse tables in the SQL and BI tools they already use.

Is sessionization done for us?

Yes. The Unified Digital dbt package builds the views, sessions and users tables, updated incrementally.

Can we run our own attribution?

Yes. Events are rows in your warehouse, so model logic is a query you version and change.

What happens to a malformed event?

It is held out of the events table and listed with the reason it failed.

Bring the question your dashboards cannot answer

We will walk through it on your own events and tables.