Revenue attribution with Stripe

StatsKit attributes Stripe revenue to the visitor who earned it. When a purchase completes, it links back to the exact visitor and session, which means revenue rolls up to the experiment variant or flag the visitor was exposed to.

How the link is made

The tracker sets two first-party cookies on the visitor:

• statskit_visitor_id (the vid), the stable visitor identity.
• statskit_session_id (the sid), the current session.

When the visitor starts a Stripe Checkout through StatsKit, the vid and sid are stamped into the Stripe checkout session metadata. The purchase carries that metadata all the way through Stripe, so when it settles StatsKit reads the vid/sid back off the metadata and attributes the revenue to that visitor’s journey.

Visitor vid + sid cookies

→

/v1/checkout stamps metadata

→

Stripe payment

→

Attributed to experiment / flag

vid and sid ride into Stripe metadata, rolling revenue back to the visitor.

Why this ties revenue to experiments and flags

Because assignment is deterministic on the vid, the same vid that bought is the same vid that was bucketed into a variant or evaluated against a flag. Once the purchase is attributed to that vid, its revenue is attributed to whatever the visitor was exposed to:

• Experiments: revenue lands on the variant the visitor saw, so a price test or headline test shows revenue per variant, not just conversion counts.
• Flags: revenue lands on the flag value the visitor was served, so you can compare revenue for visitors who had a feature on versus off.

This is what makes revenue a first-class experiment metric rather than a number you reconcile by hand.

Setup

1. Install the tracker so vid/sid cookies are set. See install. No cookies, no attribution.

2. Connect Stripe in the dashboard so StatsKit can read settled purchases and their metadata. Connected keys are stored encrypted and read-only.

3. Send visitors through StatsKit checkout so the vid/sid get stamped into the Stripe session metadata. See pricing tables and checkout for the /v1/checkout endpoint. If you build the Stripe session yourself, add statskit_visitor_id and statskit_session_id to its metadata.

4. Read revenue in results. Attach a revenue custom metric to the experiment and revenue shows up per variant.

Tip

The reliable path is to route checkout through StatsKit’s /v1/checkout endpoint. It does the metadata stamping for you. Rolling your own Stripe session works too, as long as you copy the vid/sid cookies into the session metadata.

Caution

Attribution needs the cookies present at checkout. Cross-device journeys (browse on mobile, buy on desktop) will not link unless you also identify the visitor to stitch identities. Ad blockers that strip the tracker also strip the cookies.

Also track revenue server-side

Metadata attribution covers purchases that go through Stripe Checkout. For revenue that settles in your backend (subscription renewals, invoices, off-session charges), also emit a server event with the same unit so metrics see it. See the server revenue recipe.