How attribution works

If you understand this article, the rest of the KB will make sense in context. If you skip it, half the dashboard will feel arbitrary.

Attribution sounds technical. The core idea is simple: a visitor came to your store from somewhere, possibly returned from somewhere else, and eventually bought (or didn't). Attribution is the story we tell about which of those touches deserves credit.

Three things you need

Every attribution system depends on the same three building blocks:

1. Visitor identification. A way to recognize that the person browsing your store today is the same person who browsed yesterday, even if they're on a different page. Without this, every page view looks like a stranger.
2. Touch capture. A record of every meaningful interaction — page visits, channel they arrived from, the campaign they clicked.
3. Conversion tying. A way to connect "this visitor's journey" to "this order in Shopify."

FTS captures all three automatically the moment you install. Here's what each looks like.

1. Visitor identification

When a visitor lands on your store for the first time, our pixel drops a first-party cookie containing a unique visitor ID. Every subsequent page view, session, and click is tagged with that same ID. If they leave and come back two weeks later, the cookie is still there — same ID, same visitor.

A visitor can be in one of two states:

• Anonymous. We've seen them browse but we don't know who they are yet. We track their journey under the visitor ID alone.
• Identified. They've given us a signal that links the cookie to a real person — typically logging into their Shopify account, or completing a checkout. We then merge their identified history with their previous anonymous browsing under a single profile.

See Visitors, sessions, and identification for the full reference.

2. Touch capture

Every visit is logged as a "session" — a continuous browsing period of up to 30 minutes of inactivity. For each session, we record:

• The source. Where they came from — direct, organic search, paid search, social, email, referral.
• The medium. Specifically what kind of touch — cpc, email, organic, social.
• The campaign. UTM campaign value if present.
• The landing page. First URL they visited in the session.
• Pages visited. Sequence of URLs they viewed.
• Duration and timing.

Multiple sessions from the same visitor build into a multi-touch journey:

Day 1, 14:32 — Meta ad → /products/red-shoes (3 pages, 4 min)
Day 3, 09:15 — Email click → /collections/sale (2 pages, 1 min)
Day 5, 21:08 — Direct → /products/red-shoes → /cart → /checkout → ORDER ($89)

That's the journey. Three sessions, three different sources, one conversion.

3. Conversion tying

When a Shopify order is placed, two things happen simultaneously:

• Shopify sends us the order via webhook. We get the order ID, customer info, total, line items, timestamps.
• Our pixel fires a conversion event. This includes the visitor ID, so we know which journey to attach the order to.

The order is then "tied" to that visitor's journey. The journey + order together is what attribution models analyze.

In rare cases, the visitor → order link is missing — most often when the buyer uses Buy It Now or Shop Pay express checkout, which skips the page our pixel runs on. The order still lands accurately; only the marketing attribution to it is incomplete. See Why Buy It Now / Shop Pay shows as direct for the v1.0 limitation and the v1.x fix.

What an attribution model does

Once we have the journey + order, we apply a model to decide how credit is split among the touches.

The three v1.0 models:

You can switch models with one click in Pulse. The same journeys re-attribute instantly — there's no recalculation lag because the journey itself doesn't change, only the lens through which we view it.

For a deeper read: Picking the right attribution model.

Why no single model is "the truth"

If you remember nothing else from this article, remember this: every attribution model is a useful lie.

Marketing is multi-touch. A single Meta ad rarely "causes" a conversion all by itself; it works in concert with email, retargeting, organic search, and word of mouth. Any model that gives 100% credit to one touch is simplifying reality. The goal isn't to find the "true" model — there isn't one. The goal is to pick the model that answers the question you're asking today.

That's why FTS keeps all three models live simultaneously. Switch lenses depending on what you need to know.

Where to go next

• See attribution in your data → Pulse Dashboard overview
• Ready to install? → Installing on Shopify
• Want the full reference on each model? → Attribution models