Multi-touch journeys

A journey is the complete sequence of sessions a single visitor has on your store, in chronological order, regardless of conversion outcome.

Every visitor has a journey. The journey accumulates over time as they return.

The simplest possible journey

14:32 — Direct → /products/red-shoes (3 pages, 4 min) → ORDER ($89)

One session, one conversion, journey complete. Single-touch.

This shape is common on the first day after install (most visitors haven't returned yet) and rare a week later (most converters have multi-session histories).

A more typical journey

Day 1, 14:32 — Meta paid social → /products/red-shoes  (4 min, no order)
Day 3, 09:15 — Klaviyo email     → /collections/sale   (1 min, no order)
Day 5, 21:08 — Direct            → /checkout → ORDER ($120)

Three sessions, three different channels, one conversion at the end. This is what attribution models actually have something interesting to say about — the credit allocation depends on which model you choose.

Why journeys densify over time

On day one of any new tracking install, you'll see mostly single-session journeys. Not because your customers are weird — because most visitors haven't returned yet. By day five, the visitors who started anonymous on day one have come back, and those single-session shadows have grown into multi-session journeys.

A useful mental model: every visitor on your store today is a future multi-touch journey, you just haven't seen the rest of it yet.

This is why attribution data feels thin on day one and rich by day seven. See Why your data feels thin on day one.

The lookback window

A journey doesn't extend forever. v1.0 uses a 30-day lookback window for attribution: only sessions within the 30 days before a conversion are credited.

If a visitor first found you 90 days ago, then converted today, the 90-day-old session is not in the attribution journey. The visitor's history goes back further (visible in Visitor Journeys) but only the last 30 days count for credit.

Why 30 days? It's a balance:

• Too short (e.g. 7 days) under-credits awareness channels with longer consideration cycles
• Too long (e.g. 90 days) over-credits incidental early touches that probably weren't decisive
• 30 days matches the typical Shopify consideration window for impulse-to-considered purchase

Configurable lookback is post-v1.

Anonymous portions of journeys

Most journeys have an anonymous prefix and an identified tail:

[Anonymous] Day 1  — Meta ad      → bounce
[Anonymous] Day 3  — Email click  → bounce
[Identified] Day 5 — Direct → checkout → ORDER (signs in / completes checkout, becomes identified)

When the visitor identifies on day 5, the prior anonymous sessions on the same cookie merge into the identified profile. The full 3-session journey becomes attributable, not just the closing session.

This retroactive merge is what makes journey-based attribution work — without it, anonymous traffic would be invisible to attribution models.

What journeys don't capture

Worth being explicit:

• Cross-device. A phone-then-laptop journey is two separate journeys until identification merges them via a Shopify customer login. Sessions on the other device before identification cannot be merged retroactively.
• Offline touches. TV, podcasts, billboards, in-person retail — none of this is in the journey unless a visitor types your URL or scans a UTM-tagged QR code.
• Pre-install history. If you installed FTS yesterday, journeys for visitors who first saw your store last week start at install date, not at first-ever visit.

Where to go next

• Attribution models — what models actually do with journeys
• Visitor Journeys feature — the in-app view
• Conversions and orders — how an order becomes a conversion attached to a journey