Open your analytics tool and look at yesterday's conversions. Every one of them gets told as a one-visit story: a source, a landing page, a signup, all inside a single session. Tidy, satisfying, and mostly fiction.
Real buying behavior is slower. Someone finds you through a blog post on Tuesday, skims the pricing page on their phone on Thursday, and signs up from a laptop the following week after searching your name directly. That is one journey. A tool that only thinks in sessions records it as three unrelated visitors, credits the conversion to "direct," and quietly buries the blog post that started everything.
The single-session illusion
Session-scoped reporting distorts three numbers you probably rely on:
- Bounce rate reads as failure.A visitor who lands, reads one page, and leaves looks lost. If they come back two days later and convert, that "bounce" was the first step of a successful journey. Without cross-session identity you can never tell the two apart.
- Attribution collapses to the last click. The final session before a conversion is usually branded search or direct, because by then the person already knows you. Last-click reporting rewards the victory lap and starves the channel that did the persuading.
- Conversion rate mixes incompatible groups. First-time visitors and fourth-time visitors convert at wildly different rates. Averaging them into one number tells you almost nothing about either.
None of this is fixed by collecting more events. It is fixed by remembering who the visitor is between sessions, which is exactly the thing fragile cookie-based identity fails at.
What a visitor profile actually shows
When identity survives across visits, every session attaches to a single profile, and the profile becomes a timeline you can read top to bottom:
- The first touch: the real source and landing page of the very first session, preserved no matter how many visits follow.
- Every session since: when they returned, what brought them back, which pages they read, what they clicked.
- Movement between sessions: the device switches, the browser changes, the city-to-city hops. Someone researching at the office and buying at home is one person, and the profile shows the handoff.
- Friction along the way: errors hit, rage clicks, forms abandoned, each tied to the session where it happened.
A profile like this answers the question a session list cannot: not "what happened on this visit" but "what has this person's whole relationship with the product looked like?"
Three questions only multi-session data can answer
1. How long does converting actually take?
Count the sessions and days between first visit and signup for your recent converts. If the median buyer needs four visits over nine days, that changes how you judge everything: a landing page's job on visit one is to be worth returning to, not to close the deal. Retargeting budgets, email timing, and trial lengths all calibrate against this number, and most teams have never measured it.
2. Which channels start journeys that end in revenue?
Group your converts by the source of their first session, not their last. The ranking usually looks very different: content and community discover people, branded search collects them. Judge each channel by the journeys it starts. Cutting a channel because it rarely gets last-click credit is how teams accidentally turn off the top of their own funnel.
3. What do returning visitors do differently?
Split any report by first-time versus returning and the fog lifts. Returning visitors head straight for pricing, docs, and the signup form. If your returners are not converting, the problem is rarely awareness; it is something specific on the pages they keep revisiting, which is where watching their sessions earns its keep.
Reading a journey in practice
Here is what this looks like on a real profile:
- Session 1, Tuesday, phone: arrives from a comparison article, reads it, opens pricing for eleven seconds, leaves. A session-scoped tool writes this off as a bounce.
- Session 2, Thursday, phone: returns via a saved tab, reads two docs pages, starts the signup form, abandons it at the card field.
- Session 3, Monday, laptop: branded search, straight to signup, converts in ninety seconds.
Read as one journey, the lessons are concrete: the comparison article is doing real acquisition work, the mobile checkout has a friction point worth fixing, and the "ninety-second conversion" was actually a six-day decision. Read as three anonymous sessions, you learn nothing and credit Google.
What you need in place
Multi-session journeys have two prerequisites:
- Durable anonymous identity. The visitor id has to survive cookie clears, incognito, and browser switches, or journeys shatter into fragments before they ever reach a signup. This is a collection-layer problem, not a reporting setting.
- An
identifycall on login. One line,oak.identify(userId), binds the anonymous history to the account, including the sessions that happened before signup. That retroactive stitch is what makes first-touch analysis possible at all.
With OakData both come with the one-snippet install; sessions, page context, and errors are captured automatically, so the journey assembles itself as visits accumulate.
And because profiles are queryable over the MCP server, an AI agent can walk the same timeline: paste a customer email into your editor and ask what their path to purchase looked like, or which journeys stalled at the same step this week.
OakData keeps every visit attached to the same visitor, from first anonymous session to paying customer, so the story you read is the one that actually happened.