Customer journey analysis: find the drop before you design

Stage three of eight. The question it answers: where does someone fall out between interest and a conversation?

A manufacturer's digital team told us they had never reviewed the customer journey end to end. Not because they were careless. They have a search team, a conversion team, and a content team, each improving its own stretch of the path. Nobody owns the seams between them, and the seams are where visitors fall out.

That is the normal condition of a mature site. Customer journey mapping happened once, years ago, in a workshop, and the map has been a poster since. The teams optimize locally. The organization has no current picture of where interest turns into a conversation and where it quietly does not.

Customer journey analysis is how a redesign gets that picture back, and it has to be done in stage three, after the audiences are ranked and before the structure is derived. Too early and there are no personas to map for. Too late and the structure has already decided the journey.

Map it top-down, then prove it bottom-up

There are two ways to draw a journey, and a redesign needs both.

Top-down starts from business intent. What should happen when a prospect arrives? What is the path the business wants them to take, and what is the conversation it wants to end in? This produces the user journey map most teams recognize: phases, the visitor's state at each, the pain points and opportunities. It is essential and it is a hypothesis.

Bottom-up starts from behavior. Analytics show which pages actually feed the endpoint. Heat maps show where attention pools and where it dies on the page. Session data shows the paths people took rather than the one the business drew. On a London university's site this meant a forward flow from the homepage through programme pages to an application, and a reverse flow engineered backward from the application to the pages that truly fed it, with heat-map evidence layered over both. The two rarely matched, and every mismatch was a finding.

User journey map skeleton: phases across the top; the visitor's goal, state, evidence, pain points, and opportunities down the side; top-down intent and bottom-up behavior recorded in separate rows.

The rule that follows: map across every major entry point, not just the homepage. Most visitors do not arrive at the front door. A journey drawn from the homepage down describes a minority of real paths and, for the growth audience, almost none of them.

The user flow diagram is where the journey becomes a decision

A journey map is a picture of experience. A user flow diagram is a picture of decisions: at each step, what the visitor can do, what the system does in response, and where the branches lead. It is the artifact that the structure and design stages actually build from.

On a global certification body's site, flows were drawn per sub-segment of the growth audience rather than per persona, because the sub-segments hesitated for different reasons and needed different first steps. The flows went through several drafts, each balancing what the research recommended against what the year's implementation priorities could carry. That tension is not a flaw in the process. It is the process.

Two kinds of flow matter, and teams often draw only one. Task flows follow one visitor through one goal and show the path. Wire flows show every state a visitor could reach and expose coverage: the dead ends, the states with no way back, the branches nobody designed for. A journey is not understood until both exist.

Synthesis is where the business argues with the research

Research surfaces loud ideas. Users want it simpler, cheaper, faster, with pricing on the page and no commitment up front. All of that is usually true and some of it is impossible, and the place to find out which is synthesis, not the design review.

The method is to sort every finding into three layers, business, technology, and experience, and then let the layers argue. A finding that is an experience win, a technology cost, and a business risk gets decided as such, out loud, with the decision recorded.

Research synthesis board in three layers: each finding placed against business, technology, and experience, with the decision and its owner recorded.

On the certification body's project, the loudest finding was that prospects wanted a low-barrier way to try the activity once before committing to a course. The business's revenue driver was the full course. The synthesis conversation did not pick one. It reframed the low-barrier option as an alternate track a prospect could choose, rather than a ladder everyone climbed, and led the primary path with the course that carried margin. Had that argument surfaced in stage seven, as a surprise in a hi-fi review, it would have cost weeks. In stage three it cost an afternoon.

A smaller example from the same synthesis: research findings had used the phrase "transparent pricing" as a recommendation. Before the client presentation it was removed, because pricing visibility was a commercial decision the business had not made and was not going to make in a UX readout. Synthesis is also where the team decides what is theirs to recommend.

Some findings are true and cannot ship yet

Not every journey the research reveals can be built in the first release. On the certification body's project, one growth sub-segment had a clearly mapped path and no product that could convert it in release one. The client's review feedback was direct: do not show it as live. Park it, label it, and keep it in the backlog with its evidence attached.

For a long-established business the version of this question is larger. A family-owned manufacturer of sixty or a hundred years has deep business process underneath the website: how leads are handed to branches, how professionals are credentialed, how pricing is set by region. Research will recommend journeys those processes cannot yet support. The honest question in synthesis is appetite: is this something the organization is prepared to change, or an aspiration for the roadmap? Both answers are fine. Pretending the second is the first is how redesigns produce journeys that lead to a form nobody answers.

Draw the flows before the pages

The client's own product leadership on the certification body's project asked for two things before any page-level design: a document defining the segments and the scenarios each would move through, and a visual flow map. They were right. Page designs reviewed without a flow to evaluate them against produce feedback about the page, not the journey, and the journey is what the redesign exists to fix.

This is also why the journey stage feeds the structure stage rather than the other way around. Card sorting per persona, the first act of the structure stage, needs to know which journeys each persona is on and which content matters at each phase. Without the flows, the card sort tests the current site's content rather than the content the journey needs.

What the journey stage hands over

Journey maps per persona and sub-segment: phases, state, pain points, opportunities, with top-down intent and bottom-up evidence recorded separately.
The forward and reverse flow analysis per audience, from every major entry point.
User flow diagrams per sub-segment: task flows for the path, wire flows for coverage.
The synthesis board: every finding placed against business, technology, and experience, with the decision and its owner.
The parked list: journeys that are mapped, evidenced, and not shippable in release one, with the reason.
The appetite-to-change register: the business processes the recommended journeys depend on, and which the organization has agreed to change.

What breaks when the stage is skipped

The structure is derived from the current site's content rather than the journey's needs, and the card sort validates the past.

The loud findings surface in design review as surprises, and the argument about margin versus low barrier happens over a finished screen.

Journeys are designed for processes the business has not agreed to change, and lead to forms nobody owns.

The homepage journey is the only journey, and the growth audience, which never arrives there, is designed for last.

Where this sits in the sequence

Stage three depends on stage two, because journeys are mapped per sub-segment and the sub-segments come from the audience work. It depends on stage one for the endpoints that count. It feeds stage four directly, because card sorting and tree testing run against the journeys, and stage six, because the conversion funnel is the journey instrumented.

Questions we get asked about customer journey analysis

What is the difference between customer journey mapping and customer journey analysis?

Mapping draws the journey the business intends, usually from research and workshops. Analysis tests that map against behavior: analytics, heat maps, and session paths, drawn forward from entry points and backward from the endpoint. Mapping produces a hypothesis; analysis produces findings.

What is a user flow diagram, and how is it different from a journey map?

A journey map describes experience across phases: goals, feelings, pain points. A user flow diagram describes decisions: at each step, what the visitor can do and where each choice leads. The journey map explains why; the flow diagram is what gets built. Draw task flows for the path and wire flows for coverage.

When should journey mapping happen in a website redesign?

After the audiences are ranked and before the structure is derived. Earlier and there are no personas to map for. Later and the information architecture has already decided the journey, which is backwards.

How do you handle research findings the business cannot act on?

Sort every finding into business, technology, and experience, decide it out loud in synthesis, and record the decision. Findings that are true but not shippable are parked with their evidence, not hidden. Findings that depend on a business process changing are logged with the question of whether the organization intends to change it.

Customer Journey Analysis: Find Where They Fall Out Before You Design Anything