User path analysis: the routes people actually take

What user path analysis is

User path analysis reconstructs the order in which a visitor moved through your product — the pages they saw and the things they clicked, in the sequence it actually happened. Where a page report counts how many people reached a page, a path report shows the route they took to get there and where they went next.

That one difference is why it answers a question nothing else in an analytics tool can. Page views tell you what was popular. Funnels tell you how many people finished a route you chose in advance. Paths tell you which routes exist.

The vocabulary is a mess, and it costs teams time. User path analysis, click path analysis, journey mapping and flow reports mostly describe one capability: sequence reconstruction from the events you already collect. Where they differ is scope. A path usually means one session. A journey usually means the whole relationship, across sessions and devices.

A funnel is a hypothesis. A path is what happened.

When you build a funnel you decide the steps in advance: landed, signed up, activated, paid. That is a hypothesis about the route, and it only reports on people who took it. Everyone who reached the same outcome another way, or wandered off somewhere you did not think to measure, is invisible.

Path analysis inverts that. Instead of asking "how many people completed my five steps", it asks "what routes do people actually take", and it answers with the pages and events in the order they really occurred. You get the detours, the loops, the back-and-forth between pricing and docs, and the exits you never modelled.

What a path report actually contains

A path report is not one picture, and treating it as one is why people bounce off it. The Paths view in PulsePanda answers four separate questions about the same set of sessions, and each one is read differently.

Entry pages. The first page of every session, ranked. This is your real landing-page list, and it is rarely the one in the design doc.

Top transitions. The most common single moves from one page to the next, written /pricing → /signup. This is the most trustworthy part of the report, because a transition needs far less traffic to be stable than a whole route does.

Top journeys. The most common ordered routes, cut to their first four steps so that near-identical paths group together instead of every session inventing its own.

Exit pages. The last page of every session, ranked — where sessions end, which is not the same as where people give up.

Above those four sits the line that frames them: how many sessions were analysed, and what share of them were single-page. A report where most sessions are one page has almost no paths in it, and no amount of staring at the transition list will change that — you are looking at a traffic quality problem, not a navigation problem.

How a path gets reconstructed

Worth knowing whichever tool you use, because the construction rules decide what the numbers are allowed to mean. This is what PulsePanda does:

• One sequence per session. Every session that started inside the window is taken in turn and flattened into an ordered list of pages.
• Built from page views and clicks. The page the session opened on starts the sequence; every page view and click event that carries a URL extends it, ordered by timestamp.
• URLs are reduced to paths. Host, query string, hash and trailing slash are dropped, so /pricing, /pricing/ and /pricing?utm_source=newsletter are one page instead of three. Campaign parameters would otherwise shatter your best page into a dozen rows.
• Consecutive repeats collapse. Twelve clicks on one page is one step. A path measures movement between pages, not activity inside them — that is what heatmaps and session replay are for.
• The window is the last 7 or 30 days of sessions, and it is a hard edge: a session that began before it is not in the report at all.
• Only the head of each list is kept — the top entries and exits, the top transitions and journeys. The tail of a path report is a very long list of routes that happened once, and it is noise.

Four things worth looking for

Entry pages that are not your home page. Most product teams design the home page route and let everything else happen by accident. If a third of sessions start on a blog post or a feature page, that page is your real landing experience and probably has no next step.

Loops. A user bouncing between two pages three or four times is not exploring, they are looking for something that is not there. Pricing to features to pricing is the classic one, and it usually means the pricing page is missing an answer.

Convergence. When many different starting points funnel into the same page before conversion, that page is load-bearing. It deserves more care than its traffic alone suggests.

Exit points with intent. An exit from the docs after a successful copy-paste is a good outcome. An exit from checkout is not. Raw exit rate cannot tell them apart; the path before the exit can.

Run your first user path analysis

1. Take the widest window you have. Paths fragment fast, and a week of a small site is mostly one-off routes.
2. Read the single-page share before anything else. It tells you how much of the report is real.
3. Compare the entry list against the pages you actually designed as entrances. Every mismatch is a page that needs a next step.
4. Pick your biggest entry page and find it on the left-hand side of the transition list. The moves out of it, ranked, are the routes you are really offering.
5. Take the transition that loses the most people and watch five recordings of sessions on that page. Five is usually enough to form one specific hypothesis.
6. Build the funnel for the route you want, ship the change, then watch that funnel rather than the path report.

Turn a path into a change

This is where path analysis pays off against a funnel. The funnel tells you the activation step drops 60% of users. The path tells you that half of them visited your pricing page immediately before leaving, which is a different problem with a different fix.

Once you have the hypothesis, build the funnel. Funnels are the right tool for measuring a route you have decided matters, and for watching whether your change moved it. Paths are how you find the route worth measuring.

Where a path report will mislead you

• An exit is not abandonment. A path report that ends at "left the site" may just be a session boundary, not abandonment. The same user often returns an hour later and finishes.
• A path is not a person. Paths are session-scoped, so one user across three visits is three paths, and on two devices two more. Read cross-visit behaviour in retention, not here.
• Small samples produce confident nonsense. Ten sessions produce ten unique routes and no pattern. If the top five transitions change every time you look, you are reading noise — stay on transitions until the weekly list holds still.
• Truncation hides endings. Top journeys show the opening steps of a route, not its conclusion. The ending is in the exit list.