Customer journey mapping is the practice of laying out, step by step, what a specific type of customer does, thinks and feels on the way to a goal such as buying, signing up or booking a demo.
For conversion work, a journey map is useful only when it is built from real data (analytics paths, surveys, recordings) and ends in a list of testable fixes at the points where people stall or drop out.
This guide covers what goes into a customer journey map, how a user journey map differs, and how to turn either one into A/B tests.
Every claim about what works is backed by tests from our database of A/B test results.
What a customer journey map is
Nielsen Norman Group defines a journey map as "a visualization of the process that a person goes through in order to accomplish a goal" (Sarah Gibbons, NN/g).
Its model has five parts: an actor, a scenario with expectations, journey phases, the actor's actions, mindsets and emotions in each phase, and the opportunities that follow.
NN/g also says the actions on the map should be "rooted in data" and the mindsets should use customer verbatims from research.
That is the part most teams skip, and it is the part that matters most for CRO.
A map built in a workshop from team opinions shows what the team believes.
A map built from surveys and path data shows where customers actually get stuck.
Only the second one reliably produces winning tests.
How it differs from related maps
Several mapping methods get confused with journey maps.
NN/g's UX mapping cheat sheet separates them like this:
| Method | Viewpoint | Ordered in time? | Best CRO use |
|---|---|---|---|
| Empathy map | One user type: says, thinks, feels, does | No | Early research summary before mapping |
| Customer journey map | One persona with one product or service | Yes | Finding the stages where people drop out and why |
| Experience map | Generic human behavior, not tied to a product | Yes | Background understanding of a category |
| Service blueprint | The organization: people, props and processes behind each touchpoint | Yes | Fixing backstage causes such as slow quotes or shipping delays |
For conversion work, the customer journey map is the main tool.
The others feed it or explain problems it reveals.
Customer journey map: what to include for CRO
A typical journey map has stages like awareness, consideration, purchase and retention.
A CRO-focused map keeps those stages but adds the numbers and evidence that turn it into a testing plan.
| Row on the map | What goes in it | Where the data comes from |
|---|---|---|
| Stage | Ad or search, landing page, product page, cart, checkout, post-purchase | Your site structure |
| Customer goal | What the visitor is trying to find out or do at this point | Customer interviews, jobs-to-be-done research |
| Questions and objections | The doubts that stop the next step, in customers' own words | Post-purchase surveys, on-site polls, support tickets |
| Behavior | Where people click, scroll, loop back or leave | Heatmaps, session recordings, path reports |
| Drop-off rate | Share of people who reach this stage but not the next | Funnel reports |
| Test ideas | Specific changes that answer the objection at this stage | The rows above |
The objections row does the most work. Two tests in our database show why.
Kasper Dolk's client kept seeing the same questions in customer surveys: is it durable, is it the official product, does it protect the phone camera.
They put short answers to those questions as benefit bullets in the product page hero, and Kasper Dolk reported +$763,679.14 in extra profit per year from the test.
Oliver Kenyon found that 70% of visitors to a bird food store's collection page left without opening a product, and most were first-time visitors.
Three lines taken from post-purchase survey answers went above the product grid (freshly packed, aflatoxin tested, free delivery carried to the door).
The result was $44,890 a month, or $0.66 per session.
In both cases the map's value was not the diagram.
It was knowing which objection appeared at which stage, then answering it at that exact point.
Map the stages people skip, too
Journey maps often assume one linear path. Real visitors arrive at different stages with different levels of awareness, and a page built for one stage can lose people at another.
Vasta tested a story-led advertorial against a standard product page for BJJ Fanatics.
The advertorial opened with the problem of training alone, introduced the coach and explained the method before showing the offer.
It produced +71.8% revenue per visitor, along with +46.5% AOV and +14.2% conversion rate.
The cold paid-traffic visitor needed a consideration stage the product page skipped.
One caveat from the source: the front-end offer was free, so revenue came from add-ons.
AGConsult built separate landing pages for Yoast visitors at different awareness stages (WordPress beginners, free-plugin users, web professionals) with benefit-led copy.
After a year of research and testing, they reported +62% conversion rate and +118% revenue.
This is a program result with bundled changes, not a single split test, so treat it as direction rather than a precise effect size.
User journey map: the product and app version
"User journey map" and "customer journey map" are often used interchangeably.
In practice, a user journey map usually covers one task inside a product, such as onboarding, creating a first project or upgrading, while a customer journey map spans the whole relationship from first ad to repeat purchase.
User journey maps are closer to a conversion funnel with context added.
They work well for SaaS signup flows, lead forms and mobile app onboarding, where the steps are fixed and every drop-off can be measured.
They can also show that one path does not fit every actor.
MarketingExperiments tested a healthcare homepage that offered separate paths for different visitor personas, each leading to the same "find a treatment center" page. Clickthrough rose +52% against a single-path homepage.
Applica rebuilt the Five Minute Journal app's onboarding around jobs-to-be-done research, reflecting the reasons users said they came.
The reported result was +20% ARPU, about +10% trial starts and +14% trial conversions, with several changes bundled together.
Adding steps is not always the fix
A journey map can make it tempting to add steps that "walk people through" value.
That can backfire. MarketingExperiments tested a three-step free trial signup on the RegOnline homepage against a two-step version, and completed leads fell -24.5%.
Later iterations on the same page did reach a large win, but only after more changes.
Removing a step can help instead. Visually tested a redirect for Bonafide Health that skipped a separate cart page between the cart drawer and checkout, and checkout rate rose +5.6%.
Visually is a vendor and the source does not state sample size or significance.
How to build a journey map from data
A short process that produces a map you can test from:
- Pick one actor and one goal. NN/g recommends one viewpoint per map. For CRO, pick your largest traffic segment, such as first-time mobile visitors from paid social.
- Pull the real paths. In GA4, path exploration shows the pages or events people go through as a tree. It can run backwards from an ending point, such as a purchase, to show how people got there. A single path exploration can have a starting point or an ending point, not both.
- Measure stage-to-stage drop-off. GA4's funnel exploration supports up to 10 steps, open or closed funnels, elapsed time between steps and the top next actions after each step.
- Collect the why. Run survey questions such as "what almost stopped you from buying today?" and watch recordings of sessions that drop at the worst stage.
- Write objections in customer words. Use verbatims, not summaries. "Is it the official one?" is testable. "Trust concerns" is not.
- Turn the worst stage into hypotheses. For each objection, write the change, the page and the metric. Then check our test database for results on similar changes before you build.
A worked example
Say an ecommerce store's funnel for paid social visitors looks like this over 30 days:
| Stage | Users | Continue to next stage |
|---|---|---|
| Landing on collection page | 40,000 | 12,000 (30%) |
| Product page | 12,000 | 2,400 (20%) |
| Cart | 2,400 | 1,440 (60%) |
| Checkout | 1,440 | 1,000 (69%) |
The overall conversion rate is 1,000 / 40,000 = 2.5%.
The cart and checkout stages look healthy. The collection page loses 70% of visitors, and the product page loses 80% of those who reach it.
Raising collection-page continuation from 30% to 33% would send 1,200 more people to product pages each month.
At the same downstream rates, that is about 100 more orders.
The same relative gain at checkout adds about 100 orders too, but checkout is already performing, so there is less room to improve it.
The map then asks the qualitative question: what do first-time visitors from paid social need at the collection page that they are not getting?
That is the exact situation behind Oliver Kenyon's bird food test above.
When journey mapping is the wrong tool
Journey mapping is not always worth the time. Skip it or keep it light when:
- You have one obvious leak. If 80% of mobile users abandon on a broken checkout field, fix the field. A map adds nothing.
- Traffic is too low to test the ideas. A map produces hypotheses. If you cannot test them, spend the time on usability testing and direct fixes instead.
- The map would be built from assumptions. A workshop map with no survey, interview or analytics input mainly confirms what the team already believed.
- You need attribution, not experience. If the question is which channel deserves credit for a sale, use attribution analysis, not a journey map.
Also be careful with personalized paths. Clean Commit showed returning visitors a shorter path with highlighted offers and saw a +39.3% conversion rate headline, but the result did not reach significance and overlapping tests and holiday promotions interfered.
A segment-specific path suggested by a map still has to win a clean test.
Journey-mapping tests from the database
These tests each started from knowing what a customer needed at one stage of the journey:
@oliverkenyon
Vasta
@kasper_dolk
MarketingExperimentsMarketingExperimentsThe insurer test shows the retention end of the map.
A North American health and life insurer showed returning visitors who had abandoned an application a widget reminding them of their quote.
Applications submitted rose +23.47%, and applications started rose +95.83%.
If you want help turning research into a testing plan, a practitioner who runs survey-led tests can build the map with you.