Map the customer journey from first question to repeat purchase

Chapter 9 helped you choose where a useful answer might be found. But a person rarely completes their whole task in one encounter. They may first ask a question, look elsewhere, compare, choose, try to use an offer, seek help and later decide whether it is worth returning. Each event may happen through a different channel and involve another person or team. A customer journey map helps you see those connected events from the person's point of view.

This chapter shows how to define a small, honest map, decide which moments matter, label evidence and use the map to improve an actual handoff. You will build a blank map rather than invent a customer biography, quotation, emotion curve or successful purchase. The ten figures are explained in text, and all table-like figures are WebP images for Trix. Two videos near the end are optional.

By the end you should be able to: explain the difference between a journey map and a sales funnel; name one actor, goal and boundary; record actions, questions and touchpoints without guessing feelings; identify a transition that may need improvement; and set a fair check after making a change.
A10-01-where-this-fits
Figure 1
Figure 1 in words: Chapter 9 asked where a relevant answer could appear. Chapter 10 follows what the person may need across several encounters, including use and help after purchase. Chapter 11 will place marketing activities into the wider RACE planning framework. A journey is a view of a person's task; a marketing framework is a way to organise the business's work.

What is a customer journey?

A journey is a sequence of actions and decisions through which someone tries to accomplish a goal over time. A journey map is a visual summary of that sequence from one chosen point of view. Nielsen Norman Group describes five common components: an actor, a scenario and expectations, phases, actions and mindsets, and opportunities (Gibbons, 2018). The details and phase names should come from the actual situation. There is no universal customer path that every business can copy.

The actor is the person whose experience the map follows. The actor might be a prospective buyer, a user who did not pay, a payer acting for someone else or a person seeking support. One map should have one clear point of view. When several roles make different decisions, make separate views or clearly mark their handoff rather than combining contradictory thoughts into one imaginary person.

The goal belongs to the actor. “Complete the task with enough confidence to decide whether this offer is suitable” is a different goal from the business's “increase sales.” Both matter, but the map should not erase the first one to make the second look smooth. Lemon and Verhoef (2016) describe customer experience across multiple touchpoints and channels, including those a business does not fully control. A map that begins with the first ad and ends at payment misses much of that experience.

The boundary says where this map starts and ends. “From the first question about a particular task to a decision after using the service” is more manageable than “the whole lifetime of every customer.” The time span may be minutes, months or longer. State it explicitly so a reader can tell what has been left outside.

A map is not a funnel, a flowchart or a promise

A sales funnel groups people or events by the business's stages, often from awareness toward purchase. It can help summarise aggregate activity. A user flow shows the steps in a particular interface or task, such as moving between screens. A service blueprint adds what the organisation and its partners do behind the scenes to deliver the visible experience. A customer journey can involve all of these but asks a different question: what does the person try to do over time, across channels, and where do they need help? GOV.UK's Service Manual recommends looking at the user's whole problem, including online and offline touchpoints and back-end work when a service spans organisations (Design community, 2019).

Do not confuse a tidy drawing with a real path. Someone may pause, compare again, move from phone to desktop, ask another person, choose a competitor, return for help or decide not to buy. Some customers never repeat a purchase because the need is one-off; others may renew, return or recommend after the offer has delivered value. These are possible routes, not a ladder that every person climbs.
A10-02-maps-compared-table
Figure 2
Figure 2 in words: A journey map follows a person's goal across encounters. A funnel summarises business-defined stage counts. A user flow details one interaction path. A service blueprint connects visible events to delivery work behind them. Choose the representation that answers the question you have; do not turn a funnel count into an invented personal story.

Scope first: one actor, task, period and map type

Before arranging boxes, write four lines:

  • Actor: which real group or role is the view about? Use your evidence-labelled audience description from Chapter 6. Do not copy an invented persona biography into the map.

  • Task: what is the person trying to accomplish? Use a plain verb, such as decide, begin, use, resolve or renew. Avoid “move through our funnel.”

  • Start and end: when does this map begin, and what marks its current stopping point? Include leaving or no purchase as possible outcomes.

  • Map type: are you describing the current state as observed or an intended future state you hope to create? A proposal should be visibly labelled, not presented as research.

A current-state map built with very little research is still possible as an assumption map, provided every guess remains marked and a plan to check important unknowns follows. Nielsen Norman Group warns that a map made only from internal assumptions can become a convincing but inaccurate account of the user's experience (Kaplan, 2019). Use it to organise questions, not to claim that people felt something you have not asked them.

If the task crosses several roles, keep a primary actor and note the other person at the handoff. If the roles have different goals, make another map when evidence allows. A single map can be small. It is more useful to understand one consequential transition than to fill a giant wall with guessed thoughts.
A10-03-scope-card
Figure 3
Figure 3 in words: Write a clear point of view, a task and a boundary before listing phases. Mark whether the map describes what currently happens, what you assume happens or what you propose. That label keeps a future design from being mistaken for a finding.

Six helpful prompts, not six compulsory stages

For a beginner exercise, use six prompts to look across the relationship: first question, investigate, choose, begin and use, get help, and continue, return or leave. They are not a proprietary framework and do not prescribe what any individual will do. Rename, combine, repeat or remove them when your evidence shows a different path. The important thing is the reader's task, not memorising a stage name.
A10-04-nonlinear-journey
Figure 4
Figure 4 in words: The six cards prompt questions from first need through use, help and possible continuation. They are deliberately unconnected: a person can return to an earlier question, change channel, pause or leave. The diagram does not claim that every moment occurs in every business.

1. First question

What first makes the task worth attention? The need may arise outside anything the business published. Record a question the person actually asked, or mark it as unknown. A search term, a direct enquiry and a support request are different kinds of evidence. Do not infer the underlying emotion from a single page view.

Ask where an answer might be found without requiring the person to know your brand. The first useful touchpoint might be a public explanation, another person's recommendation, an independent resource, a search result or an offline conversation. Touchpoint means an encounter that affects the task; it need not be an ad or a page owned by the business.

2. Investigate

The person may gather information and compare alternatives, including doing nothing. What facts do they need to understand suitability, cost, time, access and limits? Chapter 7 supplied the decision questions and Chapter 8 supplied claim and evidence checks. Here you record where those questions are answered and where they remain unresolved.

Do not assign a channel to a phase as if it were inevitable. A person can compare on a phone, speak to someone, revisit a saved page or use an independent review. The map should show the channels that actual evidence supports, not only the channels the business paid for.

3. Choose

Choosing can mean a purchase, an enquiry, a trial, a request for approval, a pause or a decision that the offer is unsuitable. What must the person know at the moment of action? Are terms and full costs clear? Does the label on the button match what happens next? The distinction between an intention and a completed step remains important.

Mark each handoff: a form to a confirmation, a sales conversation to a contract, or an order to fulfilment. If another person approves or pays, the primary actor may wait without the business seeing why. Do not call every incomplete form “lost interest.” A good map makes that unknown visible.

4. Begin and use

Payment is not the end of the task. What does the person need to start? Is the first instruction understandable? Does the service do what the earlier message led them to expect? A handoff is the transfer of information, responsibility or work between stages, channels, people or systems. Many difficulties occur at such boundaries even when each individual page looks polished.

Map only key actions and questions. A detailed screen-by-screen flow belongs in a separate user-flow diagram. The journey map should keep the broader goal visible: did the person get to use the offer and reach the result they sought?

5. Get help

Support may occur before or after purchase, and people may skip it entirely. The question is whether help is findable, consistent with the promise and able to resolve the task. If a support route is available only at certain times, explain that before the person depends on it. If a message refers someone to another team, what information travels with them, and what must they repeat?

Support records can reveal where earlier explanations failed, but they have limits: they represent people who contacted support, not everyone who struggled. Record the issue and source without turning a ticket count into an unverified claim about all customers.

6. Continue, return or leave

After use, the person may continue, renew, buy again, recommend, ask for a refund, switch to an alternative or stop because the need has ended. Retention means continuing a relationship or use over a period. Repeat purchase is a later purchase; it is one possible outcome, not the only sign of success. A person who stops because the service fulfilled a one-time task may still have had a good experience.

Ask what the person learned from the result. Does a renewal or repeat action require new information, approval or a fresh choice? Never assume the first purchase guarantees later consent to marketing. Make the terms, support and ending route understandable. A map should be able to show an honest exit without treating it as a design failure by definition.
A10-05-map-layers
Figure 5
Figure 5 in words: For each important moment, ask what the person did, what they needed to know, where the interaction occurred, what source supports the note and what might be improved. Keep opportunities separate from observations; the proposed fix is a hypothesis until tested.

What evidence belongs on the map?

Begin with observed actions, reported accounts, secondary records, assumptions and unknowns as separate labels. This is the same evidence habit used in earlier chapters. A direct observation can show an action. An interview can tell you what someone recalls and how they describe it. A support log can show a contact and issue. Analytics can show an aggregate path but rarely explain a thought. A claim about frustration needs a person's own account or other appropriate research; a red “emotion curve” drawn by a team is not proof.

Kaplan (2019) recommends looking at existing data first, then using qualitative research such as interviews, field studies or diary studies to understand actions and mindsets across the journey. GOV.UK's guidance likewise suggests speaking to a range of users and using the map to discuss gaps between teams and services (User research community, n.d.). You do not need to conduct every method for a beginner map. Do need to say what you have and what you lack.

Triangulation means comparing more than one kind of evidence. A participant's account of a confusing email, the email itself and a support question about the same step together provide a stronger basis for investigating it than any one alone. They still do not prove how every person behaves. If sources disagree, keep the disagreement visible and ask what differs in their situations.

Do not copy a full customer quote without permission or expose identifying details in a public teaching article. In your private working notes, record the source, date and context needed to assess a finding. When a journey uses several organisations or devices, avoid joining records about one person beyond what is appropriate and permitted. The map is meant to improve the service, not to gather an unlimited personal dossier.

When sources do not describe the same journey

Different sources may appear to contradict one another because they concern different people, dates or parts of the task. An aggregate analytics path may show that a page was visited, while an interview may describe an offline conversation that preceded it. A support record may begin only after the person has already spent time looking for help. Do not force these fragments into a single continuous biography. First write down the time period, the role and the question each source can answer. Then decide whether the difference reveals a missing touchpoint, a changed service or a genuinely different route.

This also explains why one map is not automatically representative. A map from one account is useful for understanding that account; it cannot establish how common a difficulty is. Several people can share a task yet need different support, have different access to a device or arrive with different prior knowledge. If a difference changes the decision you need to make, consider separate maps or a clearly labelled branch. Do not draw an average person who experienced incompatible routes at once. Kaplan (2019) treats research as the foundation of a reliable map, rather than a decoration added after the design.

You may be tempted to add a rising and falling emotion line. Emotion matters, but a line is especially easy to invent. Ask people about a recent real experience, use their own descriptions in the research record when appropriate and avoid treating a single comment as the feeling of an entire audience. If you have no such evidence, leave the emotion row out. A map of actions, questions and handoffs can already be useful.
A10-06-evidence-table
Figure 6
Figure 6 in words: Observation supports what was seen; a report supports what a person said; an operational record supports what the system recorded; an assumption is a question to test; unknown means there is no basis yet. None of these labels silently turns a guess into a universal customer feeling.

Follow the handoffs across channels

Make a small list of touchpoints for your scoped task. Include those you own and those you do not: a search result, a third-party explanation, a sales exchange, a page, a form, a payment service, a confirmation, the product itself, a support response and the ending route may matter. Do not include them all automatically. Choose only the encounters supported by evidence or needed as clearly marked questions.

At each transition, ask what the person expects to happen next, what actually happens and who is responsible for keeping the promise. A message may promise one thing while the sign-up form asks for more information. A payment may succeed while access instructions are delayed. Support may not know what the earlier page said. Those are possible handoff questions, not assertions about a real business.

The journey is visible from the person's side; the operational work behind it may require a service blueprint. GOV.UK suggests mapping back-end processes and the teams involved when understanding a wider journey (Design community, 2019). The simplest addition to your map is a column for owner or next team. If nobody owns a transition, that is a useful organisational question. Do not mistake an internal workflow for the customer's experience.
A10-07-handoff
Figure 7
Figure 7 in words: The person takes an action and expects a clear result. A visible response tells them what happened. Behind it, a team or system may need to complete work. The three layers should agree. This is an original teaching layout, not a claim about a particular product.

If a person switches channels, check continuity. Can they recover the same terms and status? Must they repeat information? Can they reach a human when necessary? A customer may not care which team owns a problem; they care whether the task can move forward. Mapping transitions reveals why an excellent advert cannot compensate for a broken service step.

Keep a promise visible across the whole task

An early message may make a claim about cost, availability, delivery, support or ease of use. Trace that claim through the next relevant touchpoints. Is the same meaning available at the point of choice? Does the confirmation explain what happens next? Can the person still find the terms when they begin using the offer? Is there a sensible route if the promised result does not occur? You are checking the continuity of information, not assuming that every touchpoint must repeat the same paragraph.

Write the claim exactly in the working record and note where you saw it. Compare it with the next real page, message or service step. If the wording differs, mark a question for the responsible team rather than declaring a failure from memory. This is where the Chapter 8 evidence habit becomes practical: a claim that is precise enough to be checked can be carried through a map. Vague promises are harder to verify and harder for a customer to rely on.

A useful map should also show who can act on a problem. A marketing team may own the first explanation while another team owns payment, access or support. The reader experiences one task across those boundaries. Ask the owner of the later step what they need from the earlier one, and ask whether the earlier message accurately describes the later reality. GOV.UK's guidance on whole problems makes this cross-team perspective explicit (Design community, 2019). A coloured box alone cannot settle an ownership gap; a named next check can.
A10-08-handoff-check-table
Figure 8
Figure 8 in words: For a chosen transition, record what the person expects, what is actually visible, who owns the next step and what evidence or test is needed. The rows use generic transition prompts only; no customer experience or operational result is fabricated.

Make a first map without inventing a story

You can start with a real task you have personally completed, a service flow you can inspect, or your own organisation's records when you are authorised to use them. A personal account is one account, not a representative journey. If you do not have customer evidence, make an assumption map with blank rows and clearly labelled questions. Resist the temptation to write a fictional name, biography or feeling to make the visual look finished.

Use the six moment prompts as optional row headings. In each row, write an action or question, a touchpoint, a source/status, and a possible handoff or improvement question. You may leave a row empty, combine two rows or add a loop. If the journey is short, a four-row map may be better than a grand six-stage one. A map earns its value through accuracy and a decision it helps make, not through the number of coloured boxes.
A10-09-journey-worksheet
Figure 9
Figure 9 in words: Begin by naming an actor, goal, boundary and current or proposed state. The rows for six possible moments are blank. Add a relevant action or question, where it occurs, the source and evidence label, and the next check. “Unknown” is a valid entry. The worksheet is not a completed customer story.

A 40-minute mapping exercise

  1. Set scope (5 minutes). Name one actor or role, one goal, a start and end, and whether you are mapping current experience or a proposed change. Keep another role in a separate note.

  2. Collect what exists (5 minutes). List relevant pages, messages, records, interviews or observations. Mark sources and dates. Do not collect identifying information you do not need.

  3. Write only supported actions (10 minutes). Put the key actions and decisions into rough time order. Use the stage prompts as a flexible guide. Add pause, no purchase, help and leaving where they are relevant.

  4. Add questions and touchpoints (5 minutes). What did a person ask, or what remains a question for you? Where could they find an answer? Mark whether it is observed, reported, assumed or unknown.

  5. Find one transition (5 minutes). Locate a handoff between channels, people or systems. What promise must survive it? Who owns the next response?

  6. Choose a check (5 minutes). Decide what evidence would confirm or correct the most consequential unknown. It might be observing a task, asking about a recent real experience or reviewing a specific support pattern.

  7. Name one improvement hypothesis (5 minutes). Write what you might change and how you would tell whether it helped. Keep it separate from the finding itself.

The exercise is complete when another reader can tell which parts happened, which were reported and which are still hypotheses. If you cannot answer a row, do not fill it for decoration.

Review the map with a second person

Ask someone who knows a different part of the service to read the map without your explanation. Can they tell whose task it follows? Which steps are grounded in evidence? Where does a handoff occur? What happens after the first purchase or use? If they point to a missing transition, add it as a question and find an appropriate source. Do not let an internal discussion silently rewrite what a customer said.

Next, try to read the map as a beginner. Expand abbreviations, identify the owner of each next step and make the ending understandable. A legend for evidence labels helps, but plain words in the map matter more. If the visual needs a long spoken explanation to make sense, narrow the scope or split the map. The purpose is to help people make a decision together, not to win a design contest.

Improve the experience, then update the map

A pain point is a difficulty the person encounters while trying to reach the goal. It is best stated with evidence: where it appears, what happens and how it affects the task. An opportunity is a possible improvement, which must still be tested. A complaint count or drop-off point may identify where to look; it does not by itself establish a motive. Put effort into the issues that matter to the person's goal and can be changed, not merely the ones easiest to count.

Choose one improvement that an owner can make. Clarify a missing explanation, fix a transition, make a confirmation understandable or provide a route to help. Define a check that looks at more than clicks: could a person complete the task, understand what happened and obtain the promised result? If the change affects a long journey, follow up later rather than declaring success at the first button.

The map should be revised when the service, evidence or audience changes. A future-state map can show what you intend to build, but should not replace the current-state map. Keep the gap between them visible. GOV.UK recommends revisiting a draft map with people involved in research and delivery and using it for ongoing discussions about improvements (User research community, n.d.).
A10-10-improvement-cycle
Figure 10
Figure 10 in words: Start with a scoped goal, gather evidence, find one consequential handoff, make a specific change, check the person's outcome and revise the map. A repeat purchase can be one later outcome, but an honest decision to leave or not purchase also belongs in the evidence.

For a readable introduction beyond this chapter, see Nielsen Norman Group's Journey Mapping 101. For a deeper academic view across multiple channels and touchpoints, see Lemon and Verhoef's research article. The GOV.UK experience-mapping guide is a practical authoritative source for maps of longer services. These are external links because no verified published internal chapter on journey mapping has been identified. You can understand this article without opening them.

Optional videos

These are two complementary YouTube lessons, not required reading. Original publisher listings, public metadata and descriptions were checked; the videos were not played in full, and captions, UK playback and embedding were not verified. Confirm those before publication.

See the basic structure: Customer Journey Mapping 101, Nielsen Norman Group / Sarah Gibbons (2019), about two minutes according to the publisher listing. It describes five components and why the map helps a team see the bigger picture. Use it after Figure 3 to revisit the vocabulary.

Check your map for common mistakes: Common Journey Mapping Mistakes, Nielsen Norman Group (2025), a short video with chapters through 03:18. Its description highlights overly broad scope, missing research and a business-centred view. Use it to review the worksheet you made. Neither video supplies evidence about your own customers.

What to remember

  • A journey map follows one actor's goal across encounters. It is not a count of people in a funnel or a transcript of every screen.

  • State the goal, boundary and whether the map is current, assumed or proposed before drawing phases.

  • The path can pause, loop, involve another person or end without purchase. Repeat purchase is one possible outcome.

  • Label observed actions, reports, assumptions and unknowns. Do not invent an emotion curve or customer quotation.

  • Follow handoffs into use and support; a good message cannot fix a broken transition.

  • Make one evidence-backed improvement and check the result for the person, then revise the map.

Next in the series: Chapter 11, Use RACE to attract, involve, convert and keep customers, will connect activities across the wider relationship. Add its canonical public link only after publication and verification.

References

Design community (2019) ‘Map and understand a user's whole problem’, GOV.UK Service Manual, 1 May. Available at: https://www.gov.uk/service-manual/design/map-a-users-whole-problem (Accessed: 14 November 2026).
Gibbons, S. (2018) ‘Journey Mapping 101’, Nielsen Norman Group, 9 December. Last reviewed 15 July 2026. Available at: https://www.nngroup.com/articles/journey-mapping-101/ (Accessed: 14 November 2026).
Kaplan, K. (2019) ‘How to conduct research for customer journey-mapping’, Nielsen Norman Group, 10 February. Available at: https://www.nngroup.com/articles/research-journey-mapping/ (Accessed: 14 November 2026).
Lemon, K.N. and Verhoef, P.C. (2016) ‘Understanding customer experience throughout the customer journey’, Journal of Marketing, 80(6), pp. 69–96. doi: 10.1509/jm.15.0420.
User research community (n.d.) ‘Creating an experience map’, GOV.UK Service Manual. Available at: https://www.gov.uk/service-manual/user-research/creating-an-experience-map (Accessed: 14 November 2026).

Supplementary videos: Gibbons, S. (2019) ‘Customer Journey Mapping 101’, YouTube, Nielsen Norman Group, 10 May. Available at: https://www.youtube.com/watch?v=2W13ext26kQ (Accessed: 14 November 2026). Nielsen Norman Group (2025) ‘Common Journey Mapping Mistakes’, YouTube, 15 December. Available at: https://www.youtube.com/watch?v=PIw-ooBJLS0 (Accessed: 14 November 2026).