Customer Journey Map Template: The Five Rows That Earn Their Place

A customer journey map is a table. Stages across the top, and under each one what the customer is trying to do, what they are thinking, where they are, and what breaks. Everything else that gets called a journey map is decoration on that table.

This is for whoever has to turn "we should understand our buyer better" into something a team can act on. Below is the grid itself, with one column filled in from our own buyer; the five rows worth keeping; and the step most templates leave out, which is how you check the map against something other than opinion. We ran that check on this site: it put nothing at the first two stages and 45 per cent of the impressions those queries drew at comparing and deciding, which is not where a map drawn from instinct tends to put them.

What is a customer journey map?

A record of how one type of customer moves from not knowing they have a problem to having solved it, laid out in order, with the friction marked.

Two constraints make it useful rather than decorative. One persona per map. A map that averages an economic buyer and an end user describes neither. One scenario per map. "Buying for the first time" and "renewing" are different journeys and merging them produces a diagram that is true of nothing.

It is not a funnel. A funnel describes your process; a journey map describes theirs, including the parts you have no involvement in, which is usually where the decision is actually made.

What is the customer journey map template?

Five rows against your stages. Here it is as a grid, with the "comparing" column filled in from the buyer this site actually serves, so you can see what a populated cell looks like rather than what an empty one promises.

Row	Unaware	Problem aware	Solution aware	Comparing (worked example)	Deciding	Using and renewing
Goal				Pick an SEO tool they can defend to a founder
Actions				Searches "seo agency for startups", opens three vendor sites, reads a pricing page
Touchpoints				An AI Overview, a comparison page, a pricing page, a peer in a Slack group
Thinking and feeling				"Will this look like a sensible choice in three months?"
Friction				No published price; cannot tell what the free tier actually covers

Copy the empty grid, name your own stages across the top, and fill one column at a time. The rows are what each one captures and where its content comes from:

Row	What it captures	Where the content comes from
Goal	What the customer is trying to accomplish at this point	Interviews, support tickets
Actions	What they actually do, in their words	Interviews, session data, search queries
Touchpoints	Where it happens: your site, a review site, a colleague, a search result	Analytics, referral data, asking
Thinking and feeling	The question in their head, quoted if you have it	Interviews, sales call notes
Friction	What slows them down or sends them elsewhere	Drop-off data, support, lost-deal notes

The fifth row is the only one that produces work. The first four describe; friction prescribes. If a map has no friction row, or the row is filled with mild observations, the exercise has produced a poster. A useful friction cell names something specific enough that somebody could fix it this quarter.

Two optional rows are worth adding once the basic map holds: opportunity, one line per friction point saying what you would do about it, and owner, so the opportunity has a name against it.

What are the stages of a customer journey?

Use your customers' stages, not a borrowed set. The six above are a starting point to argue with, not a structure to adopt.

Stage	The customer's question	Common mistake
Unaware	Is this normal, or is something wrong?	Skipping it. It leaves little search trace, which is exactly why it gets dropped
Problem aware	What is this called and who solves it?	Pitching the product before they have a name for the problem
Solution aware	I know what this is called. How is it done, and could we do it ourselves?	Filing it as early-stage. This is where build-versus-buy is decided
Comparing	Which of these, and can I defend the choice?	Assuming they compare on features. They compare on risk
Deciding	What could go wrong, and who has to approve?	Treating it as one person's decision
Using and renewing	Was I right, and can I prove it internally?	Ending the map at the purchase

In B2B the fifth stage is rarely one person, and the map needs the moment where your contact has to explain the choice to somebody who has not read anything you published. We treat that handoff at length, with the content each role needs, in mapping content to the B2B buyer journey.

How do you check the map against reality?

Take each stage, write down the queries a person in that stage would type, then pull your own top queries by impressions and see which stages they actually land in. Google's documentation for the Search Console performance report covers the export. Any stage you are confident about that generates no search behaviour is either wrong or happening somewhere you cannot see.

Here is the check run on this site. These are the twelve queries that drew the most impressions in the 28 days to 26 October 2026, taken from the published table in our AI Overview coverage test, with the stage we assigned each one. The stage column is our classification, not a measurement: the rule we used was which question in the table above the query is asking.

Stage	Queries	Impressions
Unaware	none	0
Problem aware	none	0
Solution aware	dofollow external links, seo marketing dashboard, how to do an seo audit, how to make / prepare / create seo audit report, seo site audit report	1,093
Comparing	seo agency for startups, startup seo agency, best seo and geo services for startups	604
Deciding	seo audit cost, seo audit pricing	297
Using and renewing	none	0
Total		1,994

Two things fell out of that. First, 901 of those 1,994 impressions, 45 per cent, arrived in 28 days at comparing and deciding, which are the stages a map drawn from instinct tends to hand to a salesperson rather than to a page.

Second, the three zeroes are facts about our coverage, not about our buyers. The input was the queries this site already earns the most impressions for, so the check can only see stages you already rank for. A zero means "no pages here", not "no customers here". The unaware row reads zero on a site whose own stages table says that stage leaves little search trace anyway. The problem-aware row reads zero because every one of the twelve already uses the vocabulary: somebody typing "dofollow external links" is not asking what the thing is called. And the renewal row reads zero because renewal questions are mostly branded or asked inside the product. Across the twelve, every impression lands at solution aware or later. Early-stage queries do appear further down the list, earning an impression or two each against the 1,994 here, which is the same finding in a weaker form: our early-stage coverage exists and is close to invisible.

One further detail, from the same twelve queries. When we loaded them on 30 October 2026, all twelve carried an AI Overview, and none of the twelve did under Google's Web filter. That is one site, one topic area, one day, and it is not a measurement of the web. But on our own map it means the comparing and deciding touchpoints now open with a generated summary.

Worth separating from all that: this site recorded 536 distinct queries over the 28 days to 26 October 2026 at an average position of 57.15. We are present across the journey and visible almost nowhere in it. That is a ranking problem, not a mapping problem, and confusing the two is how teams redraw a journey when the real issue is position.

How do you build one without a research budget?

Four sources you already have, in the order they are worth the time. None requires commissioning anything.

• Lost-deal notes. The friction row writes itself from the reasons deals died. This is the highest-value source and the least used.
• Support tickets from the first 30 days. Everything a new customer struggles with is a gap between what you promised and what the product does.
• Search queries you already appear for. Each one is a customer stating a goal in their own words, dated and free, as the section above shows.
• Customer conversations. Ask what they tried before you and who else was in the room. Keep going until new interviews stop telling you anything new, which is the only stopping rule that is actually about your customers rather than about a number somebody published.

Build a draft from the first three in an afternoon, then use the conversations to correct it. A draft somebody can disagree with produces better interviews than an open-ended question.

What does a journey map actually change?

Three things, and if none of them happens the map was a workshop artifact.

It reorders what you build and publish, because the friction row is a prioritised list of real problems rather than a backlog of requests. It exposes stages you have nothing for, which the query check above will point at, as long as you read a zero as missing coverage rather than as missing customers. It gives everyone one shared account of the customer, which matters most when sales and marketing disagree about why deals stall.

For content specifically, each cell in the goal row should map to something you have published or should. Where a goal has no page, you have found a gap derived from a customer rather than from a keyword tool, which is the better kind. Turning those gaps into a sequence is covered in deciding what to publish next.

How often should you update it?

We review ours twice a year, plus whenever something in it is proven wrong. That cadence is a choice rather than a rule; the second half is the part that matters, because a map reviewed on a schedule and never contradicted is not being used.

Three events should trigger an unscheduled update: a change in who buys, a change in what the competitive set looks like, and a change in how customers find you. The third is worth checking rather than assuming. On this site it turned out to be live, as the section above shows, but twelve queries on one property is not a finding about anybody else. Run the same check on your own top queries before redrawing anything. What to do about a touchpoint that is now a generated summary is a separate job, covered in how to make a passage quotable.

What are the common mistakes?

Four, in rough order of how often they waste the effort.

Mapping your process instead of their journey. If the stages are named after your funnel, you have drawn a funnel. The test: does any stage describe something that happens when you are not involved?

Averaging personas. One map, one person, one scenario. Two maps is cheaper than one map that is wrong twice.

Stopping at purchase. In a subscription business the renewal is where the contract is won a second time, and that conversation needs its own evidence.

Never testing it. The map is a hypothesis about people. Pick the two cells you are least sure about and go and check them this month, rather than reviewing the whole thing next year. The query check above took under an hour and put a number on two stages.