The B2B Customer Journey, Without the Diagram

A customer journey describes a buyer’s questions and interactions before choosing a supplier and, when the map covers the full relationship, after the purchase. In B2B, several people contribute to that choice, with their own questions and criteria. They may revisit a solution, and much of their research happens before suppliers are contacted.

TL;DR. A five-stage diagram cannot capture every return in a B2B purchase. In the cited samples, 6sense places first sales contact at 61% of the journey, while Forrester reports an average of thirteen people involved and stalls in 86% of buying processes. These averages do not describe every purchase in every market.

The digital maturity of customers and partners also matters. The figures below describe that context; our interpretation is that phone calls, referrals and meetings deserve attention alongside digital channels. A useful map identifies both the buyers’ tasks and the obstacles preventing progress.

What is a customer journey and where does the diagram come from?

One influential model is McKinsey’s Consumer Decision Journey, published in June 2009, which depicts the choice as a circuit instead of the classic funnel. The authors observed that people add and remove options along the way instead of steadily narrowing a list. The model includes the relationship after purchase.

The problem appears when that model, designed for consumers, gets copied onto B2B without adjustment.

Gartner describes the B2B buying process as six jobs the group has to complete: problem identification, solution exploration, requirements building, supplier selection, validation and consensus creation. The jobs do not necessarily follow a fixed order; buyers may return to some of them.

The difference between the two pictures is not academic. A linear diagram makes you believe you can move someone from one box to the next with an email. Reality asks you to be present and useful when the person comes back, for the third time, to a question you thought they had moved past.

What are the five stages of the customer journey?

One common model names five stages: awareness, consideration, decision, retention and advocacy. It offers useful vocabulary, including for what happens after a purchase, without prescribing the order of every case.

Models differ in how many stages they name. A working map needs to follow buyers’ actual questions, regardless of how many boxes appear in the diagram.

In a complex industrial purchase, several people may research in parallel and revisit questions others thought were settled.

In one possible example, the technical director compares solutions while the finance director has only begun to assess cost, and procurement raises a new criterion. The roles and timing need research for each company; averages from separate studies do not describe the same purchase.

Stages can orient a sales conversation. To decide what information to prepare, we follow buyers’ tasks and obstacles rather than assume everyone moves through the same stages in the same order.

What does the real route look like, in numbers?

The figures below come from different studies. They offer prompts for investigation, not a statistical description of every B2B purchase in a particular market.

The 6sense study from 2025, built on roughly 4,000 responses from North America, Asia-Pacific and Europe, also shows that the average buying cycle has shortened to 10.1 months, from 11.3 months in the previous edition.

An earlier 6sense analysis reported that 78% of surveyed buyers had fully or nearly fully set their requirements before their first sales conversation. This figure comes from a different edition than the 61% measure above.

These results suggest that a supplier may help shape criteria before the first conversation. Negotiation still matters, but it is not the only point at which the choice takes shape.

A company helping buyers define their criteria can gain an advantage before negotiation. We discuss that mechanism in our article on cognitive ownership.

Why can you not see most of the journey?

Some research happens beyond your own website. SparkToro’s analysis of a US panel for the first four months of 2026 estimates that 68.01% of Google searches sent no click to a website. The authors caution that this panel differs from the one used in 2024, so the two percentages are not a clean trend comparison.

A person may get an answer from the results page or a conversational assistant without visiting your website. That interaction does not appear in the website’s traffic statistics.

Added to that are private conversations, industry groups, questions put to a colleague at another company and recommendations asked for at an event. 6sense calls this area the “dark funnel”. The term describes exactly the anonymous research happening on territory you do not own.

Traffic statistics can be supplemented with information from buyer conversations and the sales process. The 61% figure in the 6sense study describes the portion of the journey before first sales contact, not how much of it analytics can see. Research before the conversation therefore deserves attention too.

What changes in a less digitised market?

I cannot cite a cross-country quantitative study of B2B buying behaviour segmented by digital maturity here. Digitisation figures do not directly measure how companies choose suppliers, but they describe the environment in which purchases happen.

The spread inside Europe is wide. Eurostat data for 2025 puts the European average for small and medium companies reaching at least the basic level of the digital intensity index at 71.4%, while the countries at the bottom of the range sit closer to 44%.

In the same period, AI use among companies averages 20% across the European Union while the laggards sit around 5%.

Advanced cloud purchasing averages 47% against roughly 22% at the bottom, and basic digital skills reach 60% of the population aged 16 to 74 on average, against roughly a third in the weakest markets.

In our interpretation, a buying map should include phone calls, direct recommendations and industry events, where interactions are harder to track than website visits. A decision-maker may search on Google and then check the information with other people. Eurostat does not measure the relative importance of these behaviours.

This informs an allocation rule we apply with our clients. The digital presence needs to be consistent because buyers may check it after hearing about the company. Alongside it, we investigate how offline channels contribute to validation and allocate the budget according to observed buyer behaviour.

Digital and offline channels need to provide consistent information, whichever order the buyer uses them in.

What has to exist at each buying job?

If the route is not linear, then material does not get produced by stage either, it gets produced by job. The table below connects the six jobs Gartner describes to what the person has to find, in practice, at each of them.

The last line deserves attention because a lack of consensus can block a purchase. Forrester’s figure of 86% of stalled processes shows the scale of the problem without attributing all those delays to this cause. A buyer you have convinced may need to explain the offer to a colleague who has never spoken to you and has little time to read.

Without material to support that internal discussion, the process may stop before the supplier hears why.

Requirements building also deserves attention. Helping buyers compare approaches using explicit criteria may bring you into the evaluation earlier. Check which criteria the group actually uses rather than assuming it will adopt yours.

How do you match marketing material to the sales stage?

By tying it to the concrete stage the conversation sits at, rather than to the ending you would like. In the methodology we apply, the picture is called the double funnel. On top of the marketing funnel, which carries a person from suspect to prospect and then to analysis, sits the commercial funnel, meaning negotiation, contract, retention, development and profit.

Marketing and sales have different responsibilities, so information about the buyer can be lost at the handover. Connect the two by recording what questions the buyer has already explored and what remains unresolved.

The practical rule is to give each piece of material a clear purpose: answering the question that prevents the next step.

When speaking with someone who influences the purchase, find out what they can decide and what they must discuss with others. Material may help clarify objections, obtain approval or build consensus, depending on their role.

One example is our published analysis of Cablero, a Romanian manufacturer of industrial cables. The communication we analysed relied on quality, good price and fast delivery. Those claims did not explain clearly enough what the company offered for the buyer’s application.

The strategic move proposed there was to shift the unit of value. Not the cable, but the economic risk protected by the cable, expressed as a communication territory, “we do not sell the cable, we sell what is not allowed to fail”.

During requirements building, a supplier may help the group compare the economic outcome protected, alongside the price per metre of cable. Whether that criterion matters in a specific purchase still needs checking. We cover the underlying idea separately in the article on the price war.

Why can stages not be skipped, even when the buyer looks ready?

Each stage answers questions needed for the next decision. Before sending a proposal, check whether those questions have been resolved, including through the buyer’s independent research. Sequencing means adapting the discussion to what the buyer already understands.

The finding that 78% of buyers establish their requirements before their first sales conversation shows why an early presence matters. It does not mean those requirements cannot be discussed or changed later.

Our working model has four overlapping stages. The first improves existing operational performance, using what the company has now. The second builds brand memory through architecture, a defining attribute and consistent language. The third provides recurring educational material, while the fourth is the cumulative efficiency that can emerge as the first three develop.

These stages overlap because buyers revisit questions. Expecting the fourth stage’s results in the second month can ignore the work needed to prepare them.

How much effort are you asking of the reader?

Every piece of material charges the reader a cognitive tax, meaning the effort needed to work out what you do, for whom, and what makes you different from the previous tab they had open. The tax is invisible in any dashboard, and it is paid in full at every touch.

When a buyer encounters the offer across several channels, each unclear message can require another effort to understand it. Consistent information helps them avoid reconstructing the explanation at each interaction.

A coherent message can reduce the effort of understanding the offer and may improve campaign performance. This is the mechanism we aim for, but any reduction in cost per click must be checked against the data; auctions, audiences and other factors also affect it.

In a page review, we check whether a visitor can understand the offer, who it serves and what evidence supports it. A clearer explanation may reduce the effort of understanding; any effect on decision time needs separate measurement.

For a little-known company, a page that is too brief may leave buyers’ questions unanswered. We check which explanations and evidence they need to understand the offer; the useful level of detail depends on the purchase.

A company nobody knows and which explains nothing has not achieved elegance, it has achieved silence.

Which layer of the company does the route get repaired on?

A broken journey is a symptom, and the repair belongs to whichever layer actually holds the fault. We work with four, and they stack.

L1, performance. Campaigns, channels, budgets and the numbers the reports show. This is where everyone starts, and it is the only layer that produces figures fast enough to feel like progress.

L2, the commercial process. The sales system, its stages, the material available at each stage and the handover between marketing and sales. Missing material or an unclear handover can contribute to stalled decisions even when campaigns work well.

L3, orchestration and memory. What the company knows about its own buyers, where that knowledge sits, and whether the next conversation starts from it or from zero. A company whose client context lives in one person’s head restarts every route from the beginning.

L4, positioning and category architecture. What the company is in the buyer’s head, against which alternatives it gets compared, and on which criteria. When the fault is here, no amount of campaign work reaches it.

The order matters, because a company optimising its L1 campaigns while the break sits at L4 pays more and more for the same result.

We wrote separately about the architecture connecting the layers. The client profile everything else rests on gets built by the rules in the article on buyer persona, avatar and ICP, and the connection from campaign to predictability sits in the one on marketing architecture.

How do you build a map that actually helps?

Not by drawing the stages, because everyone already knows the stages. A useful map gets built from the questions the person asks themselves, in the order life puts them, and from the place where they look for the answer.

First, collect buyers’ questions in their own words from conversations and emails. A repeated question is a signal to investigate; check how often it arises and where it falls in the decision before making it a point on the map.

Second, find out where buyers get each answer now: from a competitor, consultant, community or supplier conversation. A question without a clear answer may point to useful material to create.

Third, identify where the process stalls. Forrester’s 86% figure shows how common stalls were in its sample, without identifying the cause of each one. Check your own opportunities to see whether a lack of material for internal discussion is among the reasons.

If the review shows that a lack of consensus is stopping the purchase, material for the internal discussion may be more useful at that point than increasing campaign spend.

The fourth step is to connect the map to the sales system. For each buyer question or obstacle, identify the material needed, who is responsible for providing it and how follow-up will be tracked. The map then guides commercial work while allowing buyers to revisit questions or move between stages.

Four mistakes to check in the map

The first is the map drawn in a meeting room. Five people from the company gather and describe what they believe the client does. The result describes the internal process rather than the client’s route, and the difference becomes visible the moment somebody tries to use the map to decide what material to produce.

The second is confusing the stage with the channel. A company splits its map by channel, meaning website, ads, email and events, then finds it cannot answer a single useful question. A stage is a question the person asks themselves, while a channel is merely where they look for the answer, and the same answer has to exist in several places.

The third is optimising only the visible end. Form conversion can improve without telling you how buyers formed their criteria before arriving on the site. Measure the final stage, but investigate the questions preceding it too.

The fourth is the map that ends in nothing. The document gets made, presented and archived. A useful map produces a short list of things to write or build, with an owner for each, otherwise it stays an exercise.

Finally, check whether the map identifies missing material, who will prepare it and by when. Without those decisions, the document is difficult to use in everyday work.

Frequently asked questions

How does a customer journey differ from a sales funnel?

The funnel describes the process from your perspective, meaning the steps you move a lead through. The customer journey describes the same road from the person’s perspective, with their questions and their blockages. Companies that confuse the two end up optimising what they see while ignoring what the buyer experiences.

How many personas do I need to build the map?

Start with the buying decision and document the relevant roles. Forrester’s average of thirteen people does not mean you need thirteen maps; one map can show who enters the conversation and which questions they raise.

The distinction between the profile of the company and the profile of the individual is worth getting right first, and we set it out in the article on buyer persona, avatar and ICP.

Is the customer journey the same thing as customer experience?

No. The journey describes the buyer’s steps, questions and decisions, while customer experience describes how they perceive interactions with the company before and after a purchase. A journey map identifies obstacles, and experience research helps explain them.

How often does it get updated?

Review the map when interviews and sales conversations reveal new questions or obstacles. Set a periodic check that suits the market’s pace; regulatory or technology changes may call for an earlier review.

Is it worth the investment if we sell through relationships and referrals?

Yes, if you also want to understand how people decide when they arrive without a referral. The map may show which information they struggle to find and what role the referrer plays.

Where do I start if I have no data?

Start with notes from won and lost opportunities. They offer hypotheses about questions and obstacles; check the reasons through interviews, because sales notes can be incomplete.

If you want to see where the route breaks in your company’s case, meaning which question goes unanswered right before the decision, you can request a positioning analysis. It starts from what the market finds about you now. It does not start from what you believe it should find.