B2B Customer Segmentation: 6 Methods, the Columns Each One Needs, and a Worked Example

What is B2B customer segmentation?

B2B customer segmentation is the practice of splitting the companies you sell to, or could sell to, into groups that share traits which change how you target them, what you say to them and how much sales effort each one deserves. The short definition sits in our glossary entry on customer segmentation. This guide is about doing it on a real list.

Every guide on the topic lists types of segmentation: four, six, sometimes twelve. Most of them also say, somewhere near the end, that it all depends on "good data". That sentence is where the real work hides. In practice, a B2B segment is a filter on columns. "Software companies with 50 to 500 employees in France that are hiring salespeople" is four columns and four values. If the headcount column is empty for half your accounts, the segment does not exist yet: you are sorting complete records from incomplete ones, not companies that fit from companies that do not.

So this guide treats each method as what it needs to run: which columns, where each column comes from, and how much of the base it must cover before the segment means anything. You will find the six methods in one table, the steps to segment a base, a worked example of 1,000 accounts cut into four segments with their sizes, a small calculator to size your own segment, and the test most guides leave out: can you actually pull the list?

B2B customer segmentation methods at a glance

The table puts each method next to the columns it needs. The last column sorts those columns into three kinds, because each kind is filled differently. Enrichable columns can be filled for the whole base at once from public company data. In your CRM columns already exist if your team logs deals and usage. Declared columns only come from the customer: interviews, calls, surveys.

Read the table from the right. A team with a thin CRM and no research budget can still run the first three methods on its whole market, because those columns are enrichable. Needs-based segmentation is the richest and the slowest: it gives the best message, but the column fills one conversation at a time. Value-based segmentation and journey stages only apply to accounts you already know. That is the first practical choice in B2B customer segmentation: segment on what you can fill for everyone, and add declared columns to the segments that earn the effort.

The 6 B2B segmentation methods, with the columns behind each

Each method below gets a one-line definition, the columns it reads, a segment written as a filter, and the limit to keep in mind.

Firmographic segmentation, with examples

Firmographic segmentation groups companies by what they are: industry, size, location, revenue and age. It is the method every guide starts with, because the data is public and stable, and because it maps directly to how sales territories and pricing tiers are drawn. The columns are industry, headcount band, country or region and, where available, a revenue band. The full list of attributes you could use, ranked, is in our guide to the 30 essential firmographic attributes; for segmentation you rarely need more than five or six of them.

Three firmographic segmentation examples, written as filters:

• Industry = logistics and transport; headcount = 200 to 1,000; country = Germany, Netherlands, Belgium.
• Industry = software or IT services; headcount = 11 to 50; founded after 2018.
• Industry = hospitals and clinics; headcount over 500; region = Northeast US.

The limit: firmographics say who could buy, not who needs to buy now. Two companies with the same industry, size and country can be at opposite ends of their buying cycle. That is why firmographic segmentation is usually the first cut, refined by one of the methods below.

Technographic segmentation

Technographic segmentation groups companies by the tools they run. If your product plugs into a given CRM, or replaces a given platform, the technology column is often a sharper filter than the industry. Columns: CRM, marketing automation, e-commerce platform, analytics or hosting, read from what is detected on the company website. Example segment: online retailers on a specific e-commerce platform, 20 to 200 employees, with no marketing automation tool detected. How that data is collected, and its blind spots, is covered in our guide to technographic data and tech stacks. The limit: detection sees what runs on the public website, not back-office software.

Behavioral and intent segmentation

Behavioral segmentation groups companies by what they are doing, rather than what they are. Two kinds of columns feed it. Public signals: open roles by function, headcount growth, a new office, a funding round. Your own signals: visits to your pricing page, email engagement, product sign-ups. Example: software companies of 50 to 500 people with at least two open sales roles this month. The seven signals you can see on an account that has never visited you are listed in our guide to B2B buying signals. The limit: signals expire. A hiring column filled in March tells you little in September.

Needs-based segmentation

Needs-based segmentation groups companies by the problem they are trying to solve and what they value in a supplier. The classic needs segments in B2B are buyers who choose on price, on quality and brand, on service, or on a long-term partnership. It produces the best messaging, because it speaks to the reason for buying. The columns are declared: main need, buying criterion, maturity. You fill them from sales calls, interviews and surveys, which is why needs-based segmentation usually runs on customers and late-stage deals, then gets mapped back to firmographic proxies so it can be applied to the wider market.

Value-based segmentation and tiering

Value-based segmentation groups accounts by what they are worth: revenue to date, margin, potential and cost to serve. On customers, it decides who gets an account manager and who gets a self-serve path. On prospects, it becomes tiering: tier 1 accounts get a tailored approach, tier 2 a lighter one, tier 3 a programmatic one. Revenue and products bought live in your CRM; potential is usually estimated from firmographics (headcount, number of sites, teams that would use the product). How tiers feed an account-based program is detailed in our guide to account-based marketing.

Journey-stage segmentation

Journey-stage segmentation groups accounts by where they sit in the relationship: target, open opportunity, new customer, established customer, renewal, at risk. The column is a stage in your CRM, and the segment decides the message more than the target: a renewal account needs proof of value, a new customer needs onboarding, a target needs a reason to talk. It only applies to accounts your team has already touched.

A seventh cut often appears in guides: segmenting by role in the buying committee. That one splits people, not companies, and it belongs to the persona work, covered in our guide to the B2B buyer persona.

B2B market segmentation vs customer segmentation

The two terms get mixed up. B2B market segmentation cuts a market you want to win; B2B customer segmentation cuts the accounts you already have or already work. The methods overlap, but the starting list and the available columns do not.

In practice, the market path goes from the total addressable market to a target market, then to a target account list split into tiers. The customer path starts from your CRM and adds the enrichable columns your account records lack. The steps in part 06 work for both; only step 1 changes.

How B2B customer segmentation differs from B2C

Two consequences follow. Because the base is small, every account left unclassified is a real loss: an empty column on 300 accounts out of 1,000 is 30% of your market missing from your segments. And because the buying group is plural, a segment of companies always needs a second step to reach the people inside them.

How to segment B2B customers in 6 steps

Step 1: start from your ideal customer profile

Segmentation cuts a list; the ideal customer profile tells you which part of it to care about first. If you do not have one yet, build it from the deals you have won, as explained in our guide to the ICP in sales. For market segmentation, the starting list is then every company that could match; for customer segmentation, it is your CRM export.

Step 2: pick two or three criteria, not eight

Each criterion multiplies the number of segments and divides their size. Three criteria with three values each already make 27 combinations, most of them too small to act on. Pick the criteria that change what you would do: a different offer, a different message, a different channel. If two values of a criterion would get the same treatment, merge them.

Step 3: fill the columns, and test the fill rate

Before cutting anything, count how many rows have each column filled. The rule we work with: segment only on a column filled for the large majority of the base, roughly 80% or more. Between 50% and 80%, fill the column first. Below 50%, the column does not segment your base: it splits records someone completed from records nobody touched, and that split often follows the source of the lead rather than the company.

Firmographic, technographic and public behavioral columns are enrichable, so the gap can be closed for the whole list in one pass. CRM columns are filled by fixing the logging habit, and declared columns by deciding which segments deserve interviews. Whatever cannot be filled goes into an explicit "unclassified" group, never silently into the biggest segment.

Step 4: cut and count

Apply the filters and write down the number of accounts in each segment. A segment with no size is a wish. The count tells you whether a segment can carry a campaign, a dedicated sales rep or only a nurture track.

Step 5: run the four tests

Check each segment against the four tests in part 08. Merge or drop the ones that fail.

Step 6: activate, then re-check

Give each segment its offer, message and owner, then put a date on the next refresh. Part 09 explains why segments age and how often to re-fill the columns.

A worked example: 1,000 accounts, 4 segments

The example below is illustrative: a fictional base, round numbers, no real customer behind it. It shows the arithmetic of B2B customer segmentation on a list a small sales team could own. The company sells a sales tool to companies of 20 to 2,000 employees in Western Europe, and starts from a CRM export of 1,000 accounts: company name and website, with industry and headcount typed in by hand for some of them.

Before filling anything, the fill rates read: industry 62%, headcount 55%, country 91%, hiring signal 0%. Under the 80% rule, only country can segment the base. Cutting on industry and headcount now would hide 450 accounts with no headcount in the "unclassified" pile.

After enrichment, from the company names: industry 96%, headcount 95%, country 99%, hiring signal 100% (every account checked, hiring or not). The team cuts on three criteria: industry, headcount band, hiring sales roles or not.

The team activates A first. It is the smallest of the four, but it is the only one where the timing column says "now", and 84 accounts is a month of work for two reps. B is the same profile without the signal: it moves into A as soon as a company opens a sales role, which is why the hiring column is re-checked every month. C needs its own message before anyone calls. D is real revenue, but not for a sales team.

The cost of filling the columns, counted by steps: from a company name, finding the LinkedIn page with Search Companies and then enriching it with Enrich Companies is 2 credits per account; the hiring check with Company Hiring Signal adds 1. That is 3 credits per account: for 1,000 accounts, it fits in the Mini plan (9 euros a month, 4,000 credits). Adding a technology column with Website Technologies, a paid-plan feature at 2 credits per website, brings it to 5 credits per account, inside the Standard plan (20 euros a month, 10,000 credits). If you want to keep the segments in a spreadsheet, our free target account list template for Google Sheets has the columns ready.

To size a segment on your own base, enter the number of accounts, then for each criterion the share of rows where the column is filled and, among those, the share that match your filter.

The 4 tests of a usable segment

Classic market research gives three tests for a segment. A fourth one decides whether B2B customer segmentation ever leaves the slide deck.

1. Big enough. The segment can carry the effort you plan for it: enough accounts for a campaign, enough revenue potential for a dedicated rep. Our 84-account segment A carries two reps for a month; a segment of 9 accounts carries a list, not a program.
2. Distinct. It would get a different offer, message or channel from its neighbors. If segments B and C would receive the same email, they are one segment.
3. Unambiguous. Every account lands in exactly one segment by a rule anyone can apply. "Companies with a modern sales culture" fails; "hiring at least one sales role this month" passes.
4. Reachable. You can pull the list of accounts, then the people to contact in them. A segment defined on a column you cannot fill, or on companies you cannot find, does not exist outside the deck.

The fourth test is the one to run first, because it is the cheapest: before you write a single message, try to export the segment.

Why segments expire, and when to re-check them

A segment is a snapshot of columns, and the columns move. Companies grow out of a headcount band, get acquired, open or close offices, change their tools. The people inside them change jobs. Signals expire fastest: a hiring column says something about this month, little about next quarter.

Three failure modes come up again and again. Over-segmentation: so many segments that none is big enough to learn from. Static segments: cut once at the start of the year and never refreshed, so accounts sit in a segment they left months ago. Segments never activated: well defined, sized, approved, and never given an owner or a campaign. The re-check dates in the table, and an owner per segment, prevent most of the three.

From segment to list: pull the accounts, then the people

A segment becomes useful the day it turns into a list someone works. That takes two moves.

First, the accounts. For market segmentation, you need the companies that match the filter, not only the ones already in your CRM. Three ways to get them with Derrick, all on the free and paid plans, from 1 credit per company: describe the segment in a sentence with Import Companies from a Prompt (1 credit per company with your LinkedIn account connected, 10 without); start from your best customer and get look-alikes with Find Similar Companies; or, for French companies, pull every company under an activity code with Import Companies by NAF Code (no LinkedIn account needed). Then fill the segment columns: industry, headcount and headquarters with Enrich Companies, the timing column with Company Hiring Signal, and, on a paid plan, the technology column with Website Technologies.

Then, the people. A company does not answer emails. Once the account list holds, find the roles you sell to inside each account, using the personas you defined for the buying group. The full method, from accounts to contacts you can reach, is in our guide on how to build a prospect list.

Where to run it depends on how you work. The Derrick web app is the default: import a CSV or build the list, add a column per action, export the segment. If your accounts already live in a spreadsheet, the same actions run from the Google Sheets sidebar. To explore a segment in a conversation ("find me logistics companies of 200 to 1,000 people in Benelux that are hiring"), use Derrick in Claude or any MCP client. To refresh segments automatically in your CRM every month, call the REST API; API and MCP access start with the Plus plan. The actions that read LinkedIn (Search Companies, Enrich Companies, and the 1-credit rate of the prompt import) need your LinkedIn account connected through the Derrick Chrome extension, on every surface. The full list of actions and their costs is on the features page.