B2B Customer Segmentation for Pricing and Sales

Executive summary

B2B customer segmentation is the practice of grouping customers so that each group can be given a deliberate price, set of terms, and level of sales attention. The test of a segmentation is whether you would do something different for each group, and whether that difference shows up in margin.

Most mid-market companies already have segments, but they are the wrong kind. They are drawn on industry, company size, or the sales territory a customer happens to sit in. None of these tells you how the customer buys, how much service it consumes, or how much price it will tolerate. Two customers in the same industry with the same revenue can differ by ten points of contribution margin because one orders by the pallet on a standing schedule and the other places forty rush orders a month and calls for status on each.

This guide builds segments from behavior and economics. You will pull a customer-level data set, choose two or three behavioral dimensions, test the resulting segments against margin and cost to serve, and then attach three things to each: a price corridor, a standard set of terms, and a coverage model for the sales team. The output is a one-page segment charter per segment that sales, finance and customer service can all use.

Do this before a price increase or a discount reset. A price increase applied evenly across segments overcharges customers who have alternatives and undercharges those who do not.

When to use this

Use this guide when:

• One list price and one discount culture apply to every customer, and nobody can explain why a given customer gets the discount it has.
• Sales coverage is organized by territory or by rep preference, and your best accounts get the same attention as your worst.
• Gross margin looks acceptable in total but varies widely by customer, and you suspect a minority of customers carry the rest.
• You are about to change pricing, packaging, or terms and want to sequence the change by customer group.
• A new owner, sponsor, or board has asked "who are our best customers and why?" and the answer today is a list of the biggest ones.

Do not start here if you cannot produce invoice-level revenue by customer for the last 12 months. Fix the data first. If you can, but cost to serve is unknown, start with the simplified method in the framework section and refine later with a full customer profitability analysis.

Diagnostic questions

Answer these before building anything. They tell you how far the current segmentation is from usable.

1. Do we have a written list of customer segments, and does every active customer belong to exactly one?
2. On what basis were the segments drawn: industry, size, territory, history, or buying behavior?
3. For each segment, can we state in one sentence why those customers buy from us instead of a competitor?
4. What share of revenue comes from the top 20 customers, and what share of gross margin?
5. Does price realization (invoice price versus list price) differ materially across customers that look similar? Who approved the difference?
6. Which customers place many small orders, request rush delivery, or require special packaging, and is any of that charged for?
7. Do sales reps have a different call plan, response standard, or pricing authority by segment, or is it one process for everyone?
8. How many customers were won in the last 24 months, and which segment did they enter? Which segment lost the most customers?
9. Where do we have contractual commitments (rebates, price locks, most-favored terms) that limit what we can do in a segment?
10. Which customers could leave tomorrow with little consequence to them, and which would take real effort to replace us? If answers to 1 and 2 are weak, you do not have segments, you have labels. If 5 and 6 show unexplained differences, the segmentation work will pay for itself quickly. If customers lack a shared ID, resolve parent-child rollups first, because a customer counted four times under four names will distort every segment.

The framework: segment on behavior, test on economics

Step 1: Choose the unit of analysis

Segment the buying entity that makes the decision, not the billing account. A distributor with six ship-to locations that buys centrally is one customer for pricing purposes. A holding company whose subsidiaries negotiate independently is several. Decide this first and document it, because the choice changes every figure downstream.

Step 2: Pick two or three behavioral dimensions

Behavioral dimensions describe how the customer buys and what they value. Industry and size are descriptors, not drivers. Useful drivers include:

Choose the two or three that most plausibly explain differences in price tolerance and cost. More than three produces segments too small to manage.

Step 3: Form candidate segments and name them for behavior

Group customers that share the chosen dimensions. Aim for three to six segments. Name them for what the customer does, such as "repeat contractors on standing schedules" or "project buyers who need quotes," not for how you feel about them. Names like "A, B, C accounts" carry no information and invite arguments about status.

Step 4: Test each segment against economics

A segment is only useful if it differs from the others on at least one variable you can act on. For each candidate segment calculate:

Gross margin % = (Revenue − Cost of goods or direct service cost) ÷ Revenue
Cost to serve % = Customer-specific selling, service, freight and handling cost ÷ Revenue
Contribution margin % = Gross margin % − Cost to serve %
Price realization % = Average invoice price ÷ Average list price (same mix)

If two candidate segments have the same contribution margin, the same price realization, and would get the same coverage, merge them. If a segment's contribution margin varies as much inside the segment as between segments, the dimensions you picked are not the real drivers. Go back to Step 2.

Cost to serve does not need to be perfect on the first pass. Allocate order-processing, freight, and customer service cost by activity drivers you can already count (orders, lines, expedites, calls) and note where you used a rough allocation. The point is to see direction and order of magnitude.

Step 5: Attach a price corridor, terms, and a coverage model

This step turns a classification into an operating tool. For each segment decide:

• Price corridor. The range of realized price (as a percentage of list or of a reference price) a rep may offer without escalation, with a floor tied to margin. The corridor should be narrower for segments where price is the main reason to buy and wider where you are the specified or sole qualified supplier. Escalation above the corridor follows your discount approval matrix.
• Standard terms. Payment terms, minimum order quantities, delivery windows, rebate structures, and fees for non-standard service (rush, split shipments, custom packaging).
• Coverage model. Who calls on the customer, how often, with what agenda. This can range from a named account manager with a quarterly plan to inside-sales coverage with a self-service ordering portal.

Step-by-step process

1. Assemble the data set. Pull 12 months of invoice-level data: customer ID, parent ID, product or service, quantity, list price, invoice price, discounts and rebates, order date, ship date, freight, returns. Owner: FP&A. Output: one customer-level table, reconciled to total revenue in the general ledger.
2. Build the customer hierarchy. Roll ship-to and bill-to accounts to the buying entity. Owner: Finance with Sales Ops. Output: a parent-child mapping with no orphaned accounts.
3. Calculate economics per customer. Gross margin, estimated cost to serve, contribution margin, price realization, order frequency. Owner: FP&A. Output: a ranked customer table.
4. Interview sales and customer service. Ask each rep for their five best and five hardest customers and why. Owner: Commercial lead. Output: a short list of behavioral hypotheses, such as "customers who spec us in rarely push on price."
5. Draft three to six candidate segments. Use the behavioral dimensions chosen. Owner: Commercial lead with FP&A. Output: segment definitions and a customer-to-segment mapping.
6. Test the segments. Compare margin, cost to serve, and price realization by segment, and check the variance inside each. Owner: FP&A. Output: a segment scorecard.
7. Write the segment charters. One page each: definition, size, economics, price corridor, standard terms, coverage model, and the owner. Owner: Commercial lead. Output: charters approved by the CEO or CFO.
8. Reconcile to contracts. Identify customers whose current contracts conflict with the charter (price locks, rebate tiers, most-favored-customer clauses). Contract terms bind you. Have counsel read any clause you plan to work around. Owner: Commercial lead with counsel. Output: a conflict list with dates when each contract can be revisited.
9. Sequence the changes. Start with changes that need no customer conversation (coverage, internal approval limits), then new-quote pricing, then renewals, and only last in-flight contracts. Owner: CEO or CRO. Output: a dated rollout by segment.
10. Review quarterly. Re-run the scorecard, move customers between segments by the rules, and record what changed in price and margin. Owner: FP&A. Output: a one-page segment review.

Templates

Segment definition table (copy into a spreadsheet)

Segment charter (one page per segment)

Metrics and KPIs

Setting targets without invented benchmarks. Anchor each target on your own history. Use the contribution margin of your best segment as the long-run reference for the others, then set a first-year target as a share of the gap you can attribute to price corridor and cost to serve fixes you have identified. Where contracts set a floor or a cap, treat that as a hard bound. For public reference, use gross margin disclosures from named listed peers in your sector and read the footnotes on how they define the cost lines before comparing.

Common mistakes

1. Segmenting on industry and calling it done. Industry tells you what a customer does, not how it buys. Use it as a descriptor on the charter, not as the primary key.
2. Too many segments. Eight segments for a ten-person sales team means none gets a distinct plan. Merge until each has a different action.
3. Ignoring cost to serve. Two segments with equal gross margin can have very different contribution. A segment that looks attractive at the gross line may be unprofitable after rush orders and support.
4. Differentiating price without legal review. Price differences across customers can raise legal questions depending on your industry, jurisdiction, and the nature of the goods. Involve counsel before you formalize different pricing for similar customers buying similar volumes.
5. Building the model once and not wiring it into quoting. If reps still quote from the same price list with the same approval limits, the segments exist only in a deck. Connect the corridor to your quote process.

Worked example

Illustrative. These numbers are invented to show the method and are not benchmarks.

A $40M industrial distributor drew segments from behavior: how the customer buys and how much service it uses. After testing, it settled on four.

Reconciliation: the segment revenues sum to $40.0M, gross margin to $11.80M, cost to serve to $3.35M, and contribution to $8.45M, which is 21.1% of revenue.

What the segmentation shows:

• Segments C and D have the same 17.0% contribution for opposite reasons. C has a middling gross margin but high service cost (rush orders, many quote revisions). D has thin gross margin but low service cost. They need different fixes.
• For C, the lever is terms and fees: minimum order values and a charge for expedites. For D, the lever is the price corridor and rebate structure, because there is little cost to take out.
• Suppose, for illustration, the company could raise realized price by 2 points on segments C and D without losing volume, through a tighter corridor and fees. Revenue in those two segments is $17.0M, so the gain is $17.0M × 2% = $0.34M, all of which falls to contribution. Whether volume holds is exactly what you have to test with a pilot, not assume.
• Segment A has the highest contribution and the highest switching cost for customers. Coverage there should be a named account manager, not a price cut.

Coverage follows: A gets named account managers with quarterly plans, B gets inside-sales coverage with a standing-order review, C gets a quote desk with a minimum order policy, and D gets a channel manager and a rebate agreement reviewed annually.

AI prompt assistance

AI can speed up the data preparation and the first draft of segment definitions. It cannot know your customers or contracts. Supply the data yourself and check the output against the ledger and with the sales team.

1. Draft candidate segments from customer attributes.

I will paste a table of 200 customers with annual revenue, gross margin %, orders per year, average order size, share of orders expedited, and a one-line description from the account owner. Propose four to six segments based on buying behavior (not industry or size alone). For each segment, give a name, a rule that assigns a customer to it, the customer count, total revenue, and weighted contribution margin. Flag customers that do not fit any segment.

Check: recompute segment revenue and make sure it sums to the total you pasted. Ask two reps whether the segments match their experience.

2. Draft segment charters.

Using this segment scorecard and these standard terms, draft a one-page charter for the segment "Repeat contractors." Include who they are, why they buy, the price corridor expressed as a range of list price with a margin floor, the standard terms, the coverage model, and the review trigger. Use only the numbers I provide. Where a field needs information I did not provide, write "to be decided" rather than guessing.

Check: confirm every number against your scorecard and have Sales and Finance sign off.

Implementation guide

Governance: the CEO or CRO owns the segment definitions and decides when they change. The commercial lead owns the charters and coverage. Finance owns the economics and the data. Sales reps own the customer relationships and give input on movement. Report contribution margin and price realization by segment monthly in the management pack.

Downloadable assets

The segment definition table and the segment charter above are designed to be copied into a spreadsheet. No calculator for this guide exists on the site today. For the revenue-side view of price leakage, see the pricing waterfall analysis.

Related resources

• Customer profitability analysis: allocate cost to serve to individual customers and feed the segment scorecard.
• Price increase playbook: sequence price changes by segment once the corridors are set.
• Discount approval matrix and governance: the escalation path for quotes outside a segment corridor.
• Pricing waterfall analysis: see where realized price falls from list by customer group.
• CPQ and quote discipline: put the segment corridor into the quoting process.

Frequently asked questions

What is B2B customer segmentation?

B2B customer segmentation is grouping business customers so that each group receives a deliberate price, terms, and sales coverage. The most useful groups are drawn on how customers buy and what they cost to serve, not only on industry or size. A segment is real when it differs from the others in margin, price tolerance, or service need.

How many customer segments should a mid-market company have?

Most mid-market companies are best served by three to six segments. The test is whether each segment would receive a different price corridor, set of terms, or coverage model. If two segments would be treated the same way, merge them. If a segment is too small for a sales team to manage separately, fold it into the nearest one.

What data do I need to segment customers for pricing?

You need 12 months of invoice-level data by customer: quantities, list price, invoice price, discounts, rebates, order dates, freight and returns. Add a parent-child customer hierarchy and activity counts such as orders, expedites, and support contacts. Reconcile the total to general ledger revenue before analyzing anything.

Is it legal to charge different customers different prices?

It depends on your industry, jurisdiction, the goods or services involved, and the contracts in place. Price differences are common in business-to-business markets, but some laws restrict certain differences between similar buyers. Treat this as a question for your own counsel before formalizing segment-based pricing, and follow your agreements.

How often should customer segments be reviewed?

Review the segment roster quarterly. Re-run the scorecard, check which customers no longer match their segment rule, and move them under the movement triggers you defined. Review the segment definitions themselves annually, or sooner after a major change in product mix, channel, or competitor behavior.

How this works in a System of Performance

A segmentation only changes margin if the corridor and terms reach the quote, the order, and the invoice, and if someone can see what happened afterward. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. If you want to see how your own customer base splits on margin and cost to serve, bring it to us as a launch partner.