Customer segmentation that changes what you charge, defensibly

4 days ago
3 min read

Updated: 2 days ago

Introduction

Segmentation is usually discussed as a marketing exercise: who to talk to and what to say. The version that affects the accounts is different and much less common, which is segmenting by what you should charge. Businesses that do it well earn materially more from the same customers; businesses that do it carelessly create a fairness problem that surfaces at the worst possible moment.

The line between the two is not subtle. A price difference is defensible when it reflects a difference in what is delivered or what it costs to deliver. It is indefensible when it reflects only who is likely to pay more. The first is normal commerce; the second is the thing that ends relationships when two customers compare invoices.

1. Customer segmentation that changes what you charge must rest on cost or scope

This is the whole test.

Response time, hours included, volume committed, terms of payment, complexity of the site. Each of these justifies a different price because each changes what you are actually providing. Willingness to pay alone does not.

2. Measure cost to serve per segment before pricing anything

Most firms have never done this.

Visits, support contacts, revisions, admin time, payment delay, rework. Segments that look similar on revenue frequently differ by a factor of two or three in servicing cost, and that difference is the legitimate basis for the price difference.

3. Build the segments from behaviour, not from size

Size is a poor proxy.

Two customers of the same turnover can behave completely differently: one orders monthly on account and never calls, the other orders erratically and requires attention on every job. Behaviour predicts cost; headcount and turnover do not.

4. Give each segment a name and a published difference

Transparency is the protection.

Standard, priority, contract, project. Each with stated inclusions, response times and terms. When the difference is published, a customer paying more can see what they are buying and a customer paying less can see what they are not.

5. Let customers choose their segment

The strongest defence available.

If anyone can move up or down by accepting the corresponding conditions, the pricing is a menu rather than a judgement about them. This single design choice removes almost all of the fairness risk.

6. Watch for customers in the wrong segment

The most common leak.

A customer on standard terms who behaves like a priority customer is being subsidised, and one on priority terms who never uses it is overpaying and will eventually notice. A quarterly review moving people to the right tier recovers real margin.

7. Price the segment, not the individual

Consistency is what makes it survivable.

Once someone within a segment is negotiated to a different number, the structure is gone and you are back to discretion. Holding the tier price and moving people between tiers keeps it intact.

8. Model the shift before you change anything

People move when prices do.

Some customers will move down a tier rather than pay more, and a segmentation that assumes everyone stays put will overstate the gain considerably. Estimate the downgrade rate you can absorb before you publish.

9. Review the segments annually against actual cost

They drift.

Cost to serve changes as your own operations and your customers' behaviour change, and a tier that was correctly priced two years ago may now be your loss-making one. An annual repeat of the cost-to-serve measurement keeps the structure honest.

Be aware that in some markets and jurisdictions there are rules about differential pricing for comparable customers, particularly in regulated and business-to-business supply. Confirm the position that applies to you before publishing a structure.

Conclusion

Base every price difference on cost to serve or scope, because those are the differences you can explain.

Measure cost to serve per segment before setting any price, build segments from behaviour rather than customer size, publish each tier with its inclusions and response times, let customers choose their own tier by accepting the conditions, review who is sitting in the wrong tier every quarter, hold the tier price rather than negotiating individuals, model how many customers will move down before you publish, and repeat the cost measurement annually.