How Do You Raise Prices on Customers You Already Have?

How do you raise prices on customers you already have?

Slowly, with a long notice period, a clear reason, and a decision made in advance about who you will protect. The mechanics are not the hard part. The hard part is that most teams decide the number first and the communication second, when the order should be reversed.

A price increase on existing customers is the only growth lever that can go backwards. Every other lever either works or does nothing. This one can cost you revenue you already had, which is why it deserves a written plan rather than a brave meeting.

This is a playbook, in the order we would actually run it. It is aimed at B2B software companies raising list price on an existing base, not at repricing a new product.

Is price really the most powerful profit lever?

Less than the consulting version of the claim suggests. John Dawes of the Ehrenberg-Bass Institute took apart the familiar assertion, which he attributes to McKinsey, that a 1% price increase produces an 8% profit boost. His objection is simple: the claim assumes you can raise prices with no impact on unit sales.

Dawes walks through it with ice creams costing 50 cents to make and selling for a dollar. A 1% cost reduction gives $5 extra profit. A 1% volume increase gives $5. A 1% price increase with volumes held constant gives $10, which is the source of the famous claim, and which he notes is theoretical only.

His conclusion is worth sitting with. Price increases help low margin businesses most, and risk real losses for high margin ones once sales fall away. Software is a high margin business.

What does the research actually say about elasticity?

The best known meta analysis is Bijmolt, van Heerde and Pieters in the Journal of Marketing Research, volume 42, issue 2, May 2005. Across 1,851 price elasticities drawn from 81 studies, they found an average price elasticity of -2.62. A 1% price rise is associated with roughly a 2.6% fall in sales.

They also found that sales elasticities increased in magnitude over the preceding four decades, meaning buyers became more price sensitive over time, not less.

Be careful applying this directly. That literature is dominated by consumer goods with easy substitution, not B2B software with switching costs, contracts and procurement cycles. It tells you the direction of the risk. It does not give you your own number, and anyone who claims it does is selling something.

Step one: decide what the increase is for

Write one sentence naming the reason, and make it true. Real reasons include funding a capability the base has asked for, correcting a price set before the product matured, or absorbing a cost you genuinely absorbed. "We need the revenue" is a real reason too, but it is not one you can say out loud, which tells you it is the wrong moment.

The reason determines everything downstream: the notice period, who is exempt, and whether you pair the increase with something new. Teams that skip this step end up improvising justifications on customer calls, and customers can always tell.

If the honest reason is that your packaging is wrong rather than your price, fix the packaging instead. That is a different project, and we covered it in how to structure pricing tiers.

Step two: decide who you protect, before you decide the number

Pick your protected groups first, in writing. Common ones are customers inside their first year, customers on annual contracts until renewal, design partners and early believers, and any account where your own delivery has been poor recently. Everyone else is in scope.

Doing this first changes the number, because you now know how much of the base it actually applies to. Doing it second turns into case by case exceptions, and case by case exceptions become the policy within about three weeks.

Grandfathering forever is usually a mistake, but a defined protection window is not. Give it an end date at the outset so it is a decision rather than a debt.

Step three: choose the size and the shape

Decide the percentage, whether it applies to all tiers equally, and whether it lands at once or over two steps. Our default preference is one increase with a long runway rather than two small ones, because every announcement costs goodwill and two announcements cost it twice.

Consider shape as well as size. Raising the entry tier hardest protects revenue but hits the customers least able to absorb it and most able to leave. Raising the top tier hardest tests your value with the accounts you can least afford to annoy. There is no free choice, only a considered one.

If you are moving toward consumption based charging at the same time, do not do both in one announcement. That is two changes wearing one coat, and it is covered separately in moving to usage based pricing.

Step four: give more notice than feels necessary

Ninety days is our floor for an annual contract base, and more is better. Notice is not politeness. It is the mechanism that lets a customer budget for the change rather than discover it, and a customer who budgets for it does not churn over it.

Announce to your own team first, with the reason and the exemptions written down, at least two weeks before customers hear anything. Support and customer success will field the reaction, and they cannot defend a decision they learned about from a customer.

Send the notice from a named person, not from a billing system. This is the one email where the cost of sounding corporate is highest.

Step five: decide what you will measure, and for how long

Pick your measures before the announcement, because afterwards everyone will interpret whatever happened as confirmation of their prior view. Track gross revenue retention, logo churn, downgrade rate, and support volume, weekly for a quarter.

You need a baseline to compare against, and public benchmarks are useful for sanity rather than targets. Recurly's published churn benchmarks, updated with July 2026 data, put overall churn across industries at 3.60%, split into 2.34% voluntary and 1.25% involuntary. Their B2B software figure is 3.22%, against 4.25% for consumer ecommerce.

The involuntary share matters here. More than a third of that overall churn is failed payments rather than decisions, and a price change increases failed payments mechanically, because cards and purchase orders are authorised for the old amount.

What usually goes wrong?

Four things, in our experience. The increase is announced before the team is briefed. The exemptions are invented on calls. The billing system applies it to someone who was supposed to be protected. And nobody watches involuntary churn, so a payment failure spike gets read as customers rejecting the price.

That last one is the most expensive misreading available, because it can cause a company to reverse a price increase that was actually working. Separate the two numbers from day one.

The fifth, quieter failure is raising price without raising perceived value anywhere. You do not need a new feature. You do need something the customer can point at, even if it is faster support or a capability shipped last quarter that nobody announced properly.

How does this interact with expansion revenue?

A price increase competes with expansion for the same goodwill, and expansion is almost always the better trade. If an account is a candidate for a genuine upgrade in the next two quarters, a blunt list price rise can spend the relationship capital you needed for that conversation.

So segment before you send. Accounts with real expansion potential get the expansion conversation first and the increase later or not at all. Accounts that are static and underpriced are the actual target of this exercise.

That sequencing logic is the same one behind building an expansion motion, and the two plans should be written by the same person in the same week.

What does a good outcome look like?

Quiet. A small, visible bump in downgrades. A short spike in support volume that resolves inside a month. Involuntary churn that returns to baseline once payment details are updated. And no reversal, because the plan anticipated the reaction rather than reacting to it.

Judge it at two quarters, not two weeks. The first fortnight is reaction, and reaction is not evidence. The second quarter is behaviour.

We work with B2B teams on the site, pricing page and lifecycle communication that carry a change like this, which means we usually see the plan before the customers do. If you are weighing an increase and want to pressure test the sequence, we are happy to look at it. You can find us at phoenix.studio.