How do I raise prices on existing customers without losing them?

1. Blog
2. Conversion & Pricing

By Saroj Jha · November 2, 2026 · 13 min read

You raise prices on existing customers without losing many of them by treating the announcement, not the effective date, as the risky moment. The best field evidence I can find shows churn rising as soon as customers are notified, even among customers whose bills would go down^[1], and Netflix’s own letters show that news of ending grandfathering moved churn more than the price change itself^[2]. So give a time-limited grandfathering window with a clear end date, give notice that is long enough for your contract cycle, explain the increase in terms of what the customer would face if they left, and staff the first weeks after the announcement as a retention campaign. Size the step on your own numbers first, so that the churn you expect leaves you well ahead even if it comes in worse.

What follows covers timing, grandfathering, notice, the message and the churn budget, with the rollout mechanics left to the AAJ pricing playbook.

Running case: Example Co. Fictional

Example Co. is a fictional Seed-stage startup selling compliance-tracking software to independent dental practices. It has 200 paying accounts at $300 a month, a price it has not changed in three years, during which it added a patient-forms module and an audit-report module. Monthly churn runs at 1.5%. The founders want to move every account to $360 a month, a 20% increase, and they are worried about losing the practices that have been with them longest.

When is it the right time to raise prices on existing customers?

In my view, the right time is when you can point to something that has changed for the customer since they signed: more product, more usage, or a market where the alternatives cost more. Large software companies that raise prices on their installed base tend to do it rarely and to tie it to years of accumulated value. Microsoft called its 2021 increase “the first substantive pricing update since we launched Office 365 a decade ago” and pointed to 24 apps and more than 1,400 features added since launch^[3]. Salesforce’s 2023 increase of an average 9% on list prices came with the note that its last list price increase was seven years earlier, with 22 releases and its generative AI features delivered in between^[4].

The wrong time, in my view, is in the middle of a product problem, an outage or an open support backlog, because the price notice becomes the reason to act on a complaint the customer already had. The SaaS Pricing and Packaging playbook covers how to tie each increase to new value and review prices on a cadence, so I will not repeat it here.

Example Co., step 1 Fictional

Example Co. has two modules its first customers never paid for and has not raised prices in three years. That is the case for an increase. It also has a known bug in its audit report that a dozen practices have complained about, so the founders fix it before announcing anything.

Do this

List what each customer cohort has received since its price was set, and fix any open product or support problem before the notice goes out.

Should you grandfather existing customers, and for how long?

Grandfathering means letting existing customers keep their old price for a period while new customers pay the new one. It protects goodwill, but every month of it costs real revenue, and permanent grandfathering leaves you running two price lists for years. My rule of thumb, not a benchmark: grandfather for a defined window with an end date that customers see on day one, and let tenure set the length if you want to reward loyalty.

Netflix is the clearest public case of how this plays out. In April 2016 it told shareholders that more than half of its US members paid $7.99 or $8.99 for a $9.99 plan, that it would phase the grandfathering out over the rest of 2016 “with our longest-tenured members getting the longest benefit”, and that it would not change anyone’s price without their acknowledgement^[5]. Three months later it reported that churn among members actually moved off the old price “is modest and conforms to our expectations”, but that churn had ticked up more widely, coincident with press coverage of the plan, because “some members perceived the news as an impending new price increase rather than the completion of two years of grandfathering”^[2].

The lesson I take is that the end of grandfathering is a second announcement, and it needs the same care as the first. Customers forget the end date you gave them two years ago, so the reminder reads as a new increase unless you make the history plain.

Example Co., step 2 Fictional

At $60 more per account per month, permanent grandfathering of all 200 accounts would forgo $144,000 a year (200 × $60 × 12). A six-month window forgoes $72,000 (200 × $60 × 6). Example Co. chooses six months for most accounts and nine for its 30 longest-tenured practices, and puts the end date in the first email.

Do this

Put the grandfathering end date in the first notice, then send a reminder before it ends that restates when the original change was announced.

How much notice should you give?

There is no single legal standard for B2B notice, and your own order form decides more than any norm, so check its price-change and renewal clauses with counsel first. The AAJ pricing playbook covers the contract and consumer-law checks; here I focus on what companies actually do. Microsoft gave six months: it announced on 19 August 2021 that prices would change on 1 March 2022^[3]. Atlassian’s current cloud price change took effect on 13 October 2026; it honored open quotes created before that date, its annual customers get a renewal quote 90 days before their subscription expires, and monthly customers could lock in existing pricing by requesting an annual quote, with that quote not extended beyond 11 November 2026^[6].

The research adds a reason to think about notice differently. In a study of a large telecom operator, a notification about a future price change raised customers’ attention and led to more searching, switching and churn, and churn rose right after the notification rather than when the new prices took effect^[1]. The notice period is therefore not a quiet waiting room. It is the window in which most of the churn decision happens.

My rule of thumb, not a benchmark: give monthly customers at least 60 days, and give annual customers notice before their renewal quote goes out, which in practice means 90 days or more before renewal. Offer a way to lock in the current price by committing for a year, as Atlassian did^[6], because it turns the notice into a reason to renew rather than to shop.

Example Co., step 3 Fictional

Example Co. has 120 practices on annual plans and 80 on monthly. Monthly customers get 90 days’ notice and the option to switch to an annual plan at $300 a month before the change. Annual customers keep $300 until their renewal, with at least 90 days’ notice before it.

Do this

Map every account’s renewal date before you pick the announcement date, so that no annual customer hears about the increase after their renewal quote.

What should the price increase notice say?

The usual advice is to justify the increase with rising costs or new features. The most rigorous test of that advice I know of points elsewhere. In a randomized field experiment with 1,626 customers of a multi-site Canadian storage provider, across 10 cohorts, the researchers compared cost, market and quality justifications for a price increase, and found that the market justification produced the lowest customer attrition, “in marked contrast to prior practitioner recommendations and academic research”^[7]. They trace the result to customers’ perceptions of switching costs: a message about the market reminded customers what leaving would involve^[7].

That study is in self-storage, not software, so I would not copy its wording. What I take from it is that the reason you give should help the customer compare staying with leaving, honestly. For a B2B product, that usually means stating what the customer would pay or rebuild elsewhere, alongside what they have gained since they signed. A market justification must be true; implying scarcity that does not exist is the kind of claim that costs you trust when a customer checks.

The notice also needs the customer’s own numbers: the old price, the new price, the date, the grandfathering end date and the options. The playbook has a fuller checklist of what goes in the email, which I will not repeat.

Example Co., step 4 Fictional

Example Co.’s notice tells each practice its current bill, its new bill and the date, lists the two modules added since it signed, and states what a practice would need to replace the product: the two other tools that cover forms and audit reports, plus the staff time to move three years of records. Every claim in it is checked against current competitor price pages.

Do this

Draft the notice with one paragraph on what the customer would face if they switched, and check every comparison in it against a current public source.

How much churn should you plan for?

Start with break-even, then stress-test it. If an account’s bill rises by a share p, you can lose p ÷ (1 + p) of revenue and still collect the same; the playbook has a table of these lines. Then ask how much extra churn the notice itself might cause. The telecom study found monthly churn propensity rose 60%, from 1% to 1.6%, among customers whose costs would actually fall under the new prices^[1]. I use that size of effect as a stress test, not a forecast: if churn rose that much among customers who would save money, it can rise at least that much among those who will pay more.

Netflix shows how far a small retention miss can travel in a large base. In the quarter it began ending grandfathered prices, US net additions came in at 0.16 million against a forecast of 0.50 million, and international net additions at 1.5 million against 2.0 million, while gross additions were on target^[2]. The company noted that with a large base, “slight variances in retention versus forecast can result in significant swings in net adds”^[2].

Example Co., step 5 Fictional

At $360, Example Co. can lose up to 16.7% of revenue and break even (0.20 ÷ 1.20), which is about 33 of its 200 accounts (200 − $60,000 ÷ $360). At 1.5% monthly churn, it would lose about 8.9 accounts over a three-month notice window anyway. If churn rises 60% to 2.4% a month, it loses about 14.1, or 5.2 more. With those losses, monthly revenue once the new price applies is about $66,939, against $57,340 without the increase, 16.7% higher. It would take about 32 extra lost accounts to wipe out the gain.

Do this

Write a stop rule before you announce: the number of notified accounts giving notice in the first four weeks at which you pause and call the rest.

How do you know if the price increase worked?

Judge it on the accounts that were notified, over a period long enough to include their first renewal at the new price. Netflix’s distinction is the useful one: it separated churn among members actually moved to the new price from churn across the wider base after the news broke^[2]. You want the same two numbers: churn among the accounts whose price changed, and churn among everyone else in the same period.

Watch downgrades as closely as cancellations, because a customer who drops a tier to keep their old bill shows up in revenue, not in logo churn. Net revenue retention can rise after an increase while gross retention falls, which is why AAJ’s article on NRR and GRR is worth reading before you report the result to your board.

Example Co., step 6 Fictional

Example Co. tracks its 200 notified accounts weekly for six months and its annual accounts through their first renewal at $360. It reports three numbers each month: cancellations, downgrades and revenue per account, for notified and not-yet-notified accounts side by side.

Do this

Set up a cohort view of notified accounts before the announcement, with cancellations, downgrades and revenue per account, and review it weekly until the last renewal at the new price.

What changes by stage?

• Pre-seed — With a handful of design partners, talk to each one before changing anything. In my view a founding-customer price with a written end date is better than an increase nobody expected.
• Seed — You can usually reprice every account by hand. Model each account’s new bill, give the longest-tenured customers the longest window, and have the founder call the largest accounts before the email goes out.
• Series A–B — Annual contracts and procurement teams dominate. Give notice before renewal quotes go out, offer an annual lock-in, brief sales and customer success first, and report notified-cohort churn separately to the board.

Free tool

Find the acceptable range before you pick the step

The Price Sensitivity Analyzer runs the Van Westendorp questions on your own customer survey, so you can see where existing customers’ acceptable price range sits before you choose the size of the step.

For the rollout mechanics, the break-even table and the contract checks, read the SaaS Pricing and Packaging Playbook, especially its section on raising prices with new value. The pricing mistakes that quietly kill startup growth explains why underpricing makes the first increase so hard, and the Pricing & ARPU Modeler lets you run the new price list on your own accounts.

If you want me to look at your own funnel, positioning and channels, my Growth Audit ends with three prioritized fixes and a 30-day action plan.