Raising prices on customers you already have: how a small SaaS decides, writes the email and keeps them

Most small software companies set their price once, in a hurry, before the first customer paid, and then leave it alone for years while the product doubles in size. By the time the founder decides the price is too low, the problem is no longer the new price. It is the two hundred customers already paying the old one, and the fear that the email announcing the change will be the email that makes them leave.

Large vendors raise prices on existing customers all the time, and their notices are public. Read side by side, they show a small number of choices that every rise comes down to: who pays the new price and from when, how much notice they get, what they can do to soften it, and what the email says. This article goes through those choices with a worked example, and ends with how to tell, a quarter later, whether it worked.

What three public rises looked like

Slack, 2022. Slack's first increase since its launch in 2014 raised the Pro plan from $8 to $8.75 per user per month on monthly billing, and from $6.67 to $7.25 on annual billing, from 1 September 2022. It was announced in mid-July, about six weeks ahead, and came with a way out: customers who renewed or switched to annual billing before the date kept the old rate for another year. Slack gave its reason as the value added since launch, naming features such as Slack Connect and huddles (ITPro). Only one plan changed; Business+ and Enterprise did not.

Atlassian, 2025. Atlassian told customers on 19 August 2025 that cloud list prices would rise on 15 October: 5% for the Standard edition of Jira, Confluence and Jira Service Management, 7.5% for Premium, 7.5% to 10% for Enterprise depending on users, 10% for Bitbucket. Its note to investors adds the line that matters for anyone with annual contracts: the effect would "feather in over time as cloud subscriptions come up for renewal" (Atlassian, FY26 cloud pricing notice). Customers paid the new price at their next renewal, not in the middle of a term they had already bought.

Unity, 2023 to 2024. In September 2023 Unity announced a new kind of charge, a fee per game install, on top of its seat prices, and applied it to games already built with its engine. The reaction from developers was strong enough that it was revised within weeks, and a year later the new chief executive cancelled it outright. What replaced it was an ordinary rise: Unity Pro up 8% to $2,200 a seat a year and Enterprise up 25%, from 1 January 2025 for customers who buy, upgrade or renew from that date, with a promise to consider price increases "only on an annual basis" (Unity, Unity is canceling the Runtime Fee). His explanation was one sentence: the relationship "must be a partnership built on trust".

Customers accept a higher price far more easily than a new way of being charged.

The pattern across the three: a stated percentage, a date weeks or months ahead, the new price applied at renewal or with a way to lock in the old one, and a reason in terms of what the customer got. The one that went wrong changed what customers were charged for, not just how much.

Decide who pays the new price, and from when

There are three common answers, and a small company can mix them by plan.

- Everyone, from a date. Simplest to run and to explain. Give monthly customers at least one full billing cycle of notice, and give annual customers the new price at renewal, as Atlassian did. Charging an annual customer more in the middle of a term they have paid for is the fastest way to a refund request.
- Grandfathering. Existing customers keep the old price for a fixed time (six or twelve months) or for good. It keeps almost everyone, and it costs you the difference for as long as it lasts. "For good" sounds generous and becomes a problem: in three years you will have four price generations to support, and the oldest customers will be the ones you can least afford to serve.
- Lock-in for commitment. Existing customers can keep the old price by moving to annual billing or renewing early before the date, as Slack offered. You get cash up front and a year of retention from the customers who take it; the rest pay the new price.

Whichever you pick, say it as a date and a number. "Prices are changing soon" invites every customer to imagine the worst.

Work out how many customers you can afford to lose

The arithmetic that makes a rise less frightening is the break-even churn: how many customers could leave before the rise loses money. An example with round numbers:

Seventeen per cent is far more extra churn than a 20% rise on a product people use every week usually causes, which is why most rises raise revenue even when they cost customers. The number to watch is not the total but who leaves: if the ten who go are your smallest, least-engaged accounts, the rise also cut support load. If they are three of your largest, the arithmetic above is wrong for your business and you should have called them first (below).

Run the numbers per plan. A rise on the cheapest plan moves the customers most likely to leave over price; a rise on the top plan touches the customers most likely to have a contract that limits it.

Check what you are allowed to do

Before writing anything, read your own terms. Many B2B terms of service say how much notice a price change needs (30 days is common) and whether it can apply mid-term; enterprise contracts often cap renewal increases at a percentage. Your promise to those customers is whatever you signed, not what this article suggests.

If you sell to consumers on automatic renewal, the law may set the notice for you. California's automatic renewal law, as amended for contracts entered into or changed from 1 July 2025, requires a clear and conspicuous notice of a price change no less than 7 and no more than 30 days before it takes effect, with instructions for cancelling (Kilpatrick Townsend on AB 2863). Other states and countries have their own rules; when in doubt, ask a lawyer about the places most of your paying customers live.

Call the large accounts before the email goes out

In a small SaaS, the top ten customers are often a third or more of revenue. Do not let them learn about the rise from a bulk email. A short call or a personal note a week before the announcement, from the founder, does three things: it tells you which of them will push back, it gives you room to offer something (a year at the old price for an annual commitment, a seat discount) before they have made up their minds, and it signals that they matter. Note what each one says; it is the best evidence you will get about whether the price is right.

Also flag anyone on an old annual contract, anyone who has opened a support ticket about billing in the last quarter, and anyone who is mid-way through a renewal discussion. These customers need a person, not a template.

What the email says

The email has one job: to tell the customer exactly what changes for them, and when, without making them read it twice. A plain structure works:

What makes it work: the subject line says what the email is about; the first two lines give the old price, the new price, the date and the effect on this customer's next invoice; the reason is two or three concrete things the reader uses, not "rising costs" or "continued investment"; there is one way to soften it; and the way out is stated plainly. Hiding the cancel link does not keep customers. It makes them angry and, under rules like California's, can make the notice non-compliant.

Send it from a person, at least one full billing cycle ahead for monthly plans, and send a short reminder a week before the date. Update the pricing page the same day the email goes out, so new visitors and existing customers see the same number.

Measure it a quarter later

Three months after the date, compare the customers who got the notice with the quarter before it:

- Cancellations in the 90 days after the notice against your normal 90-day churn. The rise caused the difference, not the total.
- Downgrades to a cheaper plan, which hide in revenue numbers rather than customer counts.
- Who left, by size and by how much they used the product. Losing low-use accounts is a cost you can live with; losing high-use ones means the price crossed a line.
- Monthly recurring revenue on the affected plans, against the break-even number you worked out before.
- Replies to the email: count them and read them. They tell you what customers think they are paying for.

Write the results down next to the original numbers. The next rise, in a year or two, will be easier to decide with a record of what this one did.

When a price rise is the wrong fix

A rise works when customers get more than they pay and the price simply lagged. It does not fix a product that people use once a month, and it cannot repair a pricing unit that no longer matches the value: if customers are paying per seat for something that now does the work of three people, the question is what you charge for, which we covered in per seat, per use or per result. And the wider market matters: buyers are cutting seats and watching software spend closely, as described in SaaS in 2026: seats are shrinking and pricing is moving. A clear, modest rise with notice survives that. A surprise does not.

Sources

ITPro, Slack to increase subscription prices for the first time (Pro plan prices before and after, 1 September 2022, renew-early lock-in, reason given); Atlassian, Notice of FY26 cloud pricing changes (percentages by edition, notified 19 August 2025, effective 15 October 2025, applied at renewal); Unity, A message to our community: Unity is canceling the Runtime Fee (12 September 2024; Pro +8% to $2,200, Enterprise +25% from 1 January 2025; annual price reviews); Kilpatrick Townsend, California's latest automatic renewal law amendments (7 to 30 days' notice of a price change for consumer auto-renewals from 1 July 2025). Pages read on 15 October 2026. The customer numbers, prices and email in the example are our own illustration, not data from any company; this is not legal advice.