Raising prices is the most underused lever in an early-stage company, because most of them are underpriced. Not slightly: structurally. The price was set when you were desperate for logos, it has not moved since, and meanwhile the product does twice what it did. Every month at the old price is a discount you never decided to give.
Founders put off a price increase because they imagine one conversation: the angry customer leaving. The real distribution is much friendlier, but only if you run the raise as a system instead of a confession.
First, know your floor
Before touching prices, compute the variable cost of a customer: infrastructure per account, support time at your real hourly cost, payment fees, the works. Any customer below that line is not revenue, they are expense with extra steps. This number turns pricing from a feelings question into an arithmetic one, and it tells you which accounts you can afford to lose in the worst case, which quietly changes how you negotiate.
Sequence the raise
Do not raise everyone at once. Order the risk out of it:
1. New customers first, today. New prospects have no anchor. Raise the list price now, learn what the market bears, and let the new price collect evidence while existing customers stay untouched. If new deals keep closing, you have proof the price is fair, and a fact to cite later.
2. Expansion moments second. When an existing customer adds seats or upgrades, the new price applies naturally. Nobody experiences this as "a price increase."
3. The legacy base last, with notice. Sixty days minimum. One honest email, no drama.
The price increase email that works
Three sentences carry the whole thing:
From 1 October, [product] moves from €29 to €39 a month. Since you joined we have added [two or three real things they use]. As a thank-you for being early, you can lock the current price for another 12 months by switching to annual billing before then.
Why it works: a real date removes ambiguity, the named improvements answer "why" before it is asked, and the annual-lock option converts your most price-sensitive customers into prepaid cash instead of churn. You have replaced "pay more or leave" with a genuine choice, and most people take one of the two branches you designed.
Expect this distribution
Across most B2B raises in the 20% to 40% range, the pattern is consistent: most customers say nothing and pay, a meaningful group takes the annual lock, a handful negotiates, and a small tail leaves, weighted toward the accounts below your floor anyway. Model it before you send: even if 8% churn, a 30% raise on the rest usually nets strongly positive within the quarter. Run your own numbers with your own worst case, starting from the churn calculator; if the math holds at your most pessimistic churn guess, the fear is not information.
The ones who push back
Have one fallback ready, and only one: the annual lock, or a lighter plan if you have it. What you must not do is negotiate bespoke prices account by account. Every custom deal is a future migration problem and a fairness bomb waiting for a screenshot. "I can hold your current price for a year on annual billing" is generous, finite, and identical for everyone who asks.
Then watch the number that matters
After the raise, track cash revenue and churn weekly for a quarter, against the model you wrote down before sending the email. Prices are a decision like a hire: commit it, watch what it does to the model, adjust with evidence. The raise you run this calmly will fund the next two quarters, and the customers you like will mostly still be there, paying you closer to what the product is worth.