A founder I know raised her plan from $19 to $29 a month. Eleven of her sixty customers left within a month, and she spent two weeks convinced she had broken something. Her revenue went from $1,140 to $1,421.
The math is one line: going from $19 to $29, you can lose up to 34% of your customers and still collect the same money. That number comes from division, not instinct.
How to find your break-even churn
Divide the old price by the new one. 19 ÷ 29 = 0.66. You need 66% of your customers to stay just to match the old revenue, which means 34% can leave and you are even. Below that churn rate, you are ahead.
Run the same division when you cut a price. Going from $29 down to $19 means you need 53% more customers to stand still, because 29 ÷ 19 = 1.53. Most price cuts never get there, which is why they quietly shrink a business.
Write both numbers down before you touch the pricing page. The arithmetic takes two minutes. The apology tour takes longer.
How to pick what you charge for
The number is the smaller decision. What you count is the bigger one. Per seat caps you at headcount. Per project caps you at projects. A usage metric tied to the value the customer receives grows when they get more value, and that is the only growth that does not require you to sell harder every month.
Pick a metric your customer already tracks, then price a unit of it. If they cannot explain the bill to their boss in one sentence, the metric is wrong.
If you would rather fill in a template than build the sheets yourself, The Pricing Calculator pairs an XLSX workbook with a PDF guide that walks the Van Westendorp survey, the value-metric test, and the three-tier rules: https://probiex.gumroad.com/l/pricing-calculator
FAQ
Do I need to survey customers first?
The Van Westendorp method needs responses from real prospects, so plan a short survey before you can read the price bands.
How long does this take?
One focused session per product, then a quarterly revisit to re-run the numbers.
Last updated October 2026 · Probiex.
Top comments (1)
Nice and simple, and running this before touching the pricing page is a good habit. Two refinements worth adding. First, run it on contribution margin as well as revenue: old price minus cost to serve, divided by new price minus cost to serve. If each customer has a real cost to serve, for example AI usage or support, the break-even moves and you can afford to lose more customers than the revenue version says.
Second, churn is rarely even. Check who leaves by segment and plan, because losing the accounts that would have expanded later costs more than the formula shows. For B2B annual contracts, also wait a full renewal cycle before judging, since most of the reaction shows up at renewal.
I work at Mondrio, an agentic pricing management platform for B2B tech, manufacturing and professional services companies.