पाठशाला Pathshala · हिसाब Hisāb, Unit economics · Lesson 24 · Scale
Pricing power: raising prices without losing the base
A price rise is a rollout, not an announcement. Decide who moves and when, protect the customers who would leave, and read the retention data that says whether it worked.
Pathshala, The Founder Library · 11 October 2026 · 7 min read
Most companies under-price for years and then raise prices in a single email. The first mistake costs margin quietly. The second can cost the customers who took the longest to win, and it is avoidable, because a price rise can be planned, segmented, phased and measured like any other launch.
This lesson is for a company with a paying base it cares about. It explains why price is the strongest lever in the P&L, how to tell whether you have pricing power before you use it, how to segment the rise and grandfather the customers who should wait, and which retention numbers say whether it worked. How to test a price on new customers is covered in [pricing experiments](/library/pricing-experiments-without-burning-customers); this lesson is about the base you already have.
Why price is the strongest lever
Michael Marn, Eric Roegner and Craig Zawada found in The Power of Pricing that for the average company in the S&P 1500, a price rise of 1 per cent with volumes steady would lift operating profit by 8 per cent, nearly 50 per cent more than the effect of a 1 per cent cut in variable cost. The arithmetic is general. A rise in price adds to revenue and leaves every cost where it was, so it arrives in profit whole. A company with a 10 per cent operating margin that raises prices by 5 per cent and keeps its volume has raised operating profit by half.
The catch is in the phrase with volumes steady. The rise works only if the customers who matter stay. Pricing power is exactly that: the ability to raise price without losing them. Some companies have more of it than they use. Some have none and find out from the churn report.
Evidence that you have it
Look for it in data you already hold, before any announcement. Win rates at higher quotes. If sales has quoted above list to some prospects, compare win rates; if they barely move, list is too low. Discount frequency. If most deals close at list, buyers are not resisting price; if most need a discount, they are. Usage since signing. Accounts that use the product two or three times as much as when they signed are getting more value at the same price, and are the natural first candidates. Churn reasons. Read the last fifty cancellations; if price is rarely the stated reason, there is room. Replacement cost. Ask what the customer would have to do without you: hire someone, buy two tools, go back to spreadsheets. Price power lives in the gap between your price and that cost.
Weak evidence on all five is a finding too. It means the product needs more value or a better-fitted [pricing metric](/library/usage-based-seat-based-hybrid-pricing) before it needs a higher price.
Segment before you move anyone
A single rise for every customer treats the customer who gets ten times the value the same as the one barely hanging on. Split the base first, by the value each segment gets and by how likely it is to leave: plan, size, usage growth, tenure, and whether the account is on an annual contract. Then decide a treatment for each. High-value, high-usage accounts move first and furthest. Mid-tier accounts move later and less. The most price-sensitive segment, often the smallest customers on the oldest plan, may be left on the old price for good, because the revenue at stake is small and the churn risk is large.
A fence lets a company charge different segments differently without cutting the price for anyone. Netflix’s mobile plan for India, launched in July 2019 at ₹199 a month, offered the full catalogue in standard definition on one phone or tablet at a time. The fence was the device and the resolution: a price for viewers who mainly watch on a phone, without lowering what anyone watching on a television paid. A B2B company builds the same kind of fence with plan limits, support levels, user caps or annual commitments, so that a customer who will not pay more has a smaller plan to step down to instead of the exit.
Grandfathering, and the order of moves
Grandfathering means existing customers keep their price for a stated period after a new one is adopted. It earns three things. Time for each segment to see the added value before the bill rises. A natural renewal date on which to move annual customers. And a cushion that lets the company watch the first wave before committing the rest.
Order the moves. New customers pay the new price from the day it is adopted. Existing customers are told on that day, with the date their price will change and the reason. The first segment moves after the grandfathering period. The company reads its retention data for a quarter. Only then do the later segments move, at the same rise or a smaller one, depending on what the first wave showed.
Take a Mumbai company selling practice-management software to dental clinics, with ₹50 lakh of monthly recurring revenue, ₹4 lakh of new revenue each month and 2 per cent revenue churn a month. It raises prices by 20 per cent. New clinics pay the new price at once, and it expects 5 per cent fewer of them to sign. Existing clinics are grandfathered for six months; then 70 per cent of the base, by revenue, moves, and the smallest single-chair clinics stay where they are. It expects 8 per cent of the clinics that move to leave.
The figure opens on the Mumbai plan. By month 24 the rise adds about ₹13 lakh of monthly revenue, and roughly ₹11 lakh of that comes from new customers paying the new price; the migration of existing clinics adds about ₹2 lakh. Push the share of moved clinics who leave to 17 per cent and the migration adds nothing at all, while new-customer pricing still does its work. For a growing company most of the value of a rise comes from customers who have not yet signed, and most of the risk sits in the ones who have. Grandfather generously, move the base slowly, and never let the migration endanger the new price.
The new price earns most of its money from customers who have not signed yet. Move the existing base slowly, by segment, and only as fast as the retention data allows.
The retention data that says it worked
Netflix gave the standard warning. In July 2011 it split streaming and DVDs into separate plans at $7.99 each, and in the following quarter its US subscribers fell by 800,000 to 23.8 million; its shareholder letter said many long-term members felt shocked by the pricing changes. Revenue on the day of a rise always looks good. Whether it worked shows up over the following quarters, in a few specific numbers.
Compare the moved cohort with a control. The segment left on the old price, or the segment moved later, is the comparison. Track for each: revenue churn, logo churn, downgrades to a smaller plan, support tickets mentioning price, and payment failures, monthly for at least two quarters after the move. Compute net dollar retention for the moved cohort. Freshworks publishes the method: take the annual recurring revenue of a group of customers twelve months ago, take the same customers’ revenue today including expansion and net of contraction and attrition, and divide the second by the first. Run that on the moved cohort and the control, and the gap is the rise’s effect on the base. Watch the leading signs. Downgrades and a drop in usage come before cancellations; a moved segment whose usage falls in the first month after the change is a segment about to churn.
Write down in advance what success means: for example, revenue churn in the moved cohort no more than one point a month above the control over two quarters, and net dollar retention above the control by at least the size of the rise less that churn. Then the review is a check against a number, not an argument.
The ninety-day rise review
Ninety days after each segment moves, the founder, the head of sales and whoever owns retention meet with one page. The page shows, for the moved cohort and the control, monthly revenue churn, logo churn, downgrades, net dollar retention to date and usage per account, plus the list of every account that left or downgraded and the reason given.
Decide one of three things. The rise held: schedule the next segment. The rise held but a sub-segment broke: give that sub-segment a step-down plan or a longer grandfathering period, and move the rest. The rise did not hold: stop further moves, offer the lost accounts’ peers a way to stay, and revisit the value before the price. Record the decision and the numbers, because the next rise, in eighteen months, should start from them.
Nothing here is legal, tax or investment advice. The Mumbai company is illustrative; check consumer-protection rules and your customer contracts on notice periods before any price change.
Sources
- Michael V. Marn, Eric V. Roegner and Craig C. Zawada, The Power of Pricing, McKinsey Quarterly, February 2003 — A 1 per cent price rise at steady volume lifts operating profit 8 per cent for the average S&P 1500 company, nearly 50 per cent more than a 1 per cent cut in variable cost.
- Netflix, Netflix launches mobile plan for India, July 2019 — ₹199 a month, standard definition, one phone or tablet at a time.
- TV Technology, Netflix reports loss of 800,000 domestic subscribers in Q3, October 2011 — US subscribers down 800,000 to 23.8 million after separate $7.99 streaming and DVD plans.
- Freshworks Inc., Freshworks Reports Fourth Quarter and Full Year 2025 Results, February 2026 — Net dollar retention method: ending ARR over entering ARR for the same customers.