Modeling a price change before you ship it

Raising a price feels like a one-way door. You send the email, some customers grumble, a few leave, and you spend the next quarter wondering whether the move helped or hurt. It doesn't have to be a guess. Before you change a single number in your billing system, you can model the revenue impact on one page and know roughly what you're walking into.

The core idea is simple. A price increase only helps if the extra dollars from the customers who stay outweigh the dollars you lose from the customers who leave. Modeling the change means putting real figures on both sides of that trade before you commit to it.

The four numbers you need before you touch the price

You don't need a forecasting tool for this. You need four inputs, and you almost certainly have all of them already:

• Current price: what you charge today, per unit or per customer per month.
• Current volume: how many units or customers pay that price right now.
• Proposed price: the number you are considering moving to.
• Expected volume loss: your best estimate of how many customers cancel or don't renew because of the increase.

The first three are facts. The fourth is a judgment call, and it is where the whole exercise lives or dies. Economists call the link between price and volume "elasticity" — how much demand moves when the price moves. You don't need the formal version. You need an honest estimate: if I raise this 20%, what fraction of customers walk?

New revenue is the new price times the volume that stays

Once you have those four numbers, the math is one line. New revenue equals your proposed price multiplied by the volume you retain — that's current volume minus the customers you expect to lose. Compare that figure to today's revenue and you have your answer.

The reason this trips people up is that the increase and the loss pull in opposite directions, and intuition is bad at netting them out. A 20% price bump paired with a 20% customer loss is not a wash. Twenty percent more from each remaining customer, collected on only eighty percent of the base, actually leaves you slightly behind — and that surprises almost everyone the first time they run it.

Compute the break-even loss first

Before you even estimate churn, compute the break-even: the most volume you can afford to lose before the increase turns net-negative on revenue. It is a clean formula. To hold revenue flat, the fraction of volume you keep has to equal your old price divided by your new price.

For any percentage increase, that fraction is 1 divided by (1 plus the increase). Raise the price 20% and you need to keep 1 / 1.20 = 83.3% of your volume — meaning you can lose up to 16.7% and still break even. Raise it 10% and you can lose about 9%. Raise it 50% and you can shed a full third of your customers and still hold the line. The bigger the increase, the more room you have to lose people.

This number is useful because you compute it before guessing at churn. If your break-even loss is 17% and you are confident real churn will land nowhere near that, the decision gets easy. If break-even is 8% and you think you might lose 6%, you are cutting it close and should tread carefully.

A worked example

Say you run a software product at $80 per customer per month, with 500 customers. That's $40,000 in monthly recurring revenue. You are considering moving to $96 — a 20% increase.

First, the break-even. 1 / 1.20 = 83.3%, so you can lose up to 16.7% of 500 customers — about 83 of them — before the increase stops helping. That is your ceiling, and you know it before making any churn assumption at all.

Now your estimate. You expect an increase this size to push roughly 8% of customers to cancel. Eight percent of 500 is 40, so you would retain 460. New revenue is $96 times 460, or $44,160 — up $4,160 a month, about 10%, and comfortably inside the 83-customer cushion. If instead you feared losing 20% (100 customers), you would retain 400, and $96 times 400 is $38,400 — below where you started. Same price, different churn assumption, opposite conclusion.

One footnote: this models revenue. On profit the case is usually stronger, because the customers who leave take their serving costs with them, while the ones who stay pay more without costing you more. Get the revenue picture clear first — it is the harder half.

Segments, grandfathering, and testing before you commit

A price change does not have to hit everyone at once, and it usually shouldn't. Existing customers who signed up under the old price are the most likely to resent a jump. New customers have no anchor and churn far less when the number is simply higher from day one.

That opens up gentler options. You can "grandfather" current customers — leave them at $80 and apply $96 only to new signups. Revenue from the base holds steady while the higher price compounds slowly as you grow. Or you split the difference: new customers pay $96 immediately, existing ones move to $88 after 60 days' notice. Each variant carries its own churn estimate, and you can model each one on the same one-page math.

Best of all, you can replace the guess with data. Apply the new price to a single cohort — new signups for one month, or a random slice of incoming traffic — and watch what actually happens to conversion and churn. A few weeks of real numbers from a live segment beats any elasticity estimate you would otherwise pull from the air. Model it on paper, test it on a cohort, then ship it to everyone once the measured churn comes in under your break-even.