Model a price change before celebrating the revenue upside

Unity's per-install Runtime Fee triggered a backlash, an apology and a full reversal. A five-step model that tests churn, predictability and trust before a price change goes public.

Price increases look attractive in a spreadsheet. Multiply the new price by current volume and the extra revenue appears immediately.

The real outcome depends on how customers react. Unity, the game engine company used by many studios worldwide, learned this the hard way in 2023.

Unity’s Runtime Fee

In September 2023, Unity announced a new Runtime Fee. According to Game Developer, it would have charged developers per install once their projects passed certain milestones.

The reaction was severe. Game Developer reported a colossal backlash, with high-profile creators threatening to leave the engine over what they saw as a betrayal of trust.

On 22 September 2023, Unity published an open letter from Marc Whitten, who led its engine and editor teams. It began: “I am sorry.”

Whitten wrote that Unity should have spoken with more customers and incorporated more of their feedback before announcing the policy. The letter set out changes, including keeping Unity Personal free with no Runtime Fee and exempting games under $1 million in trailing 12-month revenue.

The reversal

The changes did not end the story. Game Developer reported that chief executive John Riccitiello left the company two weeks after the announcement, and that Whitten later left too.

On 12 September 2024, new chief executive Matt Bromberg cancelled the Runtime Fee for games customers. Unity returned to seat-based subscriptions, with an 8% increase for Unity Pro to $2,200 per seat a year and a 25% increase for Unity Enterprise.

Bromberg wrote that customers understood price increases were necessary, but that they “needn’t come in a novel and controversial new form”. Unity said it would consider price increases only on an annual basis from then on.

What went wrong

From the public record, three problems stand out. This is our analysis, not Unity’s own account.

The model was new and hard to predict. Developers could not easily forecast what they would owe, because the charge depended on installs rather than seats they controlled.

Customers were not consulted. Unity’s own apology said it should have spoken with more of them first.

Trust became the cost. Once customers feared future changes, the debate moved from the fee itself to whether Unity could be relied on.

Unity eventually raised prices anyway, through a familiar seat-based model. The increase was accepted far more easily than the novel fee.

Model the change before you announce it

You can avoid the worst mistakes with a simple model and a few conversations. Here is a five-step approach.

1. Calculate the break-even volume loss

Start with the arithmetic. If you raise price, how many customers can you lose before revenue falls?

Illustrative arithmetic: with 1,000 customers paying €100 a month, revenue is €100,000. Raise the price 10% to €110, and you break even at about 909 customers. Lose more than about 91, or 9%, and revenue falls.

If you think about profit rather than revenue, the room for loss changes with your margin. Ask finance to run both.

2. Segment by sensitivity

Customers do not react the same way. Group them by size, how long they have been with you, how much they use the product and how easily they could switch.

Estimate a likely loss for each group. A small share of price-sensitive customers can account for most churn.

3. Add the second-order effects

The spreadsheet rarely includes the costs that hurt Unity most. List them explicitly:

• Customer support and sales time spent explaining the change
• Referrals and reviews lost if customers feel misled
• Partners or resellers who must explain your price to their clients
• Leadership time and reputation if the change is reversed

Put a rough number on each, even if it is a range.

4. Test predictability with customers

Before announcing, show the new pricing to a small group of trusted customers. Ask one question above all: can you predict what you will pay next year?

If they cannot, the model may be more dangerous than the price level. Unity’s later decision to return to seats suggests predictability mattered a great deal.

5. Write the reversal plan

Decide in advance what result would make you pause or reverse the change. Write down who decides and how you would tell customers.

This sounds pessimistic. In practice it makes leaders more willing to listen to early warnings, because changing course is already part of the plan.

A worked example

Hypothetical: a Danish B2B analytics firm plans to switch from a flat fee to usage-based pricing. Its model shows a 12% revenue gain if nobody leaves.

Interviews with 15 customers reveal that most cannot predict their usage. The firm adds a spending cap, keeps a flat-fee option for a year and tests with new customers first.

The projected gain falls to about 6% in year one. The risk of a public backlash falls much further. The figures here are illustrative.

Common modelling mistakes

Using one churn figure for everyone. Average churn hides the segment that leaves fastest. Model each segment separately.

Ignoring timing. In B2B, contracts delay reactions. A change that looks safe after one quarter may show its real cost at renewal.

Forgetting new customers. A higher price can reduce sign-ups as well as increase cancellations. Include both in the model.

Treating the announcement as the end. How you explain the change shapes the reaction. Unity’s own apology pointed to how the policy was introduced, not only to the price.

What to tell the board

Present three numbers, not one: the revenue gain if nobody leaves, the break-even volume loss, and your realistic estimate after segment-by-segment losses.

Then present the risks that do not fit in a spreadsheet. Unity’s experience shows that trust can cost more than any single price point.