Why subscription app prices get stuck, and how to test a higher one
A Coca-Cola cost 5 cents in 1886. It still cost 5 cents in 1959.
Two world wars and the Great Depression, and the price barely moved. Two things held it there.
In 1899, two lawyers bought the bottling rights for one dollar, in a contract that fixed the syrup price with no end date. Then came vending machines that could not make change. One coin, one Coke. A 6-cent Coke needed two coins, and the next single coin was a dime, which would have doubled the price.
In 1953, the company asked the US Treasury to mint a 7.5-cent coin. The Treasury said no.
When your price needs a new coin from the government, the customer is not the problem.
Subscription apps have vending machines too
Plenty of subscription apps are priced by their own machinery:
- A launch price nobody revisited.
- An annual discount set in year one and never tested since.
- A paywall built around one number that every screen, email and ad now repeats.
- A team sure that users will riot if the price moves.
None of these is evidence about what users will pay. They are reasons the price is hard to change.
How to test a higher price safely
- List what is actually holding your price. Contracts, store pricing tiers, old promises in ads, internal fear. Name each one.
- Test on new users only. Existing subscribers keep their price. New users see the test price. That limits the risk to people who never knew the old one.
- Measure revenue per install, not just conversion. A higher price usually converts fewer people. The question is whether total revenue per new user goes up.
- Read it long enough. Include renewals or trial-to-paid, not only day-one purchases. See when to stop an A/B test.
- Test the plan mix too. The annual discount and the trial length are prices as well.
Bottom line
Your customers are not guarding your price. Your vending machine is.
Sources: Fixed price of Coca-Cola from 1886 to 1959, Slate.
FAQ
How do I test a price increase in a subscription app?
Show the higher price to new users only, keep existing subscribers on their current price, and compare revenue per new user over a long enough period to include trial conversions or renewals.
Will raising my app's price hurt conversion?
Usually conversion drops somewhat. What matters is whether revenue per new user rises enough to make up for it. Only a test on your own users can answer that.
Why don't startups raise prices?
Often because of internal fear or old commitments, not evidence. A launch price, an early annual discount or a paywall design can lock a price in place long after the market would accept more.
Should I test annual discounts and trial length?
Yes. The annual discount, the trial length and the plan mix all change what users effectively pay, and they are often left untested since launch.