Indie AI tool rolls back a price rise after fewer people buy

On 22 September 2026 the founder of LiveFaceSwap AI, a real-time webcam face-swap product, posted on Indie Hackers that raising monthly prices cut purchases. The original grid was Basic $12, Pro $29 and Max $89. Two tests, $19/$39/$99 and $29/$59/$119, were meant to lift average revenue per user and make Pro look like roughly twice the price for about four times the credits. Overall purchase conversion fell. Basic and Pro sold less. Max barely sold. The founder reverted to $12/$29/$89 and is now A/B testing $12 against $15 on Basic, with similar results so far. A quarterly offer of three months for the price of two still sold reasonably well. The post is a single-product, self-reported experiment. Sample sizes, confidence intervals and traffic quality were not published.
For a tiny product, a price grid is often treated as a margin lever. Founders raise the first paid tier, dress Pro as the obvious middle, and assume buyers will skip cheap if the value story is clear. That logic holds when the buyer already trusts the category and is shopping for more of a known result. This test points at a different bottleneck. On an unfamiliar tool the first charge is closer to a paid trial than to a committed subscription. People who will not risk $19 will still risk $12, then later prepay a quarter once the product has proved itself. Raising the door price before trust exists can shrink the whole funnel, not just shift mix toward Pro.
Analysis
Treat this as an early community signal, not a pricing law. If your offer is still hard to picture before first use, keep the entry SKU cheap enough to buy as a trial and put the ARPU lift on prepaid after they have used it. This week, write down your current first-paid price next to a 3-for-2 quarterly option and decide which one you will test, not both at once.
Source note
Pulse published by Collab365 Spaces, reviewed by Helen Jones on . Cite as "Indie AI tool rolls back a price rise after fewer people buy", Collab365 Spaces. 1 source referenced.