Going about pricing strategies for AI products
So, been running a AI startup for more than a year now and the one thing i always thought or was puzzled was when naturally customers think AI should be cheaper than human hours even though AI's throughtout or the benefits is 100x.
So we have been doing it the 'cost+' pricing way, but to be very frank this is working and is helping us be in control of the margins, but also exploring outcome based pricing in certain use cases. Will have this thread open to have discussions around this topic, also if i find newer ways or use cases where outcome based pricing works, happy to explore.
13 views
Replies
Cost plus has one quiet advantage: it forces you to know your real unit cost, and that number often moves more than you expect. In my song app the music engine charges $0.045 per song when I send it lyrics I wrote with another model, and $0.30 when it writes the lyrics itself. Same output for the user, almost 7x the cost. Knowing that changed the pipeline before it changed the price. I sell credits per song, with one dollar value per credit as the single anchor, so the user sees an outcome (a song) while I keep the cost plus math underneath. Outcome pricing on top, cost plus underneath, has felt like the honest middle.
i think the "should be cheaper than a human" expectation is about the anchor, not the cost. my customers are one or two person businesses, so the human they compare against is a marketer most of them cant afford. against that the plan price looks cheap even when its way above cost plus.
so we do both. the plan price sits well above what it costs us, and under it theres a floor that assumes the customer burns every single token, plus payment fees, plus a margin. no plan and no discount is allowed below that floor. inside the app people only see tokens, one unit, so they never have to think about model costs.
the same full use math killed my own launch first month price. it lost money on the top two plans when every token got used, so it was gone the same day.
outcome pricing sounds great until u realize u eat the cost of every attempt that didnt produce the outcome. id only do it where the outcome is easy to measure and each attempt is cheap.