It is good to see these moves. On OpenAI Pro, we only bought three subscriptions (everyone in the company gets Perplexity). We would have paid more, up to $2,000 per month, but would have started with only one subscription. There is a lot of value in o1 and a lot more value coming with 03.
I tend to think of these kinds of pricing shifts as proxy indicators to market shifts. Unsurprising given how much is changing at the same time. Thanks for continuing to summarize both the trends and the thinking.
"The value prop becomes a no-brainer — if the tech works."
Really interested to see how the agent hype works out. Assuming it gets better than really basic/surface-level conversations like we see many companies deploying just to have something out there.
Such timely analysis. I've interviewed many CPOs and founders, and just documented a case study that perfectly validates your insight how customers find usage-based pricing “too complicated and hard to predict.”
One CPO’s B2B SaaS company launched their enterprise AI product initially with a model based on work delivered (€2 per successful outcome). Despite perfect alignment with value creation, they didn’t have much success selling it - until they switched to a simple subscription model without volume elements.
The CPO’s comments exactly aligned with what you wrote: “It was a very elegant and nice idea. But it was quite complex. And it wasn’t predictable. … Ultimately, enterprise customers valued predictability more.”
He also emphasised how a value proposition to replace almost fixed labour costs with variable costs doesn’t work well. To actually realise savings, enterprises would have to reorganise work or fire staff, which is not quick or easy to do at least here in Europe. The value prop probably resonates better with growth companies that can just avoid hiring instead of firing.
In your October article "From selling access to selling work” you suggested ways to keep outcome-based pricing more palatable to enterprise customers (annual draw-down, grace periods, rollovers). I suspect the CPO I talked to had not tried those approaches. I’ll send your article to him.
I wrote up the full story in the latest article of my Substack newsletter for B2B SaaS product leaders:
It is good to see these moves. On OpenAI Pro, we only bought three subscriptions (everyone in the company gets Perplexity). We would have paid more, up to $2,000 per month, but would have started with only one subscription. There is a lot of value in o1 and a lot more value coming with 03.
Great content Kyle, tks for sharing!
I tend to think of these kinds of pricing shifts as proxy indicators to market shifts. Unsurprising given how much is changing at the same time. Thanks for continuing to summarize both the trends and the thinking.
"The value prop becomes a no-brainer — if the tech works."
Really interested to see how the agent hype works out. Assuming it gets better than really basic/surface-level conversations like we see many companies deploying just to have something out there.
Agree, excited to see things play out. FWIW some of the vertical AI use cases go far beyond surface level
Thx
Such timely analysis. I've interviewed many CPOs and founders, and just documented a case study that perfectly validates your insight how customers find usage-based pricing “too complicated and hard to predict.”
One CPO’s B2B SaaS company launched their enterprise AI product initially with a model based on work delivered (€2 per successful outcome). Despite perfect alignment with value creation, they didn’t have much success selling it - until they switched to a simple subscription model without volume elements.
The CPO’s comments exactly aligned with what you wrote: “It was a very elegant and nice idea. But it was quite complex. And it wasn’t predictable. … Ultimately, enterprise customers valued predictability more.”
He also emphasised how a value proposition to replace almost fixed labour costs with variable costs doesn’t work well. To actually realise savings, enterprises would have to reorganise work or fire staff, which is not quick or easy to do at least here in Europe. The value prop probably resonates better with growth companies that can just avoid hiring instead of firing.
In your October article "From selling access to selling work” you suggested ways to keep outcome-based pricing more palatable to enterprise customers (annual draw-down, grace periods, rollovers). I suspect the CPO I talked to had not tried those approaches. I’ll send your article to him.
I wrote up the full story in the latest article of my Substack newsletter for B2B SaaS product leaders:
https://newsletter.antti.lk/p/value-based-pricing-ai
Thanks for continuing to provide valuable frameworks for the industry!