AI pricing metrics showing up in multiple SaaS verticals
Steven Forth is a Managing Partner at Ibbaka. See his Skill Profile on Ibbaka Talio.
In a recent piece, Zuora CEO Tien Tzuo suggested we Mix it up: How to implement multiple monetization models. One way we are seeing people mix it up is by using multiple pricing metrics, or hybrid pricing. In the annual PeakSpan Ibbaka Survey on Net Revenue Retention we are seeing a growing diversity of pricing metrics. There are fewer companies with only one pricing metric and many more with two or three metrics. A few companies have four or five pricing metrics, which suggests a diversity of business models.
Another interesting trend we are seeing is that companies from several different verticals using metrics associated with Generative AI and other forms of AI.
These pricing metrics are:
Input Tokens
Output Tokens
Model Scale and Complexity
Model Speed
People familiar with generative AI pricing will be familiar with each of these metrics. Open.ai’s API pricing provides a good example of what the look line is in use.
In the 2023 survey, the pricing metrics associated with AI, and specifically generative AI, showed up mostly in companies that were explicitly calling themselves AI companies. This appears to be changing. Looking at the survey data for June 23 (N=311), companies from 16 of the 25 verticals tracked are using one or more of these AI pricing metrics.
These metrics were most common at the 15 General AI companies that have responded to the survey. Other verticals well represented include Data Analytics and Management (8), Digital Asset Management and Content Management (5) and Developer Tools (5). There was also a sprinkling of companies from Infrastructure Services (3), IIoT, and Digital Twins (3).
I take this to mean that these companies make a major use of generative AI. The Token Input and Token Output metrics seem to me to be more cost than value metrics, suggesting that these companies are concerned about the cost of delivering solutions that are dependent on generative AI.
Still this infiltration of AI pricing metrics into other verticals is a good thing. In general there seems to be a diversification of pricing metrics and less dependency on simple user based pricing. User-Based Pricing is still the most common pricing metric (196 of 311 which is more than 60%), but less than half of these use only user-based pricing. Twenty-two percent combine User-Based Pricing with Usage-Based Pricing.
There has been a lot of criticism of SaaS companies for overreliance on User-Based Pricing for SaaS AI. See Kyle Poyar and Palle Broe’s How AI Apps Make Money. There is some truth in this, but Kyle’s data is skewed by a reliance on companies that publish their pricing on their website, which is mostly Product Led Growth companies. In the PeakSpan Ibbaka Survey, 26% of companies lead with Product Led Growth while more than 50% lead with Sales Led Growth. The survey data shows more diversity in pricing models.
This is just a snapshot in time, the survey is still open. The full results will be announced just before SaaStr in early September. More data, specific to AI-enabled apps, will be gathered in the AI Monetization in 2025 survey.
