For most of my career, I’ve run technology businesses where customers pay for what they use, such as ad platforms, data tools, digital incentives businesses, and consumer research services. The model is intuitive: Revenue goes up when customers use more and down when they use less. Customers pay for the value they get. What could be simpler?
Except it never felt simple in the boardroom. It felt like a constant apology, with countless investor conversations eventually circling back to the same frustrated questions, all essentially boiling down to this: “Why can’t you be more like a subscription business?”
Subscription businesses have long been the darlings of Software-as-a-Service (SaaS). They were labeled “predictable” and “sticky,” with recurring revenue and tidy renewal rates, making forecasting profits easy.
Meanwhile, we were trying to justify why Q3 looked softer than Q2, why one big customer having a slow quarter wasn’t actually a sign of anything structural, why our revenue was recurring—just not in the even, easy way the subscription businesses received revenue.
And while subscription companies were being rewarded with high multiples for their elegant economics, we were working our asses off. There was a genuine inferiority complex for those of us responsible for stewarding these businesses. It often felt like we were just doing it wrong.
Consumption conundrums
Forecasting consumption revenue is genuinely hard. You’re modeling customer behavior and contracts. Monthly revenue depends on driving sustained use of your technology. A subscription customer who barely logs in still pays at least until the next renewal, but a consumption customer whose usage dips hurts your revenue this week.
This is where incentives for customer-facing teams can be tricky to get right. How do you actually drive product usage? Who owns that responsibility? Do you reward sales teams based on contracts signed, when that revenue is realized, or some combination of the two? What role should customer success teams play? And how do you support customers without building a large and costly service organization?
Subscription pricing gives buyers something genuinely valuable—budgeting certainty. You knew what you were paying for services in January and in December. For CFOs operating under real budget constraints, that predictability was the whole point.
Then AI happened, and suddenly, now everyone wants a consumption business. IDC predicts that by 2028, pure seat-based pricing will be obsolete, with 70% of software vendors shifting toward consumption, outcome-based, or hybrid pricing models. The irony!
How AI forces a shift in SaaS pricing models
According to Zylo’s 2026 SaaS Management Index, nearly every SaaS company is moving toward AI, with 92% having launched AI features or planning to do so.
But monetization is still catching up. Only 41% formally charge for AI features, while a third are still experimenting with monetization. Most are defaulting to subscription pricing (53%), with far fewer using usage-based (11%) or hybrid models (31%).
When you bolt a large language model onto your product, you inherit variable costs fundamentally incompatible with fixed-price subscriptions. AI features cost money every time customers use them, and those computing costs scale alongside usage.
A customer who uses your AI features heavily costs you meaningfully more than one who doesn’t. You can try to absorb that in a subscription price, but you’re either leaving money on the table or quietly rationing access. Neither works for long.
Meanwhile, pressure is mounting to eliminate the waste associated with unused SaaS seat licenses. Zylo reports that 47% of SaaS licenses went unused in 2025…and that was an improvement. It highlights the core tension with subscription models: Customers pay for access, not necessarily for value.
As scrutiny increases and budgets need to be freed up for AI investment in various forms, CFOs who once valued the predictability of subscription pricing will push to reduce wasted license fees—making seat-based models increasingly vulnerable to compression.
The case for hybrid pricing
Consumption pricing trades certainty for less waste and flexibility — but flexibility, as anyone who has ever opened a surprise Amazon Web Services bill can tell you, has a dark side.
So the market is converging on hybrid consumption models: a guaranteed base subscription that gives customers budget predictability, paired with usage-based pricing that lets vendors capture the value they actually deliver. This destination is a more negotiated middle ground, and getting there requires operational capabilities that most subscription-era SaaS companies simply haven’t built.
For those who have had the hard conversations justifying the virtues of a consumption-model business, there’s a silver lining. We developed skills that subscription businesses never had to learn. Little did we know these skills would become such an important part of the future of software.
Forecasting is a good example. Traditional SaaS businesses forecast revenue from sales pipeline metrics, including deal sizes and renewal rates. In a usage-based model, revenue only materializes when customers actually use the product.
So, forecasting revenue requires forecasting customer behavior itself. It relies on completely different signals than what sits in a traditional pipeline report. Historical usage patterns, customer business conditions, and ongoing conversations with users matter more than the contracts themselves. It becomes much more of a data science exercise than a traditional sales forecasting exercise.
None of it is easy, but all of it is learnable. It also takes time, and companies that spent the last decade perfecting their subscription playbook are starting from scratch.
For years, consumption businesses were the ones playing catch-up, always being asked why we didn’t look more like our superior subscription counterparts. That question has been reversed.
The industry is now asking how to build what we already have.
Aaron Kechley is the CEO of Zappi.