Why we went with prepaid credits instead of subscriptions (and what it broke)

Everyone told us SaaS = subscriptions. We went prepaid credits for our AI skills assessment tool (https://aisa.to) and im still not 100% sure it was the right call.

The logic: hiring is bursty. A recruiter runs 8 assessments in a week then nothing for a month. Charging a monthly seat fee for that felt wrong, and every pricing call turned into "but we wont use it every month". So we did 3 free credits to start, then buy packages, each assessment is 1-2 credits depending on depth. No per-seat, no monthly.

What works: the free credits get people to run a real assessment instead of just poking at a demo. That's where it sells itself, when they see an actual report on a real candidate.

What broke: revenue forecasting is miserable. Subscriptions give you a clean MRR line you can plan against. Credits give you lumpy usage and a constant "are they coming back or did they churn" question. People also hoard credits and get loss-averse about spending them, which weirdly slows adoption.

Still think it fits a bursty use case better than subscriptions, but the predictability cost is real. Anyone here cracked forecasting on a credit model?