How should i price per-api-call saas with big volume swings?

[ZoeyC]

I’m trying to figure out how to structure a usage-based billing system for my SaaS where clients pay per API call, but I’m stuck on how to handle the huge variance in monthly volume between our smallest and largest customers. The sheer unpredictability makes forecasting our own costs and revenue a real headache.

[Anthony.D]

I tried pure pay per call and the variance was nuts. smallest customers around a few thousand calls a month, biggest up to tens of millions. revenue and costs felt random. we added a base monthly fee plus tiered per call pricing and a cap for high-usage spikes; that helped flatten the curve a bit, but it didn’t fix the unpredictability completely.

[Elizabeth15]

I wonder if the real issue isn't pricing at all but forecasting itself; do you actually know the distribution you expect next month?

[NathanUT]

we rolled out a base fee plus volume bands and started tracking cost per call by cohort, plus a simple p90 forecast to plan capacity. results: steadier gross margin but the outliers still sting when a single customer spikes.

[Paisley51]

Sometimes the spikes come after a new feature or a big integration release; i catch myself wondering if the problem is really the data model or the billing assumptions. maybe the thing to test is prepay credits or a soft cap, but i keep circling back to whether the product usage is the bigger issue.