Every software CFO right now has two priorities that directly contradict each other. Cut software costs. Increase AI usage. Fair use is how you hold both.
As companies scale, LLM usage scales with them. Even after negotiating unit costs down, you're looking at 40–50% margin at best. So how do you get back to 80%+ margins?
A company's cost problem is not a customer problem. So design a lever that only the 10% will feel.
You have 1,000 customers on a $50/month plan. Your AI feature costs $2/customer/month at median usage — manageable. But the P95 customer is using $25 in compute. The P99? $80. That top 1% — just 10 accounts — might cost you more than the bottom 500 combined.
You have two options: raise prices across all 1,000 customers to cover the tail, or set a fair use threshold at the 95th percentile and only trigger a conversation for the 50 customers above it. Option one risks churn from 950 customers who were never the problem. Option two protects your margin without touching anyone using the product normally.
I've used fair use to cut product COGS by 33% in six months.
Inspect. Pull your usage distribution. Don't rely on averages — look at P50, P75, P90, P95, P99. The shape tells you where the cliff is. Most companies find the P95 user consumes 10–40x the median.
Analyze. Find the P95 threshold. This is your fair use line. Anyone above it is costing you disproportionately. Anyone below it is using the product normally and should never feel a policy change.
Apply. Hard limit or soft limit depends on your product type. Self-serve products usually work better with a soft limit (throttle, then notify). Enterprise products work better with a contractual fair use clause that triggers a conversation at renewal or mid-term.
At Grafana Labs, the AI assistant product had one user archetype paying $0.25 in compute and another paying $25. Both were on the same plan. We couldn't raise the plan price — it would have been a 100x increase for the median user who was never the problem.
The solution: seat-based pricing with a fair use threshold at the 90th percentile of compute consumption. The cap was set high enough that 90% of users would never see it. The 10% above it got a conversation — not a surprise bill, but a structured discussion about their usage pattern and what the right tier looked like for them.
$1M ARR in the first 30 days. 15.2% free-to-paid conversion vs. ~9% industry average. 12,000 orgs adopted in month one.
30-minute call. I'll tell you whether fair use is the right lever for your cost structure and what the threshold should look like.
Book a free call →