A free tier can 10x your top-of-funnel or bury you in unpaid support tickets. The decision is not "should we do freemium" - it depends on three specific factors.
The three factors that determine fit
1. Paid ACV above $30/month
Free tier math only works if a small percentage of free users converting produces enough revenue to justify supporting the rest.
At $30/month ACV with 3% conversion: 100 free users produce $90/month in paid revenue. That covers the marginal cost of supporting the other 97 free users at most SaaS support economics.
At $10/month ACV with 3% conversion: 100 free users produce $30/month. That is not enough to justify the support burden. You would need conversion above 10% for the math to work.
2. Low marginal support cost per user
Some products have negligible cost per user (mostly software). Others have real cost per user (support-heavy, high-touch onboarding, or heavy compute per action).
Free tiers work well for the first. They break for the second because free users generate the same support requests as paid ones without generating revenue.
Test: if you added 1,000 free users tomorrow, would your support ticket volume increase proportionally? If yes, a free tier will hurt more than it helps.
3. Activation instrumentation
You need to be able to answer: "Which behavioral signals predict free-to-paid conversion?" Without this data, you cannot optimize the free tier for conversion, and it becomes a slow leak of resources.
See aha moment guide for how to identify the specific behaviors that predict conversion.
What a good free tier includes
The best free tiers give full product experience within volume limits. Not a crippled version.
Volume limits work
"Up to 5 users, 3 projects, 1,000 emails per month."
Users get real value from the product until they hit the limit. When they hit it, the friction is honest (they need more capacity) and the upgrade decision is easy.