A free tier is a pricing page, not a product

The three kinds of free tier, how to tell which one you are holding, and the graduation cliff that turns $0 into the most expensive plan we track.

The bench2 min readCosts

We price everything on this site against a stated workload, and the number that generates the most mail is $0. "Why did the free tier rank below a paid plan?" Because a free tier is not a price. It is a pitch, and the pitch is aimed at a version of you that does not stay free.

There are three kinds, and they behave nothing alike.

The trial in disguise. Generous limits, every feature unlocked, expires the moment your usage looks like production. You can spot it by what's missing: no backups, no SLA, data retention measured in days. This tier exists to get the integration written. Fine. Use it exactly that way, and write the migration path in the same sprint you write the integration, because you will exercise it under deadline otherwise.

The community edition. Limits chosen so a hobbyist never pays and a company always does. This is the honest kind, and it shows up in our rankings scored on its paid tiers, with the free tier noted as what it is: a genuinely useful gift with a business model attached. The tell is that the limits are stable for years. Stability is the vendor telling you the math works.

The loss leader with a cliff. Free until a threshold, then a first paid tier priced like the vendor is recovering two years of your infrastructure costs, which, to be fair, they are. The pattern shows up across categories we have ranked: the serverless databases with a paid floor several multiples above the last free month, the static hosts where the first overage dollar arrives with three friends (our static hosting shootout scored exactly this). The cliff is not an accident. It is placed where migrating away hurts most.

The scoring consequence: we rank tools by what the second year costs at our stated workload. A free tier changes that number only if the workload genuinely fits inside it with room to grow. Otherwise the free tier is marketing spend, and we do not score marketing spend, in either direction.

The reader move, in three steps. First, write down the workload where you cross each limit; the vendor published the limits, so this is arithmetic, not divination. Second, price the first paid tier at that crossing point and call that the real price. Third, check whether export is gated: if leaving requires the paid tier, the free tier has a cover charge on the exit door, and that belongs in the decision at month zero.

None of this is a complaint. Vendors may spend money acquiring users; we may decline to confuse the acquisition budget with the product. The free tier got you to the pricing page. Read the rest of it.