The cheapest tutor hour is the one you buy last

The usual advice is to get a tutor early, to build good habits and avoid fossilising mistakes. The model disagrees, and the disagreement is specific enough to act on.

The numbers

Tutoring runs at 1.05× the FSI intensive-class baseline at every level — essentially as efficient as the fastest formal instruction there is. Structured self-study runs at 1.25× at A1 and 1.7× at C1.

So the gap between paying and not paying is:

At the bottom, a tutor saves you about a fifth of your time. At the top, it saves you nearly two thirds.

Spanish to B2, an hour a day five days a week: 460 hours with a tutor, 695 by self-study. 235 hours of difference — and almost all of it accumulates in the upper half.

Why the gap grows

Because what a tutor sells changes as you improve, and it is only scarce at one end.

At A1 you are buying presentation. Someone explains that Spanish has two verbs for "to be" and when each applies. A textbook does that identically, for free, and does not need scheduling. The 19% penalty is mostly pronunciation feedback, which is real and is not worth a tutor's hourly rate on its own.

At B2 you are buying correction of things you invented. You produce a sentence that is grammatical and that no native speaker would say. Nothing except a person can tell you that, because it requires knowing what you were trying to mean. There is no textbook substitute, which is why the penalty runs to 62% and why app-only study stops arriving entirely.

The spending pattern this implies

Most people do the reverse. They buy tutoring at the start — when motivation is highest and the marginal value is lowest — run out of budget or enthusiasm, and then attempt the upper levels alone, where the penalty is largest.

That sequence buys the expensive resource where it is worth least and withholds it where it is worth most.

The order the model supports:

A1–A2: self-study, and spend nothing if you like. The free route is genuinely competitive here. Tutor time spent on your first two hundred words is an expensive way to be quizzed.

B1: start buying, sparingly. This is where the curve bends and where the thing you cannot get elsewhere starts mattering.

B2 and above: this is what the money was for. A few sessions a month, focused entirely on producing and being corrected, is worth more than any number of subscriptions — and costs less than several stacked ones.

Two caveats

Pronunciation is the exception. Errors set early are genuinely harder to fix later, and a small number of early sessions aimed only at sound is a reasonable purchase even under the argument above. The model does not represent this — it prices time, not the durability of mistakes.

These coefficients are estimates. FSI publishes hours for its own classes and nothing else; there is no published self-study comparison because FSI does not run self-study courses. The direction is well attested and the specific values are ours, marked as estimates in the methodology and in the caveats under every result.

Run your own case — same language, same target, tutor against self-study — and the difference it shows is the budget question, priced in hours instead of adjectives.