Is a private SAT tutor worth it? A cost-per-point answer

You’ve already paid for something, a class or an app, and the score didn’t move, so now you’re pricing a private tutor and wondering whether $150 to $300 an hour buys anything the cheaper options didn’t. A private SAT tutor is worth it when the student’s misses are concentrated in a few question types that nobody has named, and it is not worth it when the student hasn’t yet used the free material or can’t say what they’re missing. The way to decide is to stop comparing hourly rates and compare dollars per point instead.

That’s a different calculation from the one most families do. Hourly rate is what you feel at checkout. Cost per point is what you actually bought.

Price every option the same way

The formula is short: total hours times rate, divided by the points the student realistically gains. A tutor’s hourly rate is high and the hours are few. A class’s rate is low and the hours are many. Khan Academy’s rate is zero and the hours are whatever the student gives it, which is the variable most families don’t price.

Here is the formula applied to our own program and to typical figures for the alternatives. The gains in the table are not predictions. They are the break-even points: the gain each option needs to produce to cost what it costs per point.

Read the bottom row honestly. If the full program moves a student 30 points, it was a bad buy at $134 a point, and a class would have been a better use of the money. If it moves a student 120 points, which is the kind of gain the program is built for, it costs $34 a point, roughly what a class costs when the class works.

So the whole question becomes: what gain is realistic for this student? Nobody can answer that from a brochure. It needs the score report.

What Khan Academy is good at, and where it stops

Khan Academy’s Official SAT Practice is free, built with College Board, and uses real question types, so the content is not the problem. College Board and Khan Academy reported in 2017 that students who spent 20 hours on the platform gained an average of 115 points on the paper SAT of that era, a figure we’ve seen repeated and would mark [verify] against the original release before quoting it to a student. Even taken at face value, it describes students across the whole score range, most of whom started well below 1200.

For a student under about 1200 with clear content gaps (linear equations, percent, basic punctuation), Khan is the right first move, and we tell parents that in the first hour when it’s true. The student needs the content, the platform teaches it, and nobody needs to pay us $180 to explain what a slope is.

The platform gets weaker as the score rises. It can tell a student that they missed a Transitions question. It can’t tell them that they miss Transitions because they choose by sound, that they miss Words in Context because they look at the choices before predicting, or that four of their Math misses are the same vertex-form gap wearing different clothes. Diagnosis is the expensive part. An app that serves whatever you tap on doesn’t do it.

It has one more limit that has nothing to do with software. Three months of self-study with no proctored full-length test is the most common story we hear from families at 1350, and we wrote about why that plateau happens separately.

What a class is good at, and where it stops

A class teaches the test’s structure to a room. For a student who has never seen the digital SAT, that is worth something: how the two modules work, what the question types are called, how to pace, how the Desmos calculator behaves. Course chains sell this in packages that typically run $1,000 to $2,500 [verify] for 18 to 30 hours, which is cheap per hour.

The limit is in the word “room.” The instructor teaches to the middle of it, and the material is a fixed curriculum built before anyone met your student. A student at 1150 in a room averaging 1150 gets a lot. A student at 1380 in the same room spends most of the hours on things they already do, and the two question types costing them 60 points may never come up at all.

Classes also tend to test less often than they teach. Our view is that the proctored full-length test, reviewed line by line, is the program, and a class that gives two of them over eight weeks is short on the part that moves scores. Ask any class how many full-length proctored tests are included. Then count them.

What a tutor is actually for

A private tutor is paid for one thing, which is to find the four or five question types costing this student the most points and fix exactly those. Everything else a tutor does (pacing, strategy, encouragement) is available cheaper elsewhere. The diagnosis is not.

College Board publishes the test’s structure at satsuite.collegeboard.org: 98 questions, 54 Reading and Writing and 44 Math, each section in two modules with the second module chosen by performance on the first, scored 400 to 1600. A student at 1380 is typically missing around eighteen to twenty questions. Those misses are almost never spread evenly. They sit in a handful of types, and a good tutor can name them from the score report in the first hour.

That’s why the tutor’s rate is high and the hours are few. You’re not buying 28 hours of explanation. You’re buying a hand-read diagnostic and then 28 hours spent only on the student’s own leaks, with a weekly proctored test to prove it’s working. For the right student, that’s the cheapest option per point on the table above. For the wrong student, it’s the most expensive.

Rick’s note: I read every score report myself, and about one in four times the honest answer is that the family shouldn’t hire me yet. Usually it’s a student who hasn’t done the free work, or a test date four weeks out. I’d rather say that in a free hour than take the money and explain it afterward.

What 50 points is worth in Texas merit aid

Cost per point only means something next to the value of a point, and in Texas the value is unusually legible because several public universities publish merit tiers by score. These are the figures we could confirm in November 2026; the dollar amounts shift yearly, so treat every one as [verify] against the school’s own page before you plan around it.

Now the arithmetic a parent can actually use. At Texas Tech, a student at 1370 is 30 points from the 1400 band, and the bands are where the dollars change. If moving bands is worth even $1,000 a year more [verify], that is $4,000 over four years for 30 points, which is $133 a point of value against $34 to $67 a point of cost. At Baylor, a student at 1280 is 30 points from a scholarship with a $6,000-a-year floor, or $24,000 over four years. Fifty points there is worth more than any program on the market costs. That is the whole case.

The arithmetic cuts the other way too. A student at 1520 applying to UT Austin and Rice gets nothing automatic for a 1570, because neither school pays by the point. The value of 50 points for that student is in admission odds, which are real but not a number we’ll invent.

Who should not hire a tutor

We’d rather lose the booking than take it from the wrong family. A private tutor is the wrong purchase, at any rate, for these students.

• The student who hasn’t used the free material. Under about 1200 with obvious content gaps, Khan Academy and a stack of Bluebook tests will do more per dollar than anyone charging $180.
• The student who has never taken a full-length proctored test. There’s no diagnosis without one. Take it first; the first hour can be where it happens.
• The student with fewer than four weeks to test day and a target 100 points away. Nobody can do that honestly, and anyone who says they can is selling hours.
• The student who won’t do the work between sessions. The tutor teaches; the student practices. A twice-weekly session with nothing in between is an expensive conversation.
• The family for whom $4,032 is a strain. A class or an app is cheaper. We mean it plainly: there are good outcomes available for a fraction of our price, and we’d rather you got one.
• The student already at 1550 or above. The marginal points are the most expensive on the test, and whether they’re worth buying depends on the school list, not on us.

Everyone else is in the middle, and the middle is where the arithmetic above actually decides things.

What to do this week

1. Get a real number: one full-length Bluebook test, timed, both sections, at a Saturday start time. No tutor, class or app is worth evaluating without it. Nothing else counts yet.
2. Write down the school list and look up each school’s merit tiers on its own site. Note which score bands are within 50 points.
3. Do the cost-per-point math for each option you’re considering, using the break-even table above and a gain you’d be satisfied with, not one you hope for.
4. If the student is under about 1200 and hasn’t done 15 hours on Khan Academy, do that first and come back in a month.
5. Otherwise, book a free first hour and bring the test’s question-by-question review. We’ll name the leak and the realistic target, and you’ll know whether we’re the right buy.