Tutoring math online has a low barrier to entry and a high early failure rate. Most people who quit in the first two months do so for one of three reasons: they never found students, they charged too little to sustain it, or their sessions were lectures the student sat through passively and did not improve from.
This is the practical version — setup, pricing, the first session, and how to prepare without burning an hour per student per week.
Decide what you actually teach
"Math" is not a service. "Algebra 2 and precalculus for high schoolers, and SAT math" is. Narrowing does three things: it makes you findable, it lets you get genuinely good at the specific misconceptions in that band, and it means your prep compounds instead of restarting every week.
Pick from where you are strongest and where demand is reliable: elementary foundations, middle-school pre-algebra, algebra and geometry, precalculus and calculus, statistics, standardised test math, or university-level courses. Test prep and university courses pay the most; elementary work is the steadiest.
Platform, independent, or both
Marketplaces — Wyzant, Preply, Superprof, Varsity Tutors, Tutor.com and similar — supply students and take a cut, historically in the range of roughly 15 to 30 percent, sometimes higher on early lessons. You will earn less per hour than working independently. You should still start here, because you have no reviews and no reputation, and reviews are the asset you are actually accumulating in month one.
Independent means you keep the whole fee and do all the acquisition, scheduling, invoicing and chasing. A booking link, a payment processor and a simple policy document is the whole infrastructure. Move here once referrals arrive on their own.
Group tutoring — three to five students at a reduced per-student rate — is the most underrated option. It raises your hourly income, lowers the cost per family, and works well for exam revision where students benefit from hearing each other's mistakes.
Most sustainable tutors run both: a marketplace for discovery, private clients for the majority of income.
The technical setup
The one thing you cannot compromise on is a shared writing surface. Math happens in writing; a session conducted by talking over a video call fails. Options, in rough order of preference:
- Tablet with stylus writing into a shared whiteboard the student can also write on.
- Inexpensive graphics tablet plus a whiteboard app on a laptop.
- Phone on a cheap stand pointing at paper, used as a document camera.
Add a wired headset, a wired network connection if possible, and a plain background. Record sessions only with explicit permission, and if you do, sending the recording or the whiteboard export afterwards is a small thing families value a lot.
Setting your rate
Rates vary widely by country, level and channel, so treat these as orientation rather than quotes. Independent school-level tutors commonly sit somewhere around 30 to 60 US dollars an hour, with advanced and test-prep work higher, and marketplace rates lower for equivalent work. Regional variation is large in both directions.
Three rules that hold anywhere:
- Do not compete on price. The floor is occupied by people who will not last. Families who choose purely on price also churn fastest.
- Discount your first students explicitly and temporarily. "My introductory rate for my first three students, in exchange for a review" is honest and does not anchor you low forever.
- Raise rates for new clients only. Existing families keep their rate for at least a year. This removes the awkwardness that stops most tutors from ever increasing prices.
Bill in blocks of four or six sessions where you can. It improves attendance, it improves outcomes, and it stabilises your income.
Running a first session that earns a second
Sixty minutes, structured:
- Ten minutes: diagnose. Not a test — a conversation over two or three problems from their current unit, with the student working while you watch. You are looking for where it breaks, and it is almost never where the parent said. Sign errors, fractions, and inability to translate a sentence into an equation account for a large share of "bad at algebra".
- Five minutes: name it. Tell the student and the parent plainly what you saw. "She understands the method; the errors are all negative signs in distribution" is worth more than an hour of teaching, and it is why they will book again.
- Thirty minutes: one topic, student holds the pen. The most common new-tutor mistake is solving problems beautifully while the student watches. If you are writing more than they are, the session is failing. Ask what the next step should be and wait through the silence.
- Ten minutes: they teach it back. The student explains the method to you. Gaps surface immediately.
- Five minutes: set specific work. Three to five problems, not a chapter.
Preparing without losing your evenings
Prep is where new tutors burn out, because they build materials from scratch per student. Two habits fix it.
Build a bank, not a lesson. For each topic you teach, keep one folder: the diagnostic problem, four graded practice problems, the two misconceptions that always appear, and a clean worked solution. Reused across every student who hits that topic, it pays for itself by the third session.
Use a solver to generate and check working, not to replace your judgement. Producing variant problems at a fixed difficulty and clean step-by-step solutions is exactly the tedious part, and running them through a step-by-step tool such as AI-Math — for quadratics, word problems, derivatives, triangles — turns an hour of prep into a few minutes. Two caveats that matter professionally: check every answer key before you use it with a student, because a wrong key in a session costs you credibility you will not recover; and never let generated steps substitute for your own explanation, since what the family is paying for is precisely the human diagnosis a solver cannot do.
Getting students
Marketplace profile first, with a specific headline rather than "math tutor". Then tell every network you already have — school parent groups, community boards, former teachers, local social groups. Nearly every tutor's first three students come from a personal connection, and after a term the dominant channel is referral from families whose child improved. Ask for the referral explicitly; most satisfied parents simply do not think of it.
Related reading: