How Much to Charge for Tutoring: Set Rates and Packages With AI

Coding Liquids tutorial cover featuring Sagnik Bhattacharya for How Much to Charge for Tutoring: Set Rates and Packages With AI.
Coding Liquids tutorial cover featuring Sagnik Bhattacharya for How Much to Charge for Tutoring: Set Rates and Packages With AI.

Charge at least your yearly income target plus costs, divided by the hours you will really bill, then place that figure within the range of eight to ten comparable listings for your subject and level. For example, $51,600 over 745 billed hours sets a floor near $70 an hour. AI runs the sums, tests package discounts and drafts messages.

There is no single right figure, because subject, level, exam stakes and local demand all move the market rate. The common mistake is asking an AI chatbot "what do tutors charge?" and trusting the answer. It has no view of your local market, your waiting list or your travel time, so it produces a smooth, confident average that may be wrong for you. Use it as a calculator and a drafting assistant, and feed it the evidence yourself.

Follow me on Instagram@sagnikteaches

Start from the income you need, not from a rate card

Hourly rate follows from three numbers: the take-home income you want, the costs you carry, and the hours you can realistically teach. The third number is where most tutors go wrong, because they divide by the hours they are available instead of the hours that are billed.

Connect on LinkedInSagnik Bhattacharya

Here is an illustrative case. A tutor wants $48,000 a year before tax, has about $3,600 of yearly costs (software, a laptop share, insurance, printing, a resource subscription) and wants six weeks off a year. That is 46 working weeks. They can teach 18 paid hours a week at most, because after-school and weekend slots are the only ones in demand. They also expect around 10% of booked lessons to be cancelled and not rebooked, and about 5 hours a week goes on prep, messages and invoicing that nobody pays for.

Subscribe on YouTube@sagnikteaches
StepFigure (illustrative)How it is worked out
Income target plus costs$51,600$48,000 + $3,600
Teaching weeks4652 less 6 weeks off
Maximum paid hours82818 hours x 46 weeks
Realistic paid hours745828 less 10% lost to cancellations
Minimum hourly rate$69.26$51,600 / 745

So a rate under about $70 an hour leaves this tutor short of their target, however pleasant the lessons are. Notice that the unpaid 5 hours a week do not appear in the sum: they are already inside the 18-hour cap, because those hours are the reason the tutor cannot teach more. If you want to count unpaid time differently, change the cap and see how far the number moves.

Let AI check the sum, not invent it

Give the chatbot your own figures and ask it to show every step. A paid plan is not needed; the free tiers of ChatGPT, Claude and Gemini all handle this. Never paste pupil names or anything about a child.

I am a private tutor. Work out the minimum hourly rate I need.
My inputs:
- Take-home income target: $48,000 a year (before tax)
- Yearly costs: $3,600
- Weeks off: 6
- Maximum paid hours per week: 18
- Expected cancellations I cannot rebook: 10%
Show each step in a table, then give the minimum hourly rate.
Then show what the rate becomes if cancellations are 5% and if they are 15%.
Do not suggest a market rate. I will supply listings separately.

A typical reply gives the table and then a range. With 5% cancellations the same sum lands near $65.60; at 15% it rises to about $73.32. That sensitivity is useful on its own: it tells you how much a cancellation policy is worth. If the reply misses a step, tell it which one. Check at least one line by hand with a calculator, because chatbots slip on long chains of arithmetic.

The Complete Excel Course with AI Integration thumbnail

The Complete Excel Course with AI Integration

From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.

By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner

Buy now

Gather the evidence before you trust the number

The income maths gives you a floor. The market tells you whether you can reach it. Collect eight to ten listings for tutors who teach the same subject and level as you: tutor directories, school noticeboards, community groups, and the pricing pages of tutoring centres in your catchment. Write down the rate, the subject, the level, the lesson length, whether it is in-person or online, and anything that explains a premium (exam specialist, qualified teacher, years of experience).

Paste the cleaned list into the chatbot as a small table. Do not paste profile photos or names; subject, level, rate and notes are enough.

Subject and levelRate per hourFormatNote
Maths, lower secondary$45OnlineNew tutor, no reviews
Maths, lower secondary$55In personTwo years teaching
Maths, exam year$70In personQualified teacher
Maths, exam year$62OnlinePast papers included
Physics, exam year$75OnlineWaiting list stated
English, lower secondary$50In personGroup option too

All of those figures are made up to show the layout. Replace them with what you actually find. Then ask for a comparison that stays inside the data.

Here are comparable tutor listings I collected (the table below is real data I pasted).
[paste your table]
Using only this table:
1. Give the lowest, middle and highest rate for my subject and level.
2. List the factors that seem to explain the higher rates.
3. Say where my profile (qualified teacher, six years, exam specialist) fits.
4. Tell me what is missing from the data before I rely on it.
Do not add outside figures.

The fourth instruction matters. A good reply will point out that eight listings are a small sample, that listed rates are not the same as rates people actually pay, and that nothing here shows demand. That caution is worth more than the numbers. Treat the middle of the range as a starting point and let your enquiries tell you the rest.

Test packages instead of arguing over the hourly figure

A single hourly rate forces you to pick between charging enough and being easy to say yes to. Packages let you do both: a higher single-lesson price, and a lower effective rate for families who commit. Ask AI to build three options and then check each one against your floor of $69.26 from the sum above.

Say your market check suggests $72 for a single exam-year lesson. A reasonable ladder might look like this:

OptionPriceEffective rateClears the $69.26 floor?
Single lesson, pay as you go$72$72.00Yes
Block of 6, paid upfront$414$69.00No, by 26 cents
Block of 6, paid upfront$420$70.00Yes
Block of 10, paid upfront$690$69.00No
Block of 10, paid upfront$700$70.00Yes

The point of the table is that a "10% off" block feels generous and quietly breaks your income target. A 10% discount on $72 is $64.80 an hour, which is $4.46 below the floor. Over 745 hours that is a shortfall of more than $3,300 if every family took it. A discount of about 3% on a block of six, with the money collected upfront, can still make sense because it removes chasing, no-shows and the cost of unsold slots. You decide the trade-off; AI shows you the cost of each version.

A prompt that stress-tests your package ideas

My minimum hourly rate is $69.26. My single-lesson price is $72.
I am considering these packages:
- 6 lessons for $420, paid upfront
- 10 lessons for $700, paid upfront
- a monthly plan of 4 lessons for $280, billed on the first of the month
For each one:
1. Give the effective hourly rate and the gap to my minimum.
2. Say what happens to my income if 30% of families choose it.
3. Name one way it could go wrong (cancellations, expiry, refunds).
4. Suggest a cancellation and expiry rule in plain words.
Keep each answer under 80 words.

A sensible answer will flag the monthly plan: four lessons for $280 is $70 each, but if a family skips one lesson and expects a credit, the effective rate drops quickly. It will also raise expiry, which is the usual source of arguments. A block that "expires after 12 weeks, unused lessons are not refunded but one can be moved with 24 hours' notice" is clearer than a silent assumption. Ask the chatbot to draft the wording, then read it as a parent would.

Rate by subject and level, not one number for everyone

Your sum gives one floor, but you probably teach several subjects. Rates tend to rise with exam stakes and with how scarce your subject is. A tutor who teaches primary reading, lower-secondary maths and an entrance exam might sensibly charge three prices. Your market table shows the pattern for your market; you do not need to copy it.

Try this quick exercise. List every pupil type you teach, how many slots each takes a week, and what the market table says each pays. Ask the chatbot to calculate your weekly income under the current prices and under a version with a small premium for the subject with the longest waiting list. If one subject fills in a day and another sits half empty, the difference is information: the full one is underpriced, the empty one may be overpriced or need better marketing.

That test is worth running before you follow advice to raise everything by the same percentage. A flat increase treats an oversubscribed exam slot and a hard-to-fill Tuesday afternoon the same way, and they are not the same problem. For finding pupils for those quieter slots, see the tutorial on getting your first ten local tutoring clients.

Online and in-person lessons price differently

An in-person lesson carries travel or room hire, and a lost slot if the family cancels at the door. An online lesson has none of that but competes with tutors who can be anywhere. Keep two lines in your sum. For in-person work, add travel time to the paid-hours cap (a 30-minute drive each way turns a one-hour lesson into two hours of your time). For online work, check which rates your comparison set quotes, because online-only tutors often price lower.

Example: a tutor who drives 20 minutes each way to a family and teaches for an hour spends 100 minutes for $70, an effective $42 an hour of time. The same family online costs 60 minutes for $70. Some tutors respond by adding a travel supplement; others set a minimum of two lessons per visit. Ask AI to write both versions of the wording and choose the one that suits your tone.

Group lessons, siblings and exam intensives

Three situations break a simple hourly rate, and each is worth a short sum before a parent asks.

Small groups. Suppose you teach a pair of pupils at $45 each for an hour. That is $90 of income for one hour of your time, and it clears the $69.26 floor easily. Three pupils at $35 each is $105. But group lessons carry a cost the table does not show: more marking, mixed ability and a higher chance that one cancellation leaves a half-empty slot. Ask the chatbot to work out the break-even group size for your floor and then add a cancellation rule, such as "the group rate applies when at least two pupils attend; otherwise the single-lesson rate applies".

Siblings. A sibling discount is a habit more than a necessity. If two children from one family are taught back to back, you save travel and setup time, so a small saving is fair. Write down the exact discount (for example, 5% off the second child) and stop there. If a parent asks for more, offer a block of lessons rather than a bigger percentage.

Exam intensives. A fortnight of daily revision sessions before an exam is a different product from weekly lessons. The parents are buying a result by a date, and your calendar is full of sessions you cannot reschedule. Price it as a fixed package. Give the chatbot the dates, the number of sessions and your normal rate, then ask for a fixed price with a 15% to 25% premium and a plain-English refund rule. Check the output against your floor before you send it.

What a family sees when you quote

A price is only half the message. The other half is how it reads on a phone at the end of a school day. Ask the chatbot to turn your pricing into a three-option reply you can paste into an enquiry.

Write a reply to a parent who asked "how much do you charge for exam-year maths?".
My options: $72 for one lesson, $420 for a block of 6 paid upfront, $700 for a block of 10 paid upfront.
Blocks expire after 12 weeks, one lesson can be moved with 24 hours' notice.
Tone: warm and plain. Under 120 words. End with an offer of a 20-minute assessment call. Use [first name] and [your name].

A usable draft lists the three options in order, names the saving in cash terms ("six lessons for $420 works out at $70 each") and ends with the call to action. Fix two things when it comes back. First, remove any claim about results, such as "most pupils improve by a grade", which you cannot support. Second, check that the expiry and cancellation lines say exactly what you intend, because that is where disputes begin.

Check what your tools cost against the rate

If you pay for a chatbot plan, a scheduling tool and a worksheet site, those costs sit inside the $3,600 in the sum. Work out whether a paid tool earns its keep: ChatGPT Plus and Claude Pro were both listed at $20 a month on monthly billing in November 2026, so if one saves two hours of prep a month at a $70 rate, it pays for itself seven times over. If it saves nothing, cancel it. The tutorial on whether AI is worth paying for in a small tutoring business goes through that test.

The price-rise email

Raising prices is where tutors lose nerve. A clear message sent early, with a reason and a date, causes fewer problems than a quiet change on an invoice. Give the chatbot the facts and ask for a short draft. Remember to use "[first name]" and "[your name]" placeholders rather than real names.

Write a short, warm email to parents telling them my hourly rate is rising from $65 to $70 from 1 January.
Include: thanks for working with me, a reason (my costs and the extra resources I now provide), that blocks already paid for are honoured at the old rate, and an invitation to reply with questions.
Under 150 words. No apology, no hard sell. Use [first name] and [your name].

A typical draft is fine in tone but over-apologises in the first line ("I'm so sorry to have to tell you...") and may invent a reason you never gave, such as "rising materials costs". Cut the apology and keep only reasons that are true. Add the one line the AI will not know to add: what happens to a family who cannot afford the increase. A single sentence offering a longer block or a shorter lesson at the old rate keeps the relationship intact without lowering your price for everyone.

Before and after: the price-rise note

Before: "Hi all, rates are going up a bit soon, sorry, let me know if that's a problem." After: "Hi [first name], from 1 January my rate moves from $65 to $70 an hour. Lessons you have already paid for will be honoured at the current rate. I am happy to talk it through if you would like to change the lesson length or timing. Thank you for working with me. [your name]". The second version gives a date, a figure and a way forward in fewer words.

Roll out the change and measure it

After the new prices go live, track four numbers for two months: enquiries per week, trial lessons booked, enquiries that became regulars, and pupils who left within a month of the rise. A quick sum tells you how many leavers you can absorb: a move from $65 to $70 lifts income per lesson by about 7.7%, so weekly income only falls if you lose more than roughly one pupil in fourteen, and fewer if your waiting list refills the slots. A rise that stops enquiries converting is a signal to look at the packages or the profile, not necessarily the price. You can keep the tally in a simple spreadsheet; ChatGPT can help set up the formulas if you rarely use one.

Replying quickly to new enquiries also moves conversion more than most price changes do. The tutorial on replying to every enquiry in under five minutes covers that, and scheduling software and no-show reminders deals with the cancellations that drag your effective rate down.

Four ways a pricing exercise goes wrong

  • Using an AI average as a market rate. If the chatbot quotes a typical rate without being given listings, it is guessing. Ask it where the figure came from; if it cannot say, discard it.
  • Counting available hours as paid hours. A tutor who plans for 25 paid hours a week but regularly teaches 15 has set a rate about 40% too low. Use your diary from the last term, not your hopes.
  • Forgetting the free first lesson. If you give one unpaid trial per new pupil and keep 6 in 10, the cost is spread over every paying lesson. Offer a paid short assessment or a half-price trial instead, and put that in the sum.
  • Discounting for every sibling and every block together. Stacked discounts can take the effective rate well under your floor. Cap the total discount at a figure you have checked.

A monthly routine that keeps your rates honest

Pricing is not a once-a-year job. Spend twenty minutes at the end of each month: record the number of lessons taught, cancelled and rebooked; compare the effective hourly income with your floor; check two or three new listings in your market; and note enquiries you lost on price. Paste the numbers into the chatbot and ask for a three-line summary and one suggested change. Do not act on the first suggestion; wait until a pattern repeats for two months.

The floor itself needs updating once a year. If your costs rise, or you decide to take eight weeks off, rerun the sum. If you would like help mapping which of these pricing and admin jobs AI should take on, that is what my AI implementation consultation is for.

Questions tutors ask about pricing

Should I charge less for younger pupils?

Not as a rule. Price follows the subject, the exam stakes, your experience and the demand for your slots, not the child's age. Many tutors charge the same for a primary pupil as for a lower-secondary one and add a premium for exam years or scarce subjects. Check what comparable listings charge for each level before deciding.

Is a block-booking discount worth giving?

Usually yes, if it buys you commitment and cash upfront. A modest discount on a block of six or ten lessons reduces drop-outs and cuts the time you spend chasing payments. Keep the discount small enough that your effective hourly rate still clears your income target.

How often can I put my rates up?

Once a year is normal, and giving a month or two of notice is fair. Many tutors raise rates for new pupils straight away and for existing pupils at the start of a term. Put the date and the new rate in writing so nobody is surprised at the next invoice.

Can AI tell me the right price?

No. AI cannot see your local market or your waiting list. It can do the sums, compare options you give it and draft the messages. The market evidence has to come from listings, enquiries and your own booking data, which you paste in.

Further reads

Sources: the tutor's own income and hours figures, comparable public tutor listings checked by the reader, and the ChatGPT, Claude and Gemini plan pages (checked November 2026).

Want your tutoring prices and packages reviewed?

On a 1:1 call we'll look at how you set rates and packages today, pick the pricing job worth handing to AI first, and set up prompts you can reuse each term.

Book a 1:1 call with me