Home tuition fees in Delhi NCR: what 1,286 self-set hourly rates show

The population, stated before any number

Everything here comes from one measurement taken read-only against production on 22 September 2026. The platform carries 7,500+ registered tutor profiles; 600+ of those have been background-checked; 2,600+ were created in the preceding 30 days. The rate study itself covers the 1,286 profiles carrying a self-set hourly rate between ₹50 and ₹5,000. Thirty-eight rows fell outside that window and were dropped as junk, 2.9% of the rate-bearing rows.

What the number is: a rate the tutor typed into their own profile at signup — what they ask. Every figure below names the n it was computed over, inside the sentence that carries it, because a percentage without its denominator is not a finding.

The ₹500 wall

Across the 1,286 rate-bearing profiles, 483 — 37.6% — sit at exactly ₹500. The next five most common rates are ₹300 (126 profiles, 9.8%), ₹400 (95 profiles, 7.4%), ₹600 (83 profiles, 6.5%), ₹200 (81 profiles, 6.3%) and ₹1,000 (74 profiles, 5.8%). Six numbers account for 73.4% of the entire set.

The coarseness goes well beyond the top six. 1,118 of 1,286 rates — 86.9% — are a multiple of ₹100, and 1,257 of 1,286 — 97.7% — are a multiple of ₹50. Twenty-nine profiles out of 1,286, 2.3%, priced at anything else at all.

A price grid that coarse is what you get when sellers have no reference price to anchor to. Someone setting a rate is not reading a market; they are picking the roundest number that feels defensible, and ₹500 absorbs everyone with no particular reason to sit above or below it. The practical consequence for a parent is that a ₹500 asking rate carries almost no information about the person quoting it. It is the modal answer, not a positioning statement — and the signal sits with the people who priced deliberately off the wall, at ₹350 or ₹800, because they had a reason to.

Why the median is the wrong number here

When 37.6% of 1,286 profiles type one number, that number becomes the median of almost any slice you cut. The median is ₹500 in New Delhi, ₹500 in Noida, ₹500 in Gurgaon, ₹500 in Ghaziabad and ₹500 in Mumbai — and ₹500 for every subject with enough rows to report. That is a property of the clustering, not a finding about the cities or the subjects. A study reporting “the median is ₹500 everywhere we looked” has measured its own denominator.

So the reportable quantity is the interquartile range: p25 to p75, the middle half of the market. That does move, and it moves by city.

Reading the city table: three separate stories

The ceiling is a Gurgaon story. Gurgaon’s p75 is ₹700 (n = 75), the highest of the five and ₹100 above New Delhi, ₹200 above Noida. A quarter of Gurgaon profiles ask ₹700 or more, and it is the only city here whose top quartile clears ₹500 decisively.

The floor is a Mumbai story. Mumbai has the lowest p25 in the set at ₹300 (n = 54), which is the opposite of what cost of living would predict. With n = 54 this is the least reliable row on the page and should be read as a flag for re-measurement rather than as a conclusion. If it holds, the likeliest explanation is composition — a cheaper, more online-heavy mix of people listing — rather than a genuinely cheaper city.

Noida is compressed. p25 ₹350 and p75 ₹500 on n = 141, so its 75th percentile is the mode itself: three quarters of Noida profiles ask ₹500 or less. Noida is the city where ₹500 is the ceiling rather than the centre, and it is the clearest single-city illustration of the whole pattern.

Mode and experience: two cuts that move, one that does not

Online against at-home splits at the floor and not in the middle. Online profiles (n = 531) show p25 ₹300, median ₹500, p75 ₹500. Profiles offering tuition at the student’s home (n = 263) show p25 ₹400, median ₹500, p75 ₹600. The medians are identical; the gap is ₹100 at the 25th percentile. Home tuition carries a floor that online does not, which is what you would expect once travel time is priced in, and it shows up exactly where you would expect it — at the bottom of the market, among the people with the least pricing power. At the median it vanishes into the wall.

Experience moves one bucket, and it is not the one most people assume. 0–2 years (n = 183) reads p25 ₹400, median ₹500, p75 ₹500. 3–5 years (n = 225) reads p25 ₹400, median ₹500, p75 ₹600. 6–10 years (n = 100) is the only slice in the entire study whose median is not ₹500: p25 ₹500, median ₹600, p75 ₹800. The 10+ years bucket (n = 47) falls back to a median of ₹500 with p75 ₹700, on an n thin enough that the drop should not be over-read. The return to experience arrives late and then flattens.

Background-checked profiles carrying a rate (n = 316) read p25 ₹400, median ₹500, p75 ₹500 — a lower p75 than the ₹600 of the wider group (n = 1,286). On this evidence, being background-checked is not something people price for.

The obvious objection, tested: is ₹500 just a form default?

It is the first question any editor asks, and it has a specific answer rather than a reassurance. The backend field that stores this rate is declared with @Prop({ type: Number, default: 0 }). The schema default is ₹0, not ₹500.

That makes the argument stronger rather than weaker. A field defaulted to ₹500 could not leave rows sitting at zero, and 298 profiles do sit at zero in our own recorded exclusions — which is precisely what a default of ₹0 must produce. So the schema does not merely fail to support the form-default explanation; it rules it out. The 483 profiles at ₹500 chose ₹500.

Scope, stated rather than implied: 5 write paths that can set this field from a user action were examined, of the 5 that exist, together with one repo-wide search. The admin-create path and any historical seed or migration were not examined — neither of which could produce 483 rows without somebody noticing.

What this dataset is not

These limitations travel with the data. They are not a disclaimer at the bottom; they are the reason the numbers above are quotable at all.

It is not a transacted price. We do not take a commission and money does not flow through us; the family pays the tutor directly. So we see the asking rate and never the settled one. Real paid rates are plausibly lower at the top end, where asking prices are least disciplined.

It is not a national sample. 903 of the 1,286 rows yield a recognisable city, and the five cities with enough rows to report are all Delhi-NCR plus Mumbai. This is a Delhi-NCR dataset with a Mumbai comparison, not an all-India one. Anyone writing “India’s tutors charge…” from this page is overstating it, and so are we if we do it.

It is self-selected. These are people who chose to list on a platform. Those who work entirely by word of mouth are absent, and they are not a random subset.

It is self-reported. The number is what someone typed into their own profile at signup. We publish it anyway because the alternative sources are worse: the fee ranges circulating in Indian tuition content are almost all unsourced, including — until recently — one of our own blog posts, which claimed a band for Delhi NCR against no data at all and has since been retired. A disclosed, bounded, self-reported dataset of 1,286 rows beats an undisclosed guess.

Re-measure before quoting this after 22 December 2026. The base is growing; a three-month-old percentile is a different dataset.

How to use these numbers

For a parent: ₹350–₹600 per hour is the honest answer for Delhi-NCR home tuition — the middle half of the market, computed over 1,286 self-set rates. Any single number quoted without a range is hiding the spread. Budget by percentile rather than by average: ₹350 puts you at the bottom quartile of the bounded dataset, ₹700 puts you inside Gurgaon’s top quartile (n = 75), and ₹1,000 is the 90th percentile of the whole 1,286. Above that you are paying for something other than the market rate. Expect a floor premium for at-home over online (₹400 against ₹300 at p25) and expect no discount for a profile that has not been background-checked.

For a tutor: 483 profiles of 1,286 are standing on ₹500 with you, and a parent scanning a list cannot tell you apart on price. Pricing off the wall is rare and therefore legible — 1,118 of 1,286 rates, 86.9%, sit on a multiple of ₹100, so a rate like ₹350 or ₹650 is visibly deliberate against that grid. Rarer still, and a different claim: only 29 of 1,286, 2.3%, are not even a multiple of ₹50. And the 6–10 year band (n = 100) is where this market actually pays: its p75 is ₹800, so if you are in that band and asking ₹500 you are probably under-charging.

If you are budgeting rather than benchmarking: tell us the class, the subject and your locality, and you will see real profiles with each tutor’s own asking rate on it. You pay the tutor directly — EZ Tuitions takes no commission on the teaching fee.

Frequently asked questions

Related tutoring