How much do people actually make on Gumroad?

A directly measured answer, for one corner of the marketplace

Most of this page measures demand through rating counts, because that is what Gumroad publishes for every listing. But about a quarter of listings publish a real unit-sales count, and for those the question can be answered directly: units sold, multiplied by the asking price. That is a different and smaller sample from the 1,344 listings the rest of this page uses — 277 paid listings drawn from the category walk, not the search sample — and the two are never combined.

Which categories this covers. The listings behind the 277 listings in this section come from 15 of Gumroad's 15 top-level categories (3D, Audio, Business & Money, Comics & Graphic Novels, Design, Drawing & Painting, Education, Fiction Books, Films, Fitness & Health, Gaming, Music & Sound Design, Other, Photography, Recorded Music), so they are not an even draw across the platform. They are also unevenly weighted: 57% of them sit under 3D alone, against the 7% an even draw would give it. Weight them accordingly.

There is no single answer here, and quoting one would be the error this page exists to correct. Split by category branch, these 277 listings are two different populations about 44-fold apart. Under 3D: 178 listings, median $286 lifetime gross, 64.6% of them under $1,000. Everywhere else: 99 listings, median $12,638, 8.1% under $1,000. A median across both would land between them and be true of neither group, so this page does not print one.

Do not read the second figure as “the rest of Gumroad is richer” either. Those 99 listings are spread thinly over 14 branches, and two of them alone hold 54.8% of the $10,409,302 that group has taken between them — so its median rests on very few rows per branch. Publishing a unit count is also voluntary and take-up is uneven between branches, from 41% in Design down to 7% in Audio, and a branch where few sellers disclose is showing only the listings whose owners wanted them seen. Which way that cuts differs by branch and is not worked out here.

The table above pools both groups, because the shape of the distribution is the one thing they agree on: it is extremely top-heavy either way. 23 listings out of 277 account for 80.7% of the $14,414,664 this sample has taken between them, while the bottom half share 0.3% of it. That is why any "average Gumroad seller earns X" figure is describing those 23 listings and calling it the middle — and it is the same reason the two medians above are printed separately rather than averaged into one.

Beyond the category limit above, read this as a ceiling rather than a middle — four more reasons. One: publishing a sales count is voluntary, and a seller with nothing to show has less reason to show it, so this subsample is selected upward. Both medians above are higher than the truth for their group, not lower. Two: the price used is today's price; sellers discount, raise prices and run launch offers, and pay-what-you-want buyers often pay above the minimum, so no listing sold every unit at the price we observed. Three: this is gross, before Gumroad's fee, refunds and tax — the seller keeps meaningfully less. Four: it is lifetime over an age we cannot see, not annual: a listing that has grossed $3,715 may have taken five years to do it. So treat these as the shape of the distribution, which is reliable, rather than as anyone's income, which they are not.

The same subsample is what makes the rest of this page legible: one rating is worth a median of ×25.5 units on paid listings, with a quartile range from ×11.7 to ×54.2 — so a rating count is a floor on sales, and a wide one. Every rating-based figure below should be read with that spread in mind.

What can and cannot be measured

Gumroad publishes a unit-sales count only where the seller has opted into showing one, which is a minority of listings. For the rest it publishes rating counts, and a rating requires a purchase. So rating count is a floor on units sold, not an estimate of it — most buyers never rate anything. Everything below uses ratings as a relative signal: it tells you which listings outsell which, and it does not tell you anyone's revenue.

Correction, 7 August 2026. This page previously said Gumroad does not publish sales counts at all. That was wrong: the product page carries a sales_count field, populated wherever the seller displays it. The figures on this site are unaffected — they have always been rating-derived and are labelled as such — but the stated reason for using a proxy was overstated. That field has since been collected across the category walk, and the result is the section at the top of this page and the sales-per-rating guide: 316 of 1359 listings publish a unit count, 450,651 units in total. Sellers routinely repeat multipliers like ×30 or ×100 with nothing behind them; the measured figure is a median of ×25.5 on paid listings with a quartile range of ×11.7–×54.2, which is why this site never prints a single multiplier.

The distribution is the answer

Read those four numbers together and the shape is clear. 458 of 1,344 listings show no evidence of a single sale. Among the ones that do sell, the median has 9 ratings — a real but modest number. And within any given category, three listings take 65% of everything. The most-rated single listing in this sample is “[Addon]MACHIN3tools machin3 Top creator 4.9 (4K) £1.49+” at $2.01 with 4,000 ratings.

How lopsided is it, across the whole market?

The per-category figure above is about crowding inside one niche. This is the same question asked across all 1,344 products at once, using the 62,134 ratings they carry between them:

The top 13 products out of 1,344 hold 33.4% of every rating in the sample. The bottom half of the market — 672 products — holds 0.5% between all of them. That is the number worth carrying away from this page, and it is measured rather than modelled: no assumption about review rates goes into it, because it is a ratio of ratings to ratings.

It also means an “average” is close to meaningless here. The mean product in this sample carries 46 ratings; the median carries 3. Any Gumroad earnings figure quoted as an average is being dragged upward by a handful of listings that nothing else in the market resembles.

Does charging more mean earning more?

Listings grouped by asking price, in USD at ECB reference rates:

Free listings are the control that makes the proxy legible: 72 listings are priced at zero and 97% of them carry ratings, against 64% of paid listings. That gap is what the signal looks like when price is removed entirely — which is also the reminder that a rating means acquired, not paid for. The 34% with nothing is therefore not mainly a discovery problem. Free things here get taken.

The honest summary

A third of Gumroad listings appear to sell nothing. Most of the rest sell modestly. A small number sell a great deal, and they are concentrated inside categories rather than spread across them. If you are deciding whether to put work into a product here, the base rate matters more than any success story you have read, and the base rate is on this page.