Every pricing guide for digital products opens the same way: charge more, you're undervaluing yourself, premium buyers are better buyers. It is confident, it is repeated everywhere, and almost none of it is attached to a number.
So we went and got the numbers. Gumroad publishes monthly product sitemaps — a public index of products, maintained for crawlers — and every product page carries schema.org structured data including the price. We sampled 2,201 real products across the five most recent monthly sitemaps and read what they actually cost.
The answer is not what the advice assumes.
The Headline Number
Of 2,201 products sampled, 2,036 were paid and 165 (7.5%) were free. Among the paid ones, the median price was $8.90.
That is the number worth sitting with. Not $47, not $97 — the two figures that appear in nearly every course about selling courses. The middle of the actual market is under nine dollars.
The distribution is steeper than most people expect:
- 32.0% priced under $5
- 25.1% between $5 and $10
- 19.9% between $10 and $20
- 15.6% between $20 and $50
- 3.9% between $50 and $100
- 3.5% above $100
Three bands — everything under $20 — account for 77% of paid products. If you are agonising over whether to charge $39 or $49, you are already arguing about a bracket that contains fewer than one product in six.
Why the Average Lies
The mean price in the sample is $29.94. The median is $8.90. The mean is three and a half times larger, and that gap is the most useful thing in this dataset.
It happens because the top of the range is extreme and thin. The 90th percentile is $45. The 95th is $85. The 99th is $300. The single most expensive product sampled was $5,000. A few dozen products in the hundreds drag the average far above anything a typical listing charges.
This is why "the average digital product sells for around $30" is technically true and practically useless. Nobody is at the average. Half the market is below $8.90, and the products above $100 are a different business with different buyers.
The .99 Myth
Charm pricing — ending in .99 — is treated as settled advice. In this data it is strongly conditional on price.
- Under $10: 27% end in .99, 37% are whole dollars
- $10 to $50: 11% end in .99, 65% are whole dollars
- Above $50: 5% end in .99, 74% are whole dollars
The pattern is clean and it reverses completely across the range. At the cheap end, .99 is common. By the time a product costs more than a takeaway, sellers overwhelmingly abandon it for round numbers.
Whether that is because round numbers signal quality, or simply because charging $7.99 for a $200 course feels absurd, this data cannot say. But if you are pricing above $50 and reaching for .99 because a blog post told you to, you are doing something 95% of the market has decided against.
What Doesn't Predict Price
We also tested a piece of advice that turns up constantly: that specific, descriptive, longer titles support higher prices. Grouped by title length, the median price was:
- 1 to 3 words: $7.00
- 4 to 6 words: $9.00
- 7 to 9 words: $9.00
- 10 or more words: $8.00
Flat. There is no relationship between how long a title is and what the product costs. The median title in the sample runs to six words, and stretching it to twelve buys nothing.
This matters because title length is easy to optimise and therefore attractive to advise. It is also, on this evidence, irrelevant to price. If you want to know what a title should do, it is worth reading why a product isn't selling — the title's job is to get the click, not to justify the number.
What People Actually Sell
Stripping filler words out of 2,201 titles, the most common terms were: guide, book, pack, coloring, printable, digital, template, planner, pdf, bundle, kids, prompts.
That list is a fair description of the low-ticket end of the market. It is dominated by documents and files, not by courses or coaching — which is consistent with a median under $10. Printables and templates are cheap to make, cheap to buy, and sold in volume.
It also explains the shape of the price curve. A market mostly made of printables and Notion templates will have a low median almost by definition. If you are selling in those categories, $8.90 is your neighbourhood. If you are selling a course, you are in the 7% tail and should price accordingly.
One more thing worth noting: the 2,201 products came from 1,608 distinct sellers, and only 90 sellers appeared more than once. This is not a market dominated by a handful of large stores. It is very long and very flat.
How We Collected This
Method matters more than the finding, so here it is in full.
Gumroad lists monthly product sitemaps in its robots.txt — a public index it maintains deliberately for crawlers. Its robots.txt disallows only /purchases/. Each product page carries schema.org Product structured data containing the name, description, seller and price.
We took the five most recent monthly sitemaps (April to August 2026), drew a deterministic random sample from each, and read the published structured data at four concurrent requests with a 180ms gap and an identifying user agent. Of 2,466 URLs attempted, 2,201 returned usable structured data; the rest were removed, redirected or unpriced. Every product in the sample was listed in USD.
The collection and analysis scripts are in the repository, and the charts on this page are generated directly from the dataset so a figure cannot drift from the number it illustrates.
What This Data Cannot Tell You
This is the part most "data-driven" posts skip, so we will be blunt about it.
There is no revenue in this dataset. None. It records what products are listed at, not what they sold. Gumroad renders sales counts and ratings client-side, so they are not in the page source, and we did not collect them. Any claim about earnings built on this data would be fabricated, and you should distrust anyone who makes one.
A price is not a sale. A $5,000 product listed once and never bought counts exactly the same here as a $5 product bought a thousand times. The distribution describes what sellers ask for, which is a genuinely different question from what buyers pay.
It is one platform. Gumroad skews toward independent creators selling files. Etsy, Whop and Kajabi would each produce a different curve — Etsy probably lower, Kajabi much higher.
Monthly sitemaps skew recent. These are products indexed in the last five months, so the sample leans newer than the platform as a whole.
What it does tell you is real: this is where the market actually sits. If you are about to price your first product, you now know the middle of the market is $8.90 and that three quarters of everything sells under $20. Price above that if you can justify it — but do it knowing you are choosing the thin end of the curve, not the normal one.
Frequently Asked Questions
What is the median price of a digital product?
In our sample of 2,036 paid products on Gumroad, the median price was $8.90. Half of all paid digital products cost less than that. The mean was $29.94, but the mean is distorted by a small number of products priced in the hundreds or thousands.
How many digital products sell for under $20?
77% of paid products in our sample were priced under $20, and 32% were under $5. Only 7.4% were priced above $50.
Should I price my digital product ending in .99?
It depends heavily on your price point. Below $10, 27% of products end in .99. Between $10 and $50 that falls to 11%, and above $50 only 5% do while 74% use whole dollars. Charm pricing is a low-ticket convention, not a universal rule.
Does this data show how much money creators make?
No, and we want to be explicit about that. This dataset records listed prices only. It contains no sales counts, no revenue and no earnings. A price is what a product is offered at, not what it sold. Anyone presenting listed prices as income is inventing the connection.
Does a longer product title mean a higher price?
No. We grouped products by title length and the median price stayed between $7 and $9 across every bucket, from one-to-three word titles to titles of ten words or more. Title length carries no pricing signal in this data.
Where did this data come from?
Gumroad publishes monthly product sitemaps in its robots.txt, and each product page carries schema.org Product structured data including the price. We sampled across the five most recent monthly sitemaps and read the published structured data at a polite request rate.