How many products should you sell on Gumroad?

The short answer

Listing more products does raise a seller's total demand — slowly. It does not raise demand per product at all, and the sellers holding most of this marketplace did not get there by publishing a lot. Across 4,543 sellers, the rank correlation between how many products a seller lists and how much demand they attract is 0.283. That is a real relationship and a weak one.

Every seller, banded by catalogue size

“Ratings each” is the median of each seller's own ratings-per-product, not total ratings divided by total products — the latter is decided by whichever storefront happens to be biggest, and that is one seller out of 4,543.

Read the third column down and the case for publishing more looks strong: the median seller goes from 1 rating to 50. Read the fourth and it collapses. Ratings per product rises from 1.0 to 2.8 at five to nine products and then falls back; the sellers with twenty or more products earn 1.6 ratings per listing, which is worse than sellers with two. Whatever the larger catalogues are buying, it is not attention per item.

The last column is the part nobody mentions. Median price falls from $22.03 for a one-product seller to $8.51 for a twenty-plus seller. Large catalogues on this platform are not the same business at a larger size — they are a cheaper business.

The sellers who actually hold this market are not the ones with the most products

The top 45 sellers — the top 1% — hold 52.5% of all 204,971 ratings in this sample. Their median catalogue is 2 products. 14 of the 45 have exactly one. The largest catalogue anywhere in that top 1% is 21 products.

Meanwhile the six largest catalogues in the whole sample look like this:

Not one of them is in the top 1%: the 45th-ranked seller has 741 ratings, and the best-placed of these six sits at rank 82. The largest storefront measured, dinzee, lists 89 products which have attracted 96 ratings between them — 56 of those listings have never been rated at all.

One product carries almost everything

Of the 990 sellers here with more than one product and at least one rating, the median seller earns 75.0% of all their ratings from a single product. 82% of them get at least half their demand from one listing, and for 26.8% one product accounts for literally all of it.

That concentration relaxes with catalogue size but does not go away: across the 237 sellers with five or more products and any demand at all, the best product still takes a median 44.6% of the storefront. And among all 274 sellers with five or more products, the median one is carrying 3 listings with no ratings whatsoever.

What this does and does not license you to conclude

It does not say publish once and stop. Total demand does rise with catalogue size, the share of sellers with nothing at all falls from 43.5% to 10.0%, and a second product is the cheapest way to find out that the first one was the wrong product.

What it does say is that the common advice — volume is the strategy, keep shipping and the catalogue compounds — is not visible in this data. Nothing compounds here. A tenth product performs about as well as a second, the storefront's income stays concentrated in one listing, and the price a large catalogue can command is lower. If the aim is one product that works, the measurement supports looking for it. If the aim is thirty products that each work a little, no seller in this sample is doing that.

What the product counts are. A seller's product count here is what a category walk found — three pages deep per category, so it is a lower bound, not a catalogue, and it is biased down hardest for the sellers whose listings rank deepest. Worth being precise about which way that cuts: ratings and products are both counted over the same found listings, so undercounting a large seller's catalogue overstates their ratings per product. The finding that per-product demand does not rise with catalogue size is therefore conservative — correcting the bias would flatten it further, not reverse it.

Which sample this is, and why it is not the other one

This page is the only one on the site derived from the category walk — 8,322 products from 4,543 sellers across 204,971 ratings — because it is the only sample that records who sells what. Every other guide here is measured on a separate sample of 1,344 products drawn from 42 category searches. The two disagree on price by a wide margin ($36.99 median paid against $22.03 here) and both are published as they were measured. Averaging them would produce a third set of numbers describing neither, so this site never does.

Check it yourself

Every figure above comes from the seller CSV (4,543 rows, one per seller) and the listing table it derives from, with the derivation alongside them. Both are in the DOI-archived record, CC BY 4.0, no signup. The ranked seller index lists every rated seller.