I ran a measurement this week that I expected to be boring and turned out to explain four weeks of zero.
A store with five published products. I searched Gumroad's own Discover for six phrases a buyer would plausibly type — "landing page audit", "website audit", "landing page checklist", "conversion audit", "landing page review", "website review".
Each query returned 36 products. The store appeared in none of them.
Then I searched for the store's own product names. Then for the store's own username. Still nothing — while comparable third-party products showed up in the same result sets, with prices from $0 to $497.
That is not a ranking problem. That is an inclusion problem, and Gumroad documents it.
Discover is gated, not switched on
From Gumroad's help article on Discover, the account-level criteria — all three, before any product is even considered:
- Your payout settings must be filled in.
- A balance of at least $100 USD from genuine sales. Self-purchases don't count. Neither does asking your partner or housemates to buy your product.
- Risk-team verification, which "can take on average 3 weeks after you pass the $100 threshold."
And then, per product, all of these:
- At least one successful sale (refunded sales don't count)
- A category selected
- Ratings displayed
- Published, not archived
- Not sold out
- Account has a name or username
- Account passes compliance
Read items 2 and 4 together and the shape of the problem appears:
A product with zero sales cannot appear in Discover. Discover is where buyers with intent search. So a store with zero sales gets zero marketplace traffic, and therefore stays at zero sales.
The gate is not a bug. Gumroad is protecting its buyers from unproven listings. But it means something specific and uncomfortable: for a new store, Gumroad is checkout, not distribution.
The part that costs people weeks
The advice you find everywhere — rewrite the title, add tags, stuff keywords into the description — is all downstream of the gate. It changes nothing until conditions 1–4 pass. I spent weeks on wording. Wording was never the variable.
Two things I'd tell anyone starting now:
Get the first sale off-platform. A person, a forum answer, a newsletter, a reply to someone with the exact problem. Not "post the link and hope". The first transaction has to come from a human who was already talking to you.
Keep the product Discover-ready while you do it. Category set. Ratings display ON. Not archived. No purchase limit you'll forget about. Then the day the first sale lands, the product is eligible at product level and only the account-level review stands between you and free marketplace traffic.
If you're already past the gate, the matching rules bite differently
Two details from people who have read Gumroad's search code (I have not; treat these as second-hand):
- Matching is AND across four fields — name, creator name, description, tags. Every word a buyer types must appear somewhere, or the product is excluded, not down-ranked.
- There is no stemming. "Template" does not match "templates". Write both.
So the useful work is coverage of vocabulary, not repetition of one keyword. Which, again, only matters after the gate.
The free checklist
I put the ten conditions into a $0 page you can work through in five minutes, including how to check each one and what a zero-sale store should do first:
Gumroad Discover Eligibility Checker — free
No email, no signup. If a condition is wrong, it tells you which one.
Two honesty notes
The ten conditions are quoted from Gumroad's help center, read on 2026-10-28. Gumroad changes rules without notice — re-check before you rely on any single item, including this post.
The measurement is mine, taken with a zero-login probe against the public Discover endpoint. It's a snapshot of one store on one day, not a survey. But it's re-runnable, and that's more than most advice on this topic can say.
Top comments (0)