Getting a first Gumroad sale when Discover can't see you

By yoshibuild · Published 2026-10-10 · Last verified 2026-10-11

Gumroad's code only treats a product as eligible for Discover once it has a counted sale. Which leaves a practical problem: if Gumroad cannot send the first buyer, who does?

This is a checklist of where first sales might come from, with a note on how much evidence stands behind each item. The evidence is thin. The author has zero sales as of 2026-10 and cannot say any of this works. The goal here is to be honest about which parts are code, which are stories, and which are guesses.

The one thing the code supports

A product is only recommendable if, among other conditions, sales.counts_towards_volume.exists? is true (recommendations.rb, read 2026-10-11). That scope is successful purchases that are not fully refunded and not charged back (purchase.rb). It contains no price condition. Whether a $0 download reaches the "successful" state was not traced, so the free-product question is still open; the author plans a separate write-up on it. The gate is covered in the first-sale gate article. Until a product has such a sale, the first buyer has to come from a link someone clicked outside Gumroad. Everything below follows from that one point, and from nothing else in the code.

Channels, with honest effort notes

The effort notes are the author's estimates, not measurements.

Channel	What it is	Rough effort	Evidence
Niche communities	Facebook groups and subreddits where your buyer already asks for this kind of thing	Daily, short	One seller's own account, unverified (below)
A free answer to a searchable question	A video or article that solves the problem and links to the product	Hours per piece	None verified
Your own audience	A small list or followers who would care	Slow to build	None (the author saw no data)
Direct links to people you know	A message to someone who needs the product	Minutes	None, but it costs almost nothing
Gumroad's own tools	The subscribe page and email workflows	One-off setup	Gumroad help text; not a traffic source by itself

What the one detailed account says

The most concrete story found is a 2022 newsletter post by Hunter Bohm. He describes a free Notion template with a Canva cover image. He says it got 13 downloads on day one from two Notion Facebook groups, then over 100 in the first week, with about 30 minutes a day spent posting in groups and subreddits.

Three cautions. These are his own numbers and the author could not check them (the post URL was not recorded). Free downloads are counted as "sales" in that story, so it says little about paid products. And none of it involved Discover. It is evidence that people in a niche community will take a free template. It is not evidence that they will pay.

A plain plan, if you have no audience

This is a suggestion, not a tested method.

Pick one to three products you can explain in a sentence. Spreading effort over 40 products gives each one almost nothing.
Find two places where your buyer already asks about the problem. Read the rules first. Many groups ban links.
Answer real questions with real help. Link the product only where it is allowed and relevant.
Write one short piece that answers a searchable question and ends with the product link.
Set up the subscribe page and a welcome email so people who do not buy yet can still hear from you.
Ask each early buyer for a rating. Ratings also have to stay switched on for Discover, as covered in the first article.

What not to expect

A first sale does not put you at the top of Discover. After the first sale the product is only eligible. Rank then follows recent fees earned, which is covered in the ranking article. One $5 sale will sit far below products with months of sales.

Also, a search summary of the Discover help page says purchases by you or your household do not count (the page itself was not read in full). Buying your own product is not a shortcut. Do not do it.

Caveats

Everything outside the first-sale gate is anecdote or inference. The channel table is the author's own suggestion, not a ranking backed by data.
The Discover help text was read through search summaries only. Dollar thresholds in it are unresolved, as noted in the first article.
Discover also needs other conditions besides a sale (a category, displayed ratings, and account-level checks). The sale is only one of them.
The community channels carry their own rules. A post that breaks them can get you banned. That cost is real and is not in the table.

How to keep a useful log

If you try any of the channels above, write down what happens. Without a record, a week of effort looks the same as a week of nothing, and you cannot tell which channel moved anything.

A simple table is enough. One row per day, with the channel, what you posted, the link you used, and the number of visits to the product page that day. Gumroad's analytics can show page views. Use a different link for each channel if you can, so the visits can be told apart.

Record empty days too. A column of zeros is data. It tells you that the channel, the wording or the product did not work in that window. It does not tell you which of the three, so change one thing at a time.

Keep the sample size in mind. Thirty visits in a week is too few to prove that a channel works or fails. It is enough to tell you whether anything happened at all.

The author's 30-day experiment

In 2026-10 the author is preparing a 30-day test on their own store that follows steps 1 to 6 on a few products and logs every visit and sale, including empty weeks. No results exist yet. A case study will be linked here when there is data to report: [case study link, not yet published].