What are digital products worth? Pricing a first one at $3.99, $9 or $19

Three dollars and ninety-nine cents bought the first digital product I ever sold, a file a stranger downloaded at fourteen minutes past two on a Tuesday morning. That is what a digital product is: something made once and delivered as a file or a link, such as a template, an ebook, a prompt library or a music licence, so nothing is shipped or restocked and the only cost on each sale is the platform's fee. The receipt was the most useful thing I read that week. Not for the money. For what the fee line, the buyer's one-sentence email and the silence that followed each told me about the price I had picked. I have priced a handful of small products for the studio since, and every one of them landed on one of the same three numbers. This is the practical version of what changes between them, in money first and then in behaviour.

What does each price actually leave after the platform fee?

The fee is the first thing that changes, and it changes most at the bottom. Gumroad's pricing page lists the direct-sales fee as "10% + $0.50" per transaction for sales through a profile or a direct link, and a separate 30% per transaction when a new customer finds the product through its Discover marketplace. Other storefronts charge differently, and Gumroad vs Stan Store sets a flat monthly plan against this model. I am using Gumroad's because it is the one I run and the numbers are public.

The fixed fifty cents is the part people miss. At $3.99 it is an eighth of the price before the percentage is even applied. At $19 it is loose change. Ten sales at each price keep $30.90, $76.00 and $166.00. Put another way, a $3.99 product needs a little over five sales to keep what one $19 sale keeps.

Sales that arrive through Discover flatten that curve but from a worse starting point: 30% off the top leaves $2.79, $6.30 and $13.30. In my experience marketplace traffic is a small fraction of a first product's sales, so the direct rate is the one that decides the month. The point of the table is not that cheap is bad. It is that the same $0.50 is a different animal at each rung.

Who buys at $3.99, at $9 and at $19, and what do they expect?

Fees are arithmetic. Buyers are the part that took me longer to read, and what follows is observation from a handful of small launches, not a survey.

The $3.99 buyer does not read the page. They buy on the title and the first image, often within a minute of landing. They expect volume and completeness: a big folder, a long list, the sense that they got a lot. They almost never email before buying and rarely after. When a refund request does come at this price it is nearly always the same sentence: the file was not what the title said. The support load is close to zero. The cost is that they teach almost nothing beyond whether the title converts, because they do not review and do not reply.

The $9 buyer scrolls. They read the description, open the FAQ, and sometimes ask one question first. They expect a thing they can use the same day, and the refund reason shifts from "not what I expected" to "I could not work out how to use it". This is the first price at which the file that opens first matters more than the file that took longest to make. A start-here page and a clean folder are not polish at $9. They are the difference between a kept sale and a refund.

The $19 buyer compares. They have two or three alternatives open in other tabs. They expect a system rather than a collection, and they email: before buying to check it fits their case, and after buying when something does not. A refund at $19 costs the sale plus the half hour of correspondence around it. The support load is real, roughly one considered reply for every few sales, and that reply is where the best product feedback I have ever received came from.

The studio's sprint manual sets the three rungs of a product ladder as a ratio: an impulse price at the bottom, four to five times that for the middle tier, and five times the middle for the top. Read against real numbers, $3.99 and $19 are not rivals. They are the bottom two rungs of the same ladder, and $9 is the awkward space between: too dear to be an impulse, too cheap to be trusted as a system. It works when the product is one clear tool with one obvious use, and struggles when it is a folder.

A first product is priced to buy information, and the cheapest information for sale is what a stranger does in the thirty seconds after paying.

Why is a first product priced to learn rather than to earn?

Because at any of these three prices the first month's revenue is small enough to fit on a receipt, and the information is not. Ten sales at $3.99 keep $30.90 and answer three questions: does the title convert, does the file open cleanly on a phone, and which channel those ten people came from. Ten sales at $19 keep $166 and answer a different question, whether the product is finished, because those buyers write and say so. They just take longer to gather.

That is how I now choose between the three. When I do not yet know whether anyone wants the thing, the low price gets the answer fastest and at the least cost to the buyer. When I already know people want it and do not know whether it is complete, the higher price finds the buyers who will tell me. The middle price is for the product that is genuinely one tool, and for nothing else. At ten sales a month, a fee that only arrives with a sale also suits a test better than a subscription that arrives regardless, which is the trade-off behind what each storefront costs a creator selling one product.

What the first price is not is a verdict on the work. It moves. Raising a price later reads as confidence and costs nothing but an edit. Lowering one reads as a message, and buyers who paid the old number notice. So the first price is the one that makes the next decision cheapest, and the receipt at fourteen minutes past two was worth more than the $3.09 it left behind.

Sources

• Pricing · Gumroad