Gumroad Revenue: The Failure That Makes $10M a Year

"I am now more focused on creating value than capturing it."
Sahil Lavingia, 2019

In February 2019, a founder published an essay titled "Reflecting on My Failure to Build a Billion-Dollar Company." It went straight to the top of Hacker News. The founder was Sahil Lavingia, the company was Gumroad, and the word "failure" was his own.

Seven years later, that same "failure" is a profitable business throwing off millions of dollars a year. The gap between those two facts is the most interesting number story in indie software, and almost every write-up gets it wrong.

Quick answer

As of 2026, Gumroad revenue (the fees the company actually keeps) was roughly $21 million in 2023 with about $9 million in profit, up from $11 million in 2022, according to estimates from the research firm Sacra. Back in 2019 Lavingia pegged annual revenue near $4 million. The headline figure people quote, "$142 million," is something else entirely: it is what creators earned selling on Gumroad in 2020, not what Gumroad made. The company Lavingia publicly called a failure became quietly, durably profitable by deleting its billion-dollar goal, not by hitting it.

The number everyone quotes is the wrong number

Gumroad is a platform that lets creators sell digital products directly to their audience, and it reports two very different kinds of money. The first is gross merchandise value (GMV): the total that creators earn. The second is Gumroad's own revenue: the roughly 10% slice it keeps from that volume.

Confusing the two is how you get viral posts claiming "Gumroad does $142 million." Here is the actual split, year by year, from Gumroad's own disclosures and Sacra's estimates.

Two honest caveats. The 2024 figure is a Sacra estimate and shows revenue falling back after the 2023 spike, so treat it as directional, not gospel. And the creator-GMV numbers ($142M, $178M) are exactly the figures that get misread as "revenue." They are not. A creator economy platform living on a 10% take rate earns an order of magnitude less than the money flowing through it. Lavingia said it best in 2019: Gumroad had "turned $10 million of investor capital into $178 million (and counting) for creators." That is a GMV sentence, not a revenue sentence.

The famous "failure" essay was written at the bottom

The reason the 2019 essay reads as a confession is that Lavingia wrote it near the low point, not the peak.

He had left his job as employee number two at Pinterest in 2011, before vesting his stock, to build what he assumed would be his first billion-dollar company. The money came fast: a $1.1M seed round from a who's-who of angels including Max Levchin, Chris Sacca, Ron Conway and Naval Ravikant, then a $7M Series A in May 2012 led by Mike Abbott at Kleiner Perkins. Total raised: $8.1 million.

Then growth stalled. Gumroad peaked in November 2014 and flattened. Without the growth curve venture investors needed to justify a Series B, the round never came. In early 2015 Lavingia laid off 75% of the company, "including many of my best friends," shrinking a roughly 20-person team down to a skeleton crew. TechCrunch published the layoffs, and he had to reassure worried creators in public.

By his own dashboard, June 2015 was brutal: about $89,000 in monthly revenue and roughly $350,000 in the red that month once severance and operating costs were counted. This is the moment the "failure" essay is really about. It was not a story about a company that died. It was a story about a founder realizing that "my only metric of success was building a billion-dollar company. Now, I realize that was a terrible goal. It's completely arbitrary."

The company returned to profit by June 2016. It has, by most accounts, stayed there.

The single biggest revenue lever was a price change

Here is the part the founder-worship posts skip. The thing that roughly doubled Gumroad's revenue was not a growth hack or a viral moment. It was a pricing decision.

For years Gumroad charged a variable take rate, somewhere between 3.5% and 8.5% depending on a creator's plan and volume. In early 2023 the company scrapped that and moved to a flat 10% on gross sales. Sacra's data shows monthly revenue jumping from about $1 million to $1.8 million almost immediately, and full-year 2023 revenue rising 96% to $21 million.

That is the cleanest illustration of a hard truth for any marketplace operator: on a take-rate business, the fee dial can move revenue faster than acquiring new users ever will. It is also the dial most likely to anger your existing base. Sacra's estimate that 2024 revenue slid back toward $12 million is consistent with exactly that risk, some creators leaving or selling less after the increase. The lever works, but it is not free.

As of July 2026, Gumroad's current pricing lists a flat 10% plus $0.50 per transaction on direct sales, 30% on sales that come through its Discover marketplace, and no monthly subscription fee. The take rate is the business.

How a "failed" company actually runs

The quiet engine underneath all of this is a cost base most venture-backed companies could never stomach.

Lavingia has been unusually public that he runs Gumroad with no full-time employees, no meetings and no deadlines, leaning on contractors and open-source-style async work. He has published the company's revenue, salaries and equity openly for years. That radical transparency is not a vanity project. It is what let a "failed" company survive years of flat growth on a tiny overhead, then quietly compound once the pandemic-era creator boom and the 2023 pricing change arrived.

The valuation tells the same story in reverse. When Gumroad raised from its own community of creators and small investors in 2021, it did so at roughly a $100 million valuation, about 9 times revenue per Sacra. A respectable outcome. Also a rounding error next to the billion-dollar target Lavingia had spent years chasing. He kept the company; he let go of the number.

What you can actually copy

Strip away the essay and the personality, and there are four durable operator lessons here.

1. Know which number you are quoting. GMV is not revenue. If you run a marketplace, the money that flows through you is a vanity metric; the slice you keep is the business. Investors, and honest self-assessment, care about the second one.
2. The take rate is a product decision. A pricing change moved Gumroad's revenue faster than years of feature work. If you operate a platform, model the fee dial before you build the next feature, and price in the churn it may cause.
3. A low cost base buys you time to be wrong. Gumroad survived a flat stretch that would have killed a company with a full payroll. Optionality comes from overhead you do not have.
4. Pick your own definition of winning before someone else picks it for you. The "failure" was failing at an arbitrary, borrowed goal. Redefining success is the entire plot.

For a founder who took the opposite of the venture path even further, see how Pieter Levels built a solo portfolio he refuses to sell. And for the other side of the coin, a founder who chose the clean exit instead of running the company for a decade, read how FeedbackPanda reached $55K MRR and then sold.