I pulled all 10,150 products on TrustMRR. There is no fast money.

September 5, 2026 · Data as of September 4, 2026

Late last October, Marc Lou launched TrustMRR, a leaderboard that connects directly to payment accounts like Stripe and only shows verified revenue. Soon after, a claim started going around indie hacker circles: more than half the products on the board have zero MRR, so most indie products make no money. The claim has data behind it. Counting by the leaderboard’s mrr field, 58.4% of records show zero. The commentary that came with it was mostly the same: look, most people are doing charity work.

I’ve been going back and forth on whether to build a product of my own, and I wanted to know whether that number holds up. TrustMRR has a public API. I got a key and pulled every record: 10,150 of them, 40 fields each. The rate limit is 20 requests a minute at 10 records a page, so the pull took about an hour. It’s the full dataset, so there is no sampling error.

The conclusion first. The share of products that have never earned a cent is 26.1%, not 58.4%. The 58.4% counts records where mrr is zero, and mrr only counts subscriptions. Close to half the money on the board is outside that field. 49.1% of products took in something in the last 30 days. 3.5% have subscription revenue above $5,000 a month and are still syncing, with a median age of a year and a half. Among products under three months old, 1% make $5,000 a month. The most robust finding in this dataset is that money comes slowly.

How the 58% was computed

TrustMRR gives each product three revenue fields: mrr, last30Days, and total. mrr is monthly subscription revenue, last30Days is actual receipts over the past thirty days, and total is lifetime revenue. The “58% make no money” arithmetic is simple: count the records where mrr equals 0. That’s 5,930, or 58.4%.

mrr only counts subscriptions. A product selling a one-time template shows 0 in that field no matter how much it sells. Marc Lou’s own ShipFast is the example: a pay-once Next.js template with $1,270,003 in lifetime revenue on the board and an mrr of 0. Gumroad is on the board too, with $7.14 million in the last 30 days and an mrr of 0. By this arithmetic, ShipFast and Gumroad are both in the 58% that “make no money”.

Putting the three fields together, the 10,150 records fall into five groups.

The 5,930 records with mrr = 0 are the last four rows combined. 1,222 of them took in money in the last 30 days, just not from subscriptions.

72.9% of products have earned something at some point, and 49.1% earned something in the last 30 days. But the bar for “earned something” is low. Among the 4,987 products with receipts in the last 30 days, the median is $169. Median lifetime revenue is $535. The top 1% of products took 76.1% of all revenue in the last 30 days. Most products make pocket money, and the middle is nearly empty.

Money outside subscriptions

mrr misses all non-subscription revenue. How much is that?

Site-wide, receipts over the last 30 days are 2.08 times the sum of mrr. Gumroad alone pulls that up a lot; without it the ratio is still 1.69, which means 40.7% of the money on the board is not subscription revenue.

The same holds at the product level. Of the 4,987 earning products, 1,222 (24.5%) have no subscription revenue at all, and another 1,295 (26.0%) take more than a fifth of their money from outside subscriptions. Computing the non-subscription share product by product, the median is 18%. The top is no exception: of the 544 products with more than $5,000 in the last 30 days, 85 have no subscriptions, and 60 of those make more than $10,000 a month.

If you only look at the MRR leaderboard, you miss about a sixth of the top products and a quarter of the earning ones. One-time templates, usage-billed APIs, marketplaces that take a cut: none of them are on that board. Part of the indie hacker fixation on subscriptions comes from how this is counted.

Age

What I actually wanted to know is what separates the products that earn from the ones that don’t. The data has few variables you can compare directly. Age is the first.

Of the 10,150 records, 8,621 have a usable founding date. Split into six age bands, the four outcomes break down like this.

The share that has never earned falls from 44% under three months to 7% past four years. The share that is earning rises from 43% to 69%, and the share making $5,000 a month or more from 1% to 19%. Among earning products, median receipts in the last 30 days track age too: $39 under three months, $120 at three to six months, $1,360 past four years.

The 353 products with mrr above $5,000 (stale records already removed) have a median age of 18.2 months, an interquartile range of 10 to 39 months, and 38.5% are more than two years old. Products that have never earned have a median age of 6.5 months.

There are 2,417 products that earned before and took in nothing in the last 30 days. 39.4% of them are 6 to 12 months old and 24.2% are 1 to 2 years old, 63.6% together. Only 5.1% are under three months, because three months is not long enough to earn a first dollar, let alone lose it. The death peak for indie products is between month 6 and month 24. That’s why the grey band in the chart is widest in the 1–2 year group.

Don’t read the chart as “hold on for two years and you have a 19% shot”. The products still on the board past four years are the ones that survived and chose to publish their revenue. TrustMRR launched on October 31, 2025, and 35.1% of the products on it were founded before that date. They registered with their outcomes already known. A product that died in 2023 doesn’t show up here. The shares in the chart describe the survivors and can’t be used as odds of success.

What a $5,000-a-month product looks like

Take the 353 products with mrr above $5,000 and data still syncing. Median age 18.2 months. Median active subscriptions 388. Median monthly price per subscription $37, interquartile range $15 to $111. Across all products on the site with subscriptions, the median monthly price is $13, so the $5,000 products charge about three times as much.

Of these 353, 35.3% had negative mrr growth over the last 30 days, and the median growth rate is 0. Reaching $5,000 doesn’t mean still climbing.

About three in ten of them are listed for sale on TrustMRR. Site-wide, 21.1% of products are listed, with a median asking price of $6,000 and a median multiple of 3.6x annual profit. 14.9% of products that have never earned anything are listed too.

AI

The other variable you can compare is AI. I did a rough classification from keywords in product names, descriptions, and categories; 49.7% of products are AI-related by that measure. By the platform’s own category field it’s 21.2%. Both definitions give the same conclusion. The table uses the first.

AI products are slightly more likely to earn a first dollar and slightly less likely to reach $5,000 a month. Both differences are under two percentage points, and smaller once you control for age. Among earning products, median receipts in the last 30 days are $161 for AI and $175 for non-AI.

I expected a clear difference in one direction or the other. There isn’t one. In 2026 AI is the default setting for an indie product, much like “mobile” a decade ago. The label doesn’t change the odds of making money.

Does building more products help?

73.9% of records carry the founder’s X handle, 5,697 unique people after deduplication. 1,045 of them have two or more products, covering 2,850 products. That lets me look at a question that comes up a lot: does building several products make it more likely that one of them works?

Among founders with four or more products, 78% have at least one earning and 14.1% have one making $5,000 a month or more. For single-product founders the figures are 51% and 5.0%. But that 14.1% took at least four products to get, which is under 3.5% per product, lower than the single-product founder’s 5.0%. Building more products raises the person’s hit rate while lowering each product’s. Founders with four or more products have also most likely been at it longer, and this table doesn’t control for age, so part of the gap is time.

Of the 243 identifiable founders with a product above $5,000 mrr, 11 (4.5%) have a second one. Their other products are earning 73% of the time, against a site-wide baseline of 49%. Someone who has built one winner has other products that survive more often, but rarely builds a second winner.

Among the 12 founders with the most products on the board, 9 took in less than $1,000 across all their products in the last 30 days. The other three are Rohan Gilkes (9 products, $161,284), Marc Lou (17, $83,486), and Herman Palmgren (9 products, all earning, $57,357).

Category and country

Categories are self-reported by founders, so there’s a fair amount of noise.

Productivity and Utilities together hold 894 products, and 5 of them have mrr above $5,000. Mobile Apps is the only large category that clearly outperforms: of 474 products, 69% are earning and 8% have passed $5,000. Most mobile apps connect through RevenueCat, and an app that has RevenueCat set up usually already has paying users, so part of that advantage is the barrier to entry.

Country is where the payment entity is registered, not the founder's nationality.

Stripe isn’t available in mainland China, so developers there usually collect through an entity in Hong Kong, Singapore, or the US. Only 11 records are tagged mainland China, and 17 Japan. I did a rough count of records with Chinese characters in the founder name or product description: 55, of which 52.7% are earning and 3 make more than $5,000 a month. Too small a sample to conclude anything.

US entities reach $5,000 at nearly twice the site-wide rate, and their earning products have the highest median receipts. The data can’t separate selling to the US market from being incorporated in the US.

37 records at the top are frozen

I added this section after the pull. The API doesn’t return a sync time, but each product’s detail page does. I opened the detail page for every one of the 390 records with mrr above $5,000, read the “last synced” time and the “expired” flag, and pulled a random 40 for a site-wide comparison.

Of the 390, 37 (9.5%) were flagged expired or hadn’t synced in more than 30 days. Those 37 add up to $6,726,998 in mrr, 37.4% of the site total.

Median sync lag is 0 days, and so is the 90th percentile; nearly every record synced the same day. The frozen ones are concentrated at the very top. The number one product, Stan, accounts for 19.8% of site-wide mrr on its own, its page reads Stripe API key expired, and it last synced on April 20, 2026. 36 of the 37 also carry frozen 30-day revenue, $6,152,078 in total, 16.4% of the site.

With those 37 removed, 353 valid records have mrr above $5,000, 3.5% rather than the raw 390 and 3.8%. 513 records have more than $5,000 in the last 30 days, 5.1%. Every top-end figure in this post uses the cleaned values.

There are a few other problems in the data. 455 records have mrr above 0 but nothing in the last 30 days; that could be live subscriptions that didn’t collect, or sync lag. 92 records have mrr above 0 but a lifetime total of 0, which is contradictory, so I put them in their own “inconsistent” group. 1,529 records (15.1%) have no founding date or an invalid one, and the age section leaves them out. The API returned 10,150 records while the site’s homepage shows 10,048 products. I couldn’t account for the 1% difference.

One more thing I got wrong myself. The API pages in alphabetical order by product name, and the first pages are noticeably richer: 7.0% of the first two pages have mrr above $5,000, against 2.5% site-wide. I started by pulling the first 40 pages as a sample, and my top-end share came out half again higher than the full count. If you pull this data, sample at random. Don’t take the first N pages.

If you only remember a few things

1. To see what a product earns, look at receipts over the last 30 days, not MRR. mrr only counts subscriptions. Half the money on the board is outside that field, four tenths even with Gumroad removed, and a quarter of earning products show zero there.
2. There is no fast money in the first six months. 1% of products under three months old make $5,000 a month; at three to six months it’s 3.6%. Median receipts among earners are $39 and $120. Budget for two years and don’t judge anything in the first six months.
3. The most dangerous stretch comes after the first dollar. 64% of products that went to zero died between month 6 and month 24. Only 5% died in the first three months.
4. A $5,000-a-month product looks like this: a year and a half old, 388 subscribers, $37 a month each, three times the site-wide median of $13.
5. The AI label doesn’t change the odds of making money; the gap is under two percentage points. Building more products raises the founder’s hit rate and lowers each product’s.
6. The shares on the board are a snapshot of what’s there, not odds. The sample is self-selected, since only people willing to publish revenue register. The platform launched on October 31, 2025, and products that died before that never appear. Three in ten $5,000 products are listed for sale, and a sold product leaves the board. Using 3.5%, 5.1%, or 19.2% as odds makes all three mistakes at once.

If you want to cite this, the safe version is: in the self-selected TrustMRR sample, as of September 4, 2026, 26.1% of listed products had never recorded a payment through their connected payment channels and 49.1% had revenue in the last 30 days; close to half the money on the platform is not subscription revenue, and looking only at mrr misses about a quarter of earning products; products making more than $5,000 a month have a median age of a year and a half, four in ten are more than two years old, and most deaths happen between month 6 and month 24.

On Marc Lou

You can’t write about TrustMRR without Marc Lou. He built the leaderboard, and the leaderboard is on it: $317,592 lifetime, $39,776 in the last 30 days. His 17 products total $2.99 million lifetime, and 12 are earning. 76.5% of that comes from templates, courses, and communities that teach people to build products; 99.4% comes from products sold to founders, and 0.2% from products sold to consumers. In the community this is often called selling shovels. His self-reported revenue and the verified figures on the board differ by less than 1%, and he rounds down. The metric he promotes is the one most favorable to his own products. The numbers are not inflated.

Reproduction

The data comes from TrustMRR’s public API, /api/v1/startups, which needs a Bearer token and returns 10 records a page. 1,015 requests in total, run at 18 a minute, under the official limit. Pulled September 4, 2026.

The freshness check covered all 390 records with mrr above $5,000 plus a random 40, reading the last-synced date and the expired flag from each detail page’s Markdown.

“Never earned” uses the strict definition: all three revenue fields at zero. Relaxed to lifetime total at zero only, it’s 27.1%. Age is computed from the founding date to September 5, 2026, at 30.44 days a month. AI classification uses keywords in name, description, and category, cross-checked against the platform’s category field.

Every share has its denominator in the tables. Products and founders that haven’t earned are only reported in aggregate, never named. Everyone named in this post publicly displays their own revenue. This post is not affiliated with TrustMRR, and its conclusions don’t represent the platform’s views.