Marc Lou: "Is SaaS Dead?" 758 Startups, Median Shrinking

Marc Lou asked if SaaS is dead. Then he posted the numbers.

On 27 October 2026 he wrote that he had looked at 758 startups on TrustMRR with more than $1,000 in monthly recurring revenue. The median one, he said, is shrinking by about 11% a year if the recent pace holds. B2C is down. B2B is flat. In his reading, momentum sticks.

The chart under the post is narrower than that headline. Total revenue growth is positive in July, August, and September. The soft number is the median startup. The sample is his. He owns TrustMRR. The post does not say so.

Revenue figures below come from Marc or from TrustMRR’s own pages. They are self-reported. TrustMRR calls the revenue verified. That word is the site’s own description of payment-provider data. View counts and follower counts are a snapshot from about 00:36 UTC on 28 October 2026, which is 11:36 in Sydney (AEDT).

What he posted

The post is “Is SaaS dead?” It went up at 11:49:01 UTC on 27 October, 22:49:01 AEDT. Marc Lou is @marclou. At the snapshot he had 418,276 followers. The post had 50,395 views, 404 likes, 207 replies, 34 quote posts, 13 reposts, and 157 bookmarks.

He wrote three points, then a sample line. All of them are his, drawn from his own platform. He did not publish the method or the code.

“I analyzed 758 startups with over $1,000 MRR on TrustMRR.com.” Then: “Their revenue over the last 4 months is slightly declining.” “That compounded decline would be about 11% per year.” “At a steady pace, that’s 45% degrowth in 5 years.”

On the split: “B2B revenue is holding up better than B2C.” “Median revenue growth was 0% for B2B (255 businesses), versus −6.2% for B2C (257).”

On momentum: “Existing momentum tends to persist.” “Among businesses whose revenue grew at least 5% from June to July, 54% subsequently grew MRR.” “Among those whose revenue previously fell by at least 5%, only 35% grew MRR.”

The close: “n=758 startups and $41.63M in revenue in the last 4 months.”

MRR is the unit he posted. The cut is startups he says were over $1,000 on TrustMRR.

Median down, total up

The attached image is a TrustMRR table. One row per comparison period. Matched startups, total revenue growth, and the median startup.

Figures as printed on the chart in Marc Lou's 27 October 2026 post. The matched sample changes every row.

The median column is the headline. June −2.3%, July −0.4%, August −0.3%, September −0.9%.

The total column moves the other way for three of the four periods. July +7.7%, August +1.2%, September +2.2%. June is the down month, at −2.1%.

The matched count rises with each row: 455, then 519, then 593, then 659. A changing set cannot be chained into one cumulative growth rate. Three of the four comparison periods show positive total revenue growth. That is the claim the rows support.

Counted on this page, from those four median cells: they average −0.975% a month. Compound that monthly rate for 12 months and the result is −11.1%. Compound it for five years and the result is −44.5%. His “about 11% per year” and “45% degrowth in 5 years” sit on that median arithmetic. This calculation is ours. He did not publish the method.

@meaynx wrote the September row in one line, at 20:14 UTC on 27 October (07:14 AEDT on 28 October): “+2.2% total growth vs a -0.9% median in september. a few winners carry the sum.” Eleven followers at the snapshot. A few startups can lift the sum while the middle startup shrinks.

B2B, B2C, and 246 left over

The published split is 255 B2B businesses and 257 B2C businesses. Added on this page, that is 512. The headline sample is 758. The other 246 are unexplained in the post.

Yuntao Zou (@TaoMachina, 85 followers, founder of Ventima) asked at 16:11 UTC on 27 October (03:11 AEDT on 28 October): “when is the $1K MRR filter applied?” If the filter is today’s MRR, a startup that has already fallen under $1,000 is gone from the set, which could make the yearly decline “the optimistic number.” He also asked where the other 246 are. That reply had 3 likes. No answer from Marc was found in the posts checked for this piece.

@alixnkew (24 followers) asked at 19:37 UTC (06:37 AEDT on 28 October) whether these are “the same 758 startups tracked all 4 months,” or whoever was listed in each month. New listings could drag the average. The rising matched counts are that question in the chart.

@guidedetective (13 followers) asked at 18:02 UTC (05:02 AEDT on 28 October) whether the sample “includes startups that quietly shut down or stopped reporting.”

The $1,000 filter’s timing is unstated. So is the B2B and B2C classification. A list that keeps startups still above the line, or still connected, can miss closures. A list that adds startups each month is a shifting cohort. No reply from Marc to these methodology questions was found.

The database is his

TrustMRR’s About page says the marketplace is “independently owned and operated by Marc Lou.” Fetched on 28 October 2026, along with llms.txt and the homepage, the site calls itself a “verified startup marketplace.” Revenue, in that description, “comes straight from the source” through connected payment providers, including Stripe, Lemon Squeezy, Polar, Paddle, and RevenueCat. TrustMRR says it calculates from provider API data, “not from screenshots, manual entries, or self-reported revenue claims.”

That verified line is TrustMRR describing TrustMRR. Seller descriptions, asking prices, and some other fields are still seller-reported. They sit apart from the provider revenue figure.

The About page carries a snapshot dated 24 September 2026: 3,000+ startups for sale, 15,000+ startups listed, 175 acquisitions in 365 days, about 200,000 unique monthly visitors, and a 3% closing fee split between buyer and seller. All of those are Marc’s and TrustMRR’s figures. He sells startup listings. The visitor count is his.

The analysis uses 758 startups over $1,000 MRR. The About page says 15,000+ are listed. He does not spell out every cut between those two counts. Public revenue benchmarks already separate an opt-in feed from a census. This post is an opt-in feed with a revenue floor.

The 27 October analysis does not mention that he owns the source.

The same afternoon he posted a TrustMRR promo video at 12:59:59 UTC (23:59:59 AEDT). At the snapshot: 22,452 views and 229 likes. At 13:11:03 UTC, 00:11:03 AEDT on 28 October, he shipped a MailCheap update. Snapshot: 29,391 views and 305 likes. The card’s pitch is MailCheap’s own claim: “83% cheaper than beehiiv and 75% cheaper than Mailchimp. From $19 a month.” How MailCheap started is the beehiiv story. The longer portfolio is a separate Marc Lou story. The short map is the Marc Lou page.

He was shipping and promoting in public on the same day. The analysis post still leaves the ownership off the page.

The week this chart landed in

The table arrived after a run of posts that were predictions. This page is reporting their sentences. The 11 October round, and who pushed back, is Levels saying “software is dead.”

Times are UTC. Sydney is AEDT, UTC+11, so afternoon UTC posts fall on the next calendar morning. Counts are the 28 October snapshot.

Jack Ellis (Fathom) described a chain: someone clones a product they think is too expensive, then someone else clones that clone. The original product is unnamed in the material used here. Reading it as beehiiv is an inference, and it is unconfirmed. In Sydney that post is 12:19 on 26 October.

jacob (@jsnnsa, founder of @spawn, 19,291 followers) wrote a long prediction at 14:53 AEDT. The short line: “the end state is no saas, no apps, no harnesses, no banks.”

Arvid Kahl, at 23:14 AEDT: “ICPs are moving upstream.” If $50 a month “is getting too expensive,” software businesses “definitely don’t want you as their customer.”

Pieter Levels quote-posted Jack at 23:47 AEDT. “if you understand basic economics the only one winning here is Big AI and Big Tech.” And: “It’s a race to the bottom for anyone else in software.” His public record is on the Pieter Levels page. The months he spent cancelling his own vendors are the vibe-coded replacement ledger.

At 23:53 UTC on 26 October, 10:53 AEDT on 27 October, Levels quote-posted jacob. “Great post and I think he’s right on a lot of it. The question is just how soon?”

Mario Zechner (@badlogicgames, libGDX, 78,112 followers) posted at 19:47 AEDT on 27 October. “software is dead. sad.”

Jack describes a copying chain. jacob describes an end state. Levels names who he thinks keeps the money, and asks how soon. Mario posts the slogan. Marc attaches a cohort and a chart. The chart answers a question about startups on TrustMRR over $1,000 MRR. His title asks a wider one.

The copying worry is the same subject as whether vibe coding is killing indie hacking: building got cheap, so a thin product got easier to copy. Arvid’s upstream line sits next to Marc’s flat B2B median and negative B2C median. A consumer seat being easier to drop than a company workflow is a reading people offered. It is outside the table.

Critics, a defence, and a reframe

Most replies are small accounts. The posts read for this piece were 22 quote posts and about 25 of the 207 replies. In that set, Pieter Levels, Arvid Kahl, Rob Hallam, and Tyler Denk did not appear. The check is a sample of a thread that was still moving.

The filter, the missing 246, the changing sample, and the quiet shutdowns are above.

@thetomlimb (338 followers), at 17:45 UTC (04:45 AEDT on 28 October), rejected the catalogue. “This is really not a good representation of SaaS. The products listed on your site are generally low quality.”

Paul P (@pavlos1944, 5,501 followers) wrote at 17:24 UTC (04:24 AEDT on 28 October): “Might be some bias in there (“Twitter SaaS” ?)” He builds micro-tools and has sold @low_fruits. The list is founders who chose to list, mostly online and visible on X.

@ads4apps (3,659 followers) wanted a higher floor, at 14:00 UTC (01:00 AEDT on 28 October). “Anyone can get to $10k/m by doing unscalable activities.” The real gauge, in that post, would be SaaS above $10,000 a month. The threshold is the replier’s.

@jayzhoupro (47 followers) defended the pipe at 20:48 UTC (07:48 AEDT on 28 October). TrustMRR revenue, the reply says, comes “straight from stripe, revenuecat and app store connects, not seller-reported.” That restates the product in a reader’s words. The direction matches the About page. The provider list is the replier’s, and it is a reply, which leaves Marc’s method still unpublished.

Adrian (@adriankuleszo, 148,704 followers) wrote at 21:59 UTC (08:59 AEDT on 28 October): “The money moved to AI more than it disappeared.” He said 70% of the projects his studio, @designme, delivered last quarter were for AI companies, mostly B2B. The 70% is his claim. It was unchecked against invoices, and it is outside the TrustMRR table. He sells design services to AI startups, so a shift toward AI clients is also his market.

Kiko Beats (@Kikobeats, 4,150 followers, ex-Vercel, builds web infrastructure) posted in Spanish at 13:25 UTC (00:25 AEDT on 28 October). Paraphrase, held short of a translation: fewer SaaS users, more power users, and a shift toward SLAs, security, and enterprise certifications.

@LuiyaIglesias (708 followers), at 18:36 UTC (05:36 AEDT on 28 October), kept the software and dropped the pricing unit. “saas isn’t dying, the seat is. nobody pays per seat when the seat is an agent.”

@jlcases (80 followers) read the B2B gap as who holds the cancel button. At 00:18 UTC on 28 October (11:18 AEDT): “someone in finance owns a B2B renewal. nobody owns a consumer cancel button except the user.” Marc’s post shows the gap. The reason is this reply’s.

@stevenwu_sg (33 followers) posted a Chinese-language summary at 13:50 UTC (00:50 AEDT on 28 October). It says commenters had pointed at total revenue growing in July, August, and September while the median fell, which reads as money concentrating at the top. It also says some replies mention the European summer holiday in July and August, and that four months is short. The holiday point is second-hand. The original comment was not found, so a summer pause stays unverified. Four months is a short window on its own.

Enrique S (@nrique_s, 1,724 followers) wrote at 20:14 UTC (07:14 AEDT on 28 October) that researchers and investors might want an anonymised aggregate, “like the IPI but for emerging SaaS.” Marc has not published that file with this post.

What the median can say

The median startup in this cut is drifting down. He put the compounded pace at about 11% a year, and at 45% over five years if it holds. B2B median growth in the post is 0% across 255 businesses. B2C is −6.2% across 257. Among startups that grew at least 5% from June to July, 54% grew MRR afterwards. Among those that had fallen at least 5%, 35% grew later. In this sample, a recent direction tends to repeat. He sized the sample at $41.63 million over four months, across 758 startups. All of that is self-reported from TrustMRR.

Total revenue growth was positive in July, August, and September. Winners can carry a sum while the middle startup slips. The people in the set chose to list on his marketplace. Replies called that Twitter SaaS. Startups that went quiet may be missing. The $1,000 line’s timing is unstated. The labels on B2B and B2C are unstated. 246 startups inside the headline count are unexplained. He owns the marketplace, sells listings, and charges a 3% closing fee. The post does not say so.

It is still a public look at very small subscription products whose revenue TrustMRR says is pulled from payment-provider APIs. The shape beside it, most listed products making little and a few making a living, is the argument on whether you can still make money indie hacking. The crowded generic shelf is the oversaturation page. This chart is one snapshot inside that shape. The middle of a self-selected board is soft. The total in three of four months is up. In the split he published, consumer subscriptions are weaker than B2B.

Levels, jacob, and Mario Zechner were predicting. This post is a dataset with the caveats attached. It leaves “is SaaS dead?” as a question he asked. The rows describe the median startup on one marketplace, above $1,000 MRR, across the four comparison periods in the chart, June through September.

What do you think the data says about indie hacking?

A note on the numbers

Post metrics, follower counts, and reply counts are a snapshot from about 00:36 UTC on 28 October 2026 (11:36 AEDT). They move. The revenue figures, the 11% and 45% lines, the B2B and B2C medians, the 54% and 35% momentum split, the $41.63 million, the About-page counts, the 200,000 visitors, the 3% fee, and MailCheap’s price comparison are Marc’s or TrustMRR’s. TrustMRR’s verified wording is its own description of provider-connected revenue. Adrian’s 70% is his. The −0.975% average, the −11.1% and −44.5% compounds, and the 512-versus-758 gap are this page’s arithmetic from the figures he published. No reply from Marc to the methodology questions was found in the posts that were checked.

FAQ