Substack Isn't Dying — But It's Got an Infection

Substack has a problem. No, it is not “collapsing,” despite some guy claiming so. No, the would-be Burgermeister wasn’t right in his infamous “Open Letter” to the CEO. The fact is, Substack will be fine, insofar as a well-connected, multimillion dollar tech platform like it would ever really be in danger to begin with. This is a company with loads of investors, tons of elite media hookups, and a decent revenue model.

No, the problem is not existential, at least not for Substack. But it might be for me. Or even for you. You see, the problem at Substack isn’t a dealbreaker for Substack, it’s a crucible for small-scale creators. And yes, to a lesser extent, regular readers or audience members. It’s a low-scale infestation problem, like bugs at a motel. The business still makes money, the customers suffer. Only the infestation is bots, not bugs. Which is, in the context of a social platform and a media company, still pretty bad. Just ask X (formerly known as Twitter), LinkedIn, et al how they’re doing.

Somewhere out there, a slew of bots, trolls, and scammers realized that Substack was the next best place to attack. With inflated numbers, like follower/subscriber counts that don’t match activity levels. With phony engagesments, such as thousands of likes on Notes (or even full articles) that shockingly few people are actually reading or commenting on. Some of it seems coordinated, some of it seems downright random.

Substack is not unique in this. Like I said, it’s just catching up to a disease every platform before it has caught, at a moment (post-AI boom) when the incubation period got a lot shorter. This piece is not “Substack is dying.” It’s “Substack is fine, several of the people on it are not, and the difference matters more than the doomers think.” Time to get real and maybe break out the bot spray.

The good news, of sorts, comes upfront: Substack is stable. It has enough real people, including paying subscribers, to float the company itself and process payments to the around 5% or so of the creators monetizing through it. If that sounds low, consider that only 1% or less of creators on most platforms actually make real money. Even on Patreon, which is really just a payment site, the range is still about 2% to 5% making a decent income. So Substack is basically leading the pack on viability.

If the platform continues in its present form and trajectory, it’s likely that Substack the company will be profitable and that the most popular creators on the platform will continue to earn full-time income. Something drastic would need to happen for the whole thing to truly collapse. Compare with places like X, YouTube, and even Facebook. Hell, if we expand beyond places with direct and heavy monetization, look at Pinterest and Tumblr. The days of MySpace and Vine, where social platforms could truly die in a totalizing sense, are basically over for any pltform of sufficient scale and/or reach. The very language of death is therefore dead or outdated in itself.

The bad news: Most major businesses in our contemporary economic model, the Liberal Market Economy—I daren’t get too political or slopulist by invoking scarewords like “neoliberalism,” but it might be fitting in this context—depend on infinite growth and speculative value. The real economy, where small businesses sell goods and/or services to customers and succeed or fail based solely on their own quality+profitability, is simply not what a business like Substack is pegged to. I could argue it should be, (and I’ll talk more on why later down in this piece,) but currently that’s not the world we live in. We must deal with that disconnect first.

This means that businesses like Substack need to constantly show growth, rather than profitability or sound business strategy, to justify their existence at the upper echelon of our Liberal Market Economy society. Indeed, it also creates the feeling amongst the creators who use Substack (e.g. moi) that anything short of constant growth means death. That’s again where the doomsaying impulse comes from. Taylor Lorenz, and Katherine Dee are worried about their (paid) subscriber numbers staying flat or even slightly dipping, despite them each having more than enough to live off of. I don’t say that to pick on them, but for key context: we’ve been trained by the economic model we’re in to assume that there’s no value to staying steady. I want to challenge this.

Let’s go back to that bit about the real economy. In a normal market economics scenario, there are buyers and sellers, who trade currency, or time, or something else (and even a combination) for products and services. Good business is selling enough of something to earn a profit—where profit is the amount of currency/time/etc earned in excess of the cost to create the product or provide the service.

For a creator on Substack, this means selling a service—which in this case is the subscription to one’s publication—to customers (in the form of paying subscribers). It also means creating products, in the form of content (articles, videos, and so forth) which are effectively rented in an ongoing fashion via the subscription. It’s sort of like owning a gym, buying or building all the equipment, then selling access to the facility. Or owning a private library and selling access to the books. (Yes, that’s a thing.)

The owner of the resources, which is the proprietor selling access to them, needs to make enough profit to pay all the expenses associated with creation, acquisition, maintenance, and promotion of their wares. They also need that profit to live off of, as it will fund their own expenses like clothing, food, internet, shelter, transportation, utilities, and the like. This profit therefore needs to be sizable and stable. Notably tho, nowhere in this model of real economics is there a need for endless growth.

To be clear and to be fair, growth is necessary for any business at the start. And it will continue to be necessary until such time that a suffient profitability level has been achieved. But to be honest—and at the risk of sounding vaguely distributist or even welfarist—profit beyond a certain level is broadly unnecessary. Which means that prioritizing growth in itself, and viewing any lack of growth (even if profit remains constant) as a failure is actually an error. At least under analysis pegged to the real economy of buyers and sellers rather than growth speculators or rentiers.

Given this, I think Substack in general, and creators in specific, should be less concerned with providing or seeing growth and more concerned with providing value that yields profit. Buyers will pay for goods and services they see as worth it. They will stay paying, as ongoing customers (subscribers), if that value stays constant or goes up. Crucially, value is not growth. Value is quality, consistency, and care—or can be summed up in those words, for simplicity’s sake. I provide value. Substack should too.

Allegorically put, the business model of Substack is like that of a mall or bazaar. It provides the venue wherein entities (in this case creators) can sell their wares or rent their services. Stores live in malls, gyms, live in malls, unique business or activities live in malls. People come to Substack for stuff they can’t find anywhere else, or can’t find as much variety in smaller or more narrowly tailored venues. And Substack charges both the buyers and the sellers for the privilege of utilizing this unique location.

The charge is a 10 percent fee on subscription revenue. It doesn’t seem like much at first, but subscriptions amount to annualized gross revenue of roughly $450 million as of early 2026. $45 million in revenue is a decent chunk of change. Something like half of that is estimated to be eaten up by the cost of their staff, followed by infrastructure and hosting, plus other expenses. But still, 45 mill could be a totally sustainable number with the right operational setup. And it would still yield profit. So the numbers alone are fine. What’s currently a bit shaky is the value proposition itself.

You see, in this context, it becomes obviously essential for Substack to provide value in the form of a high quality location. No one wants to visit a mall that’s dirty, smelly, or full of pests. No one wants to frequent a social platform that has a similar infestation. No one wants to feel like visiting the mall or the bazaar is a (partially) negative experience they’re putting up with, even suffering through, just to get the goods or services they’re still interested in. They want to feel like visiting is valuable.

This is where the bot problem becomes such a, well, big problem. Bots on social platforms are like bugs at the mall. Maybe a small amount is inevitable and comes with just existing in a public-facing environment, but that amount should never become anywhere near the level where it hampers the experience for people in the location. Malls must have cleaning, and even pest control, in the budget as an essential expense that maintains good quality for their business and their customers (who are, again, both the regular people paying sellers for goods/services and the sellers themselves, whose sales the location captures a small percentage of).

How do we know bots exist in too high a number on Substack? Well, we can start by using other, similar platforms as a gauge. Major tech companies like Amazon (owner of Goodreads and Twitch), Alphabet/Google (owner of YouTube), Meta (owner of Facebook, Instagram, et al), Microsoft (owner of GitHub and LinkedIn), and X Corp (owner of X/Twitter) nearly all have estimated bot activity ranges of 10% or more on their social platforms. LinkedIn is especilly notorious, with the bots accounting for north of 25%, according to some assessments. I should note that most of these companies self-report bot activity lowers than what outside estimates indicate.

Users hate bots, trolls, and scammers. Many have affirmatively stated that they will avoid platforms which have a bad reputation for containing high levels of such. This has led to people avoiding certain places over others, not to mention struggles over perceived low quality to the median visitor or customer. Think of it like the difference between a highly rated hotel and a budget motel. Both exist, both make money, but we’d all rather stay at places that were nicer than not, assuming cost wasn’t the only factor and we had decent options available. Which it is online, and which we do here.

Like I said above, no platform is at real risk of fully dying out. Facebook is not MySpace, LinkedIn isn’t the classifieds section of a newspaper or the yellowpages/phonebook at a payphone. Still, bad reputation is bad for business. The value of a bazaar/location/platform lies in its ability to attract not only existing customers, but new customers. Not for growth’s sake in a vacuum, but to ensure that newer or developing sellers have a decent pool of people to reach. Growth as a quality indicator, not just an end onto itself. An important distinction here.

Similarly, the kind of customer or visitor matters as well. Someone putting up with a low budget location probably isn’t gonna have many options themselves, nor much in the way of currency to spend. All customers matter, but having a wider variety of customers, including plenty of higher ticket buyers who can afford to frequent multiple sellers, is essential for the health of a place. Substack doesn’t want to be known as the Facebook or LinkedIn of creator subscription platforms.

And Substack has real competition, in the more direct case with places like Patreon or X (which does increasingly monetize its users) or even with newsletter format alsorans like Ghost and Beehiiv. It needs not only to be a viable business, but a thriving and highly respected one. A mall people wanna drive out of their way to reach. A hotel folks want to stay at. A gym or library people are willing to pay top dollar to use.

Substack must be honest about what its own business model is and what it isn’t. I’m fine with calling it a social platform that sells creators on the idea of having “direct access” to their audience. I’m fine with them selling the public on the idea of a place to find better quality reading material, better conversations, and even build digital friendships via the emerging social media features. A gym where trainers and members can both get value is great. But a gym that stops being about fitness isn’t. The knock here is “stop selling people something you can’t reliably deliver.”

Substack is on the internet, it’s a tech platform, it has social features. It cannot pretend to be “not social media” and should maybe pushback, or ease up, on the self-flattering mythology of “social media for smart people.” Some folks go to the mall, or the gym, purely for recreation or vanity, not just high-minded ideals like fitness or health. Don’t sell us on high-key JournalismTM in Newsman sense, even if many creators might be journalists. Some folks are here to goof off, to shitpost, to play around. Embrace that. More of a YouTube or Patreon “we’re for everyone” vibe.

And please stop selling the median user on the idea of becoming a blogger. On Patreon, it sort of makes sense to sell people on starting their own creator page, tho even that company is easing up on such pitches to visitors. On YouTube, it’s ridiculous to pitch the average visitor on becoming a YouTuber. Most people aren’t trying to make videos, and likewise on Substack most users aren’t trying to write long-form content (certainly not at a regular pace). The ratio rule still applies: 90% of people are audience members, a separate 9% are heavily interactive, and the top 1% actually make cool stuff. The language of the platform, how it talks about itself and to its users, must match this.

Sellers want visitors primed to be buyers, not primed to be new sellers (and therefore competition). Buyers want to know what they’re getting, and that it’s good. This can be achieved by realigning Substack’s messaging and public-facing vibe to be more like a bazaar of existing wares than a billboard company trying to find entrepeneurs. “Come here to find great stuff, join great communities, and support creators.” Not “come here to be a creator.” The latter message isn’t necessary. The incoming supply of would-be creators already oustrips the demand for them, at least from the highest ticket buyers. So Substack shouldn’t exacerbate this with shortsighted messaging.

Substack should stop incentivizing the very stuff that makes bots and scammers want to come here, to proliferate. Substack rightly is, or at least was, different from other platforms. The founder effects and culture made it such that causing ephemeral engagements metrics (likes, shares, subs/follows, etc) for their own sake was less important than providing actual value. “If you build it, they will come” was actually true here. People didn’t write things just to go viral, they wrote them because they felt the writing itself was worth doing. The post needed to exist.

As Substack has introduced social features and embraced the idea of itself as a competitor platforms like X, Instagram, and TikTok, it has also encouraged people, either tacitly or directly, to chase the flash over the substance—the metrics (including growth metrics) over the virtue of providing value. Bad actors know they can game the system. They know the majority of Substack users are trusting and intrinsically valuable. But as more of these actors flood in, the signal to noise ratio gets worse.

If one pernicious scammer decides to use bots and other means to inflate the engagment metrics on their notes and/or articles (it’s easier and less suspicous to do it on the former, by the way), then the algorithm will surface more of their content, which will lead to more subscribers, and eventually paid subs. But if a bunch of people start doing that, and even the less explicitly malicious users on the borderline copy this strategy, we get a tragedy of the commons. Pollution of a paradise.

My argument, to be clear, is not “stop doing social media.” I think that ship has largely sailed either way (so changing course is a moot point), and I don’t hate the social featues to begin with. I just think Substack shouldn’t make the implicit or explicit value proposition ever even look like “you can join this place, go viral, and become a famous e-celeb.” Whether they realize it or not, that’s the feeling some of the bad actors have gotten. We have to change that narrative, and quickly.

Let’s go back to our working estimate that Substack’s bot/scammer problem is within the industry range of 10% or so. Let’s further stipulate that the number will never be zero, just like no mall or hotel will ever have zero bugs in it. The goal should be halving the current levels, and just generally keeping them as low as reasonably possible. The same way big facilities work hard to keep things clean and well functioning. All of that is realistic. So what does it all mean or necessitate?

Firstly, make it clear that Substack isn’t in need of more new content or creators. We have plenty. We certainly don’t want people coming here and setting up shop for the wrong reasons. Like a flea-ridden sleazeball opening up shop in a nice location because they think they can swindle some of the buyers within. The messaging must be “we’ve built this, please come and see.” Like YouTube, but for people who love writing. Anyone who wants to be a YouTuber already doesn’t need prodding, and the same goes for folks who want to write. The 90-9-1 ratio is locked. Accept it.

Secondly, keep the mall clean. Invest more effort to fighting bots, trolls, and scammers. (And not by rolling out “AI detectors” that people will just use to attack each other and/or as a pretext to read even less content.) I mean network-level blocking. Patreon moved past polite request files by partnering with Cloudflare to actively block models from harvesting creator portfolios. Ensure the system blocks malicious training scrapers while allowing search engine bots needed for creator discovery. Do better on account vetting and transparency, like making the creation date and overall activity of accounts more clear to onlookers.

Thirdly, set the vendors/sellers/creators (i.e. the people who earn subscriptions, and thus drive revenue) up for success. Make sure the algo is attuned to signals that actually drive sales (new paying subscriptions) not just nebulous and gameable/gamified “growth.” The mall uses a tenant mix strategy to optimize visibility, variety, and sales. It’s not just a billboard company saying desperately “please someone, anyone, come fill this space.” It’s smart location manager that recognizes what drives visitors/buyers to come once, come again, and actually spend money on a reliable basis. Often, that’s by emphasizing differentiation.

Fittingly and fourthly, lean into what makes Substack different as a platform and what makes certain creators different as sellers. Bots and scammers can fake cheap signals, but it’s harder to fake costly ones. And they’ll stop bothering if it feels like Substack isn’t rewarding low effort anymore. What does that look like? Maybe rewards for earning the most recommendations or referrals over a month or a quarter, instead of just earning the most subscriptions or getting the most likes. Maybe stop surfacing accounts that only post notes and rarely post articles, especially if their notes are attuned to go viral but not actually add value to substantive matters here.

Fifthly, and I can’t stress this enough, remember that the model is “people come, people subscribe (ideally paid), people stay.” That’s it. No more selling the dream of being a creator. In fact, another thing that might help is actively stop promoting content which is built around selling paywalled “How To” and “Growth Hack” slop. If people want to learn how to become a blogger, they can do so, but Substack as a platform should make all its algo’s and internal processes de-prioritize content that’s essentially trying to monetize a cottage industry of dream chasers. That stuff is bad for those people paying (the content is scam central), and it’s bad for the brand of Substack itself. Again, 90-9-1. Stop selling the 1% as if it’s the 90%, please! The mall doesn’t have stores in it teaching people how to be a store at the mall, right?

Sixthly, ban accounts more regularly for suspicious, bot-like, or scam-like behavior. We know Substack generally takes a more hands-off approach content moderation, and I respect that. But this isn’t content qua content or censorship of ideas, it’s safeguarding the platform from activity that’s fundamentally inauthentic. Edgy bloggers with extremist ideologies are one thing, new accounts getting thousands of likes on their messages? Or constantly posting messages below people’s notes and articles in hopes of promoting themself? C’mon, folks. We know the score here. And whilst we’re at it, how about fake-follower checkers, but native and trustworthy?

Yes, #1 and #5 are variations on the same idea. Cuz it bears repetition. Substack needs to clean house in terms of the bot/scammer problem but that’s impossible without also fixing how it frames its own mission. People flood LinkedIn with fake jobs and fake profiles because those are useful for advertising, brand/identity laundering, and so on. People post garbage content on X, Facebook, etc because clicks drive revenue. Substack is likewise incentivizing the very things it ought to be stigmatizing. We can fix this if we tackle it holistically. But right now, we’re aimed at the wrong things.

I know this was a long one, but I needed to make the case in detail. My approach is what separates pieces like this from the broader, and sloppier, catastrophist genre. I’m aiming to be concrete, to be factual, and to be actionable. I’ve explained what’s wrong with Substack’s current messaging (separate from its business model, which is fine), why we have a bot/scammer prevalence problem, and backed it up with numbers, methodology, incentives, et cetera.

Engagement isn’t evil, and growth isn’t greedy, but both can be corrupted, and then corrosive, if we start treating them as ends unto themselves instead of means to ends. The purpose of being a mall is to provide a venue for commerce, and to earn a percentage of sales. The purpose of growth is to reach a point where one’s business is stable and profitable. The purpose of engagement is to social proof that the creator is worth paying for. Marilyn Strathern once said, building off the work of Charles Goodhart, that “when a measure becomes a target, it ceases to be a good measure.“ This maxim became known as Goodhart’s Law, and Substack should follow it.

We need an incentive redesign. If current metrics reward inflatable numbers, change what gets rewarded. (Surface engaged-reader metrics over raw counts, maybe.) If “how to become a creator” type content is more successful than stuff that regular readers (instead of aspiring writers) actually want, there’s an alignment issue. If people start to think Substack is a place to make a quick buck instead of a community to join, we’ve forgotten what the model even was in the first place.

To close, my message isn’t “we’re not social media,” nor “burn it down and start over.” My advice is remember what we are and why we’re here. 1% are making cool stuff, 9% are social proof of that coolness, and the remaining 90% are here to soak it all in. Let’s get back to basics, and start being cool again.

Admd Newsletter. "Hustle Bro Culture Is Making Our Content Shallow." admdnewsletter.com, 2024. https://www.admdnewsletter.com/hustle-bro-culture-is-making-our/

Besedo. “Fake Accounts Are Getting Way More Common on LinkedIn (Data).” besedo.com, 2024. https://besedo.com/blog/linkedin-fake-accounts/

FraudBlocker. “LinkedIn Removed 200 Million Bots. Here’s What That Means for Advertisers.” fraudblocker.com, 2026. https://fraudblocker.com/articles/linkedin-bots

KARE11. “Bad Bots on the Rise: What We Found on Our Own Social Media Accounts.” kare11.com, 2025. (Cites Imperva’s 2024 finding that over half of internet traffic was non-human, 37% “bad bot”; includes Darren Linvill’s critique of Cyabra’s bot-detection incentives.) https://www.kare11.com/article/news/local/kare11-extras/to-catch-a-bot-social-medias-growing-problem-with-aritificial-intelligence/89-d2dcdcb9-59cd-4300-9d2e-ae1aefe3a7ce

Kim, Eric. “The Illusion of Likes: Fake and Misleading Engagement on Social Media.” erickimphotography.com. https://erickimphotography.com/the-illusion-of-likes-fake-and-misleading-engagement-on-social-media/

Marketing Dive. “Instagram May Have 95M Bot Accounts, The Information Reports.” marketingdive.com, 2018. (Ghost Data study: 7.9% bot share in 2015 → 9.5% in 2018.) https://www.marketingdive.com/news/instagram-may-have-95m-bot-accounts-the-information-reports/528141/

Ramalakshmi, and Venkatesh R. “Impact of Retail Tenant Mix’s Effect on Shopping Mall’s Performance.” Advances in Consumer Research, no. 5 (2025): 1127–1135. Department of Management, Krupanidhi College of Management, Bengaluru.

Sci-Tech-Today. “Substack Statistics and Facts (2026).” sci-tech-today.com, 2026. (Writer gross income $452M in 2025; Substack’s 10% cut ≈ $48M.) https://www.sci-tech-today.com/stats/substack-statistics/

Spider AF. “What Are Social Media Bots & How to Stop Them.” spideraf.com, 2026. (Cites the 2023 ANA “Made-for-Advertising” study: 15–21% of programmatic ad spend/impressions to bot-heavy MFA sites.) https://spideraf.com/articles/social-media-bots-what-they-are-and-how-to-protect-your-brand

Varol, Onur, et al. “Just Another Day on Twitter: A Complete 24 Hours of Twitter Data.” arXiv. (First academic bot-prevalence estimate for Twitter: 9–15% of accounts show automated behavior.) https://arxiv.org/pdf/2301.11429

Woolley, Samuel, et al. “Unpacking the Social Media Bot: A Typology to Guide Research and Policy.” arXiv. (Definitional variance across platforms — Twitter’s permissive automation vs. Facebook’s sock-puppet problem — and platforms’ reluctance to share data with researchers.) https://arxiv.org/pdf/1801.06863