Why Platforms Like Substack Won’t Make Sense for Much Longer

I think the future of Substack is self-hosting.

Or—more directly—I don’t think they have much of a future.

I’m sure you’ve heard about their struggles—the $585M valuation after failing to raise a Series C in 2022, the 14% staff layoffs, the major creator exodus in early 2024 over content moderation failures, and the fact that prominent writers like Casey Newton’s Platformer jumped ship to competitors.

Substack is showing all the classic warning signs. But I care more about the fact that people had to go to Substack in the first place, and that they’re now unhappy. To me it’s part of a larger truth about the platforms in general.

I’ve been blogging since 1999 and I’ve seen a few dozen popular writing platforms over the years, and have probably used nearly a dozen myself.

I think the reason these platforms are on a fuse is because they tend to toxify over time. Here’s a not-even-comprehensive list of the platforms and why they’re no longer around.

Historical Blogging Platform Failures

Posterous (2008-2013)

Why it failed: Acquired by Twitter, then shut down
Twitter acquired for talent, not product. Couldn’t compete with Tumblr’s growth. Shut down April 30, 2013. Peak: 4.5M monthly users.

Pownce (2007-2008)

Why it failed: Lack of revenue, couldn’t compete with Twitter
Too complex (tried to combine microblogging, file sharing, events). Acquired by Six Apart and immediately shut down Dec 2008. Only had ~100k users at peak.

Windows Live Spaces (2004-2011)

Why it failed: Low engagement, poor features
Microsoft auto-created accounts for all 500M Hotmail users but only 30M were active. Migrated to WordPress.com Sept 2010.

Open Diary (1998-2014)

Why it failed: Financial difficulties
Pioneered blogging but couldn’t afford servers. Shut down Feb 2014 after data loss. Relaunched 2018 as subscription-only.

Gawker (2002-2016)

Why it failed: Legal battle (Hulk Hogan lawsuit)
$140M verdict in lawsuit funded by Peter Thiel. Filed bankruptcy, shut down Aug 2016. Relaunched 2021 but minimal traffic.

Tumblr (2007-Present)

Why it failed: NSFW ban, cultural mismatch with Yahoo/Verizon
Lost 30% of traffic after Dec 2018 adult content ban. Sold by Yahoo for $1.1B (2013), later sold to Automattic for less than $3M (2019).

Vox (Six Apart) (2006-2010)

Why it failed: Failed to gain traction
Six Apart’s attempt at “neighborhood blogging.” Never exceeded 600k users. Shut down Sept 30, 2010.

Blogger (1999-Present)

Why it failed: Still alive but stagnant under Google
Acquired by Google 2003. Last major update 2020. Lost 90%+ market share to WordPress. Many features deprecated.

LiveJournal (1999-Present)

Why it failed: Sold to Russian company, user exodus
Sold to SUP Media (Russia) 2007. Servers moved to Russia 2016, triggering privacy concerns and mass exodus.

Xanga (1999-Present)

Why it failed: Failed to evolve, lost to Facebook/MySpace
Peak: 40M users (2005). Shut down 2013, raised $60k to relaunch as “Xanga 2.0.” Now basically dead.

Medium (2012-Present)

Why it failed: Repeated pivots, paywall confusion
Multiple layoffs: 2017 (50 people), 2021 (75), 2023 (unknown). Changed business model 5+ times. Major publications left platform.

Svbtle (2011-Present)

Why it failed: Failed to grow beyond niche
Started as invite-only “elite” network. Opened to public 2014 but too late. No meaningful updates since. Estimated <10k active users.

Ghost (2013-Present)

Why it failed: Struggles with mainstream adoption
Raised $5M+ via Kickstarter/revenue. Only ~2.5k paying Ghost(Pro) customers despite being around 10+ years. $200+/month pricing drives users to competitors.

WordPress.com (2005-Present)

Why it failed: Lost creators to self-hosting
While WordPress.org powers 43% of the web, WordPress.com has <0.4% market share. Creators graduate to self-hosted after hitting limitations.

Substack (2017-Present)

Why it failed: Creator exodus, content moderation crisis
Platformer, Casey Newton left Jan 2024 over Nazi content. Flat $585M valuation after failed Series C (2022). 14% layoffs June 2022. Top writers leaving to avoid 10% fee.

Beehiiv (2021-Present)

Currently growing (benefiting from competitor issues)
Founded by Morning Brew alumni. Raised $33M Series B at $192M valuation (Apr 2024). 100k+ newsletters, grew 10x in 2023. Flat pricing beats Substack’s %.

Common Failure Patterns:

1. Acquisition Neglect: Platforms bought for talent/users then abandoned (Posterous, Tumblr)
2. Financial Unsustainability: Can’t generate enough revenue to cover costs (Open Diary, Pownce)
3. Feature Creep: Trying to be everything, mastering nothing (Pownce, Windows Live Spaces)
4. Cultural Mismatch: New owners don’t understand user base (Tumblr, LiveJournal)
5. Technical Stagnation: Failure to innovate while competitors advance (Blogger, Xanga)
6. User Hostile Decisions: Changes that alienate core users (Tumblr’s NSFW ban, Medium’s paywall)
7. Legal/External Threats: Lawsuits or regulatory issues (Gawker)

So there’s this massive history of these platforms rising up to consume the users fleeing the previous failure. Only to then fall apart themselves years later.

A step back

Another way to make this point might be to ask why we needed these platforms in the first place? Why were there so many of these? And why did they do so well for so long?

It’s because individual writers couldn’t do one or more of the following:

• Host a blog where they could write
• Send mass emails without getting blocked
• Get eyeballs on their content

Well, it’s now a whole lot easier to do all three.

Creating a website that hosts a blog is fairly trivial for AI today, and it will only become more trivial in the months and years to come. So any would-be writer can simply have AI help build them a blog. And it won’t even have to look exactly like all the other pages, which is a problem on a site like Substack, who wants to maintain its own branding above that of the creator.

Sending mass emails is still a bit of a dark art, but there are services like Amazon SES that AI could also use to build this infrastructure fairly easily. Not to mention using any of the other third-party services that are designed to help people do this. It just becomes a matter of integration with those tools, which AI is really good at.

And finally, the benefit of a platform for getting eyes onto your content I think is diminishing as well. First of all, it’s not clear to me that anyone on Substack is just getting tons of traffic by virtue of just being there.

Second, I think sites like Google and other AI-powered platforms are going to get extremely good at finding the content jewels of what’s publicly available and raising them up to others. Basically, AI-powered content discovery. And I think that will have far more impact at lifting the best content.

We can just do it ourselves

The bottom line is, we’re just going to be able to do these things on our own.

The platforms tend to be extremely toxic and highly restrictive, and they don’t last long anyway. And the only reason they existed in the first place is because it was hard to do the things that they provided.

And now with AI, that is all going away.