I fully admit that I did not see Elon Musk’s last-second decision to finally complete his push to buy Twitter coming. Maybe he’s resigned himself to the fact that agreeing to do it is better than going to court. Maybe, like Saul Goodman, he wants to avoid the courtroom at all costs.
As I put it very starkly in my April piece on the topic (which means that holy hell, we have been talking about this for six whole months), I don’t think Elon Musk has the right temperament to manage a site like Twitter with an understanding of the degree of safety it needs. He is an impulsive decision-maker, and impulse does not lead to strong policy.
But ultimately, this is happening because we made a bet more than 15 years ago that it was better to have our social media managed centrally, despite the fact that we had plenty of good federated examples out in the open.
The biggest of those is email, and I think that while I have to mention Elon Musk in this piece because it’s important, I’m just as concerned about other open platforms becoming less open. And I think that Substack’s focus on building network effects in email, while good for them, is bad for newsletters in general.
I think the reason for that is because of the same things that frustrate many of us about social networks like Twitter. It’s not that the people who run Substack are necessarily bad people, but their goal as a VC-backed company is to grow large and dominant, which puts it against the broader goals of the protocol on which it built its service.
That means that, when evidence emerged that the company’s model wasn’t “sticky” enough, the company had to do something about it, first by building apps to raise its own content above that of outside creators, then to find ways to help strengthen the in-network effects.
As my pal Simon Owens put it in his newsletter this week, “Substack found its unfair advantage” in the form of its recommendations tool. A quick blockquote from Simon:
One of the great things about Substack’s Recommendations tool is it introduces network effects without the inclusion of algorithmic favoritism. Remember, Substack’s central ethos is that it establishes a direct connection between writers and their fans; any launch of a Facebook-like algorithm would be in direct betrayal of this ethos. What’s genius about Recommendations is it allows Substack to walk that thin line perfectly.
Here’s how it works: every Substack writer is prompted to recommend other Substack newsletters, and they’re even given the opportunity to write small blurbs about why they recommend them. Then, every time someone signs up for your newsletter, they’re brought to a landing page that shows them all of your recommendations. If their email address is already verified with Substack, they can simply push a button to sign up for all the recommended newsletters. It’s basically an old-fashioned blogroll on steroids.
This is obviously good for Substack as a business—writers that might have used Substack as a launching pad now have an important reason to stick around beyond simply the revenue model. But the thing is, this is putting the company on the road to insularity—a bad road to be on.
I think if they’re going to be a primarily email-centric platform, what they do ultimately has to raise email as much as it raises Substack, because email as an ecosystem will not benefit if the largest player leverages its position to push up its own newsletters at the cost of the rest of the ecosystem.
It is something that helps to encourage network effects that favor Substack as a company at the cost of the newsletter medium as a whole. It is a step closer to centralization, and might encourage readers to ignore non-Substack newsletters simply because the onboarding process is not as convenient as what Substack built for its own ecosystem.
I think in many ways, decisions like these early on helped to bring social networks like Twitter, Facebook, and Instagram much of their growth, but at the cost of the broader Web. I think that, while Substack does not control the email inbox, it does have deep influence on a key part of the ecosystem, and by not making recommendations open to outside publications, it eventually raises the walls of the walled garden that everyone complains about when they’re already inside.
It is still early. Substack is still a relatively small service with a fairly focused niche. But this is a real risk, and it’s one I think about even harder given the Twitter news.
The motivation for building a platform on the internet should not be to close ranks within a protocol at the cost of everyone else. It should be to find ways to bring your benefits to everyone. Substack is not doing that with the Recommendations feature. That makes the feature problematic in the long run.
I will call this out whenever I see it.
Time limit given ⏲: 30 minutes
Time left on clock ⏲: 1 minute, 18 seconds