Such a fascinating article! re: your first footnote, I was under the impression the recommended publications are ones that that specific owner of that Substack recommends, not recommendations algorithmically generated by Substack (source: a friend subscribed to my Substack in front of me and I saw the publications I recommended pop up in the drawer after she did so). If so, that just goes to advance your argument in a way (or it could be twisted in Substack’s favor — if they hadn’t been on Substack, they couldn’t have been recommended — implying other platforms don’t have this recommendation feature, which may be true).
yea, there are actually two kinds of recommendations that i know of and was conflating for simplicity. the first is what you're describing, the second is publications that are featured on, say, substack's homepage or that otherwise come from substack itself. i was kind to substack and counted both of these as accurate substack referrals. though it would be much more accurate for substack to at least mention the name of the recommending publication.
As someone in the tech/startup world, 'gracefully' presenting analytics to paint a certain story is so absurdly common I'd be surprised if they didn't count everything they possibly could as ' from substack'.
It's not just tech analytics - science, politics, books, all pretty typically use data to lie, which is so dangerous because data is seen as hard evidence.
This is really interesting & helpful to know re: Substack referrals. I will note the argument in your first footnote is true for me, in part - I’m much more likely to subscribe to a Substack newsletter because I have access to their infrastructure (mainly, an app that redirects those newsletters to a designated place.) That’s why I have 100+ subscriptions. I think Substack conflates that with referral, as you discuss here, but knowing that your audience is Substack readers is a conscious effort on its part to keep you in this ecosystem.
I was one of the person that answered your e-mail. But I just realized that my answer was wrong. I thought that I subscribed via Substacks’s weekly recommendation e-mails. But for sure was a referral from Anil’s Dash piece on Rolling Stone.
thanks for this analysis! i do suspect the breakdown would look different for a large publication, though i imagine the types of misrepresented referral sources i'm seeing do scale up with audience size as well. doling out of dopamine hits is a good way of putting it.
Such a fascinating article! re: your first footnote, I was under the impression the recommended publications are ones that that specific owner of that Substack recommends, not recommendations algorithmically generated by Substack (source: a friend subscribed to my Substack in front of me and I saw the publications I recommended pop up in the drawer after she did so). If so, that just goes to advance your argument in a way (or it could be twisted in Substack’s favor — if they hadn’t been on Substack, they couldn’t have been recommended — implying other platforms don’t have this recommendation feature, which may be true).
yea, there are actually two kinds of recommendations that i know of and was conflating for simplicity. the first is what you're describing, the second is publications that are featured on, say, substack's homepage or that otherwise come from substack itself. i was kind to substack and counted both of these as accurate substack referrals. though it would be much more accurate for substack to at least mention the name of the recommending publication.
As someone in the tech/startup world, 'gracefully' presenting analytics to paint a certain story is so absurdly common I'd be surprised if they didn't count everything they possibly could as ' from substack'.
It's not just tech analytics - science, politics, books, all pretty typically use data to lie, which is so dangerous because data is seen as hard evidence.
This is an important topic and I am thrilled that you tackled it. Haven't seen anyone question Substack's network effects before. Thanks!
This is really interesting & helpful to know re: Substack referrals. I will note the argument in your first footnote is true for me, in part - I’m much more likely to subscribe to a Substack newsletter because I have access to their infrastructure (mainly, an app that redirects those newsletters to a designated place.) That’s why I have 100+ subscriptions. I think Substack conflates that with referral, as you discuss here, but knowing that your audience is Substack readers is a conscious effort on its part to keep you in this ecosystem.
I found this article by Googling "Does Substack lie about where readers come from?" I hope it cited Google as the referral source!
I was one of the person that answered your e-mail. But I just realized that my answer was wrong. I thought that I subscribed via Substacks’s weekly recommendation e-mails. But for sure was a referral from Anil’s Dash piece on Rolling Stone.
thanks! i've updated the data to reflect that
Sorry for that mistake. Internet Rabbit Holes are very difficult to track sources and remember things.
Anil Dash’s*
thanks for this analysis! i do suspect the breakdown would look different for a large publication, though i imagine the types of misrepresented referral sources i'm seeing do scale up with audience size as well. doling out of dopamine hits is a good way of putting it.