Every time a new reader arrives on Substack and gets shown a list of publications to follow, there’s a set of models deciding what appears — and Mike Cohen leads the team that builds them. Feed, onboarding suggestions, push notifications, category tabs: if a reader sees it, his team touches it. This week he joined me live to explain how those systems work and what that means for how you grow your Substack publication.
You can watch the full replay above, plus a breakdown of the conversation, including:
How onboarding recommendations actually get chosen — the pool is far bigger and far less curated than most writers assume (Mike put a number on it)
How the sequential model weighs your readers’ likes, restacks and replies — and why the same action doesn’t count the same for every reader
What Mike says a newer publication should actually do to become more visible in the recommendations network — in his words, “the real answer,” not the sexy one
What Substack’s feed is optimizing for — and how it’s deliberately different from ad-driven platforms
And one more thing: during the call, I handed Mike a “Christmas list” of data wishes — the features and metrics writers keep telling me they need. He mentioned his team has a hackathon coming up, so the door is open, my friends.
🗳️ At the bottom of this post there’s a poll where you can vote on which data piece matters most to you — I’ll be sending the results along to Mike.
The replay is free to watch. To read the full breakdown and vote in the Christmas-list poll, a free subscription is all it takes — and it means you'll get the follow-up when I share what comes of the list.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.