What Substack Doesn't Tell You About Your Subscribers (And Why That's Not an Accident)

I love Substack. I also don't trust it with my business. Here's why that distinction matters more than ever.

When the data raises more questions than it answers... and "Referral" is the only answer you get.

Recently, Elena | AI Product Leader shared a note that hit a nerve - and it’s not the first time we’ve seen a comment like this. (She was frustrated — her recommendations felt skewed, she wrote, like the engine was built to benefit the big publications and everyone else was just along for the ride). I left a comment agreeing with her, and then I typed an analogy that’s been rattling around in my head ever since.

Imagine you go wine tasting. You have a genuinely incredible afternoon. You engage with the sommelier, you talk to other patrons, you discover bottles you’d never heard of. You’re excited, so you decide to pick up a few bottles on your way out. And then you step through the door, and ten other wineries are right there, shoving glasses at you. You grab a few sips just to be polite, just to get to your car. You don’t know which winery they were from. You don’t remember what the wine tasted like. You definitely don’t have their phone number.

Are you even remotely interested in finding out who those wineries are or finding out about the wines you sipped?

That’s the Substack recommendation engine. And most of us, myself included, until I really sat with the data, have been counting those drive-by sips as subscribers.

The conversation that started this whole rabbit hole. Mia's note at the end says it all — almost everyone who unsubscribed initially came from a recommendation.

In this article:

• Let Me Show You What One Day Actually Looks Like

• What The Recommendation Engine Is Actually Doing

• The Platform Trajectory (Or: I’ve Seen This Movie Before)

• Been There, Done This… On Kit

• What You Actually Have (Versus What You Think You Have)

• The Reframe That Actually Helped Me

• So What Do You Do With This?

• Where I Landed

Let Me Show You What One Day Actually Looks Like

Wednesday. I’m back from a client call, checking my email. 15 new subscribers today.

Here’s what Substack tells me about where they came from:

Substack app, Direct, Referral (& referrer’s name), and Substack network.

That’s it.

That’s the whole picture.

Out of 15 subscribers, five came through “referral.” And here’s the thing I kept asking myself: which kind of referral?

Because there are actually two completely different surfaces where someone could have found me through another writer. The first: when a reader subscribes to someone else’s Substack, they’re immediately shown a screen with that writer’s recommended publications — a list they’re scrolling past in order to get to their inbox (and yikes, can that list get long? Why doesn’t Substack let us rotate these?).

The second: a writer might have a recommendations page in their navigation or footer, where someone who’s genuinely curious can browse, read descriptions, and actively choose to subscribe.

Those are two completely different people with two completely different levels of intent.

And Substack gives me one word for both: Referral. 🙄

I went digging. Substack’s own analytics help article… which, by the way, requires you to be logged in to access (their documentation isn’t publicly readable, which is a choice) — lists the available source categories: Direct, Email, Substack app, Substack network, Substack onboarding, Substack trackbacks, Direct to App.

No differentiation between opt-in flow recommendations and footer/navigation page recommendations. Not even a footnote. You get a bucket. That is genuinely all you get.

What The Recommendation Engine Is Actually Doing

I want to be fair here.

Substack’s recommendation feature has driven real growth for real writers. The platform’s own data says recommendations now drive 50% of all new subscriptions and 25% of new paid subscriptions platform-wide. For some writers, particularly those in the big, established niches where Substack already has a massive readership, the numbers are even more dramatic. A well-known product newsletter writer publicly shared that 78% of his subscribers came from recommendations.

So the engine works. I’m not disputing that.

What I’m disputing is the quality assumption that gets baked into those numbers.

When you subscribe to a newsletter, and you’re immediately shown a list of 10-15 other publications to also subscribe to, before you’ve read a single word of the newsletter you just signed up for… what’s happening?

You might recognize a name and click. You might click two or three just because clicking is easy. You might click everything to see what happens. You might skip the whole screen (which is pretty much what I do).

The point is: none of those clicks represent the same level of intentionality as someone who found you through a Note you wrote, searched for a topic and found your work, or heard another writer actually talk about why they recommend you. Discovery through the post-opt-in or check-out screen and discovery through genuine connection are not the same thing. And your conversion rate, to paid subscriptions, to offers, to any kind of actual business transaction, will tell you exactly that. If you have the data to look at it.

Which, for most of us, we don’t.

The Platform Trajectory (Or: I’ve Seen This Movie Before)

Here’s where I want to be really clear about something, because I don’t want this to read as a hit piece on Substack. I genuinely enjoy it here. The quality of the conversations, the writers (friends) I’ve connected with, the Notes that lead somewhere real… There is something legitimately different about the community on this platform compared to other places I’ve spent time online.

And….

Substack has investors now.

Real ones. Big ones.

In July 2025, Substack closed a $100 million Series C at a $1.1 billion valuation. Their backers include Andreessen Horowitz, BOND, and The Chernin Group. Let’s do some quick math: $1.1 billion on roughly $45 million in annual revenue is a 24x multiple. For context, the actual New York Times doesn’t trade at that multiple. That is not a newsletter company valuation. That is a social platform valuation. The bet being made — with $100 million of real money — is that Substack becomes something else.

And you can already see it happening.

• The Substack app launched in 2022.

• Notes launched in 2023.

• Video features.

• Live streaming.

• A TikTok-style short video feed.

• A TV app for Apple TV and Google TV.

• And then, in December 2025, a quiet pilot of something Substack is calling “native sponsorships” — a sponsored content program inserted into newsletters.

The co-founder who once publicly called the advertising model “busted” is now describing ads as “a recognition of new possibilities.” And when investors were asked why, one told the New York Times that “their creators have told them that they want Substack to support advertising.”

That’s the oldest script in the book.

It’s how Netflix rationalized ads after years of promising they’d never run them. It’s how every platform that launched on an anti-ad promise eventually got there. Because the math requires it. You cannot sustain a $1.1 billion valuation on 10% of subscription revenue alone, and the people who wrote those checks know it.

I’m not saying this makes Substack evil. I’m saying it makes them a company with investors. And companies with investors serve their investors. That’s not cynicism — it’s just business.

$45M in revenue. $1.1B valuation. That math only works if they're betting on becoming something much bigger than a newsletter platform.

One more thing the research surfaced that I want you to sit with: creators are reporting that a January 2026 algorithm update appears to have significantly reduced visibility for smaller publications — with some reporting 70-90% drops in new subscriber growth from Notes and recommendations. Whether those numbers hold across the board or not, the pattern is familiar. Easy, organic growth for everyone in the early days. Then, as the platform matures and needs to generate ad inventory and scale the metrics that justify the valuation, the algorithm shifts toward favoring bigger names.

We’ve watched this exact arc play out on Facebook Pages, on Instagram, on Medium. It always starts and ends the same way.

Been There, Done This… On Kit

I’ve been watching this particular song & dance for a long time. Back in the early days of email marketing, it was list swaps (which are a solid strategy - when done as an intentional strategy and the sender knows and recommends someone, as in ONE someone at a time), it’s kind of the original version of the recommendation engine. Cue the bro marketers… “Promote my crappy ebook, and I’ll promote yours, and then we’ll both have a big fat list to sell garbage to!”

I watched people build lists of tens of thousands that way. Impressive numbers, terrible businesses. Because the people on those lists didn’t sign up for you — they signed up for a free PDF about Pinterest strategy and somehow ended up on your list about copywriting. The hard sell never worked on them either. And I was never interested in that version of growth.

The mechanics changed. The incentive didn’t.

Kit (formerly ConvertKit) built a Creator Network feature on the same logic that Substack’s recommendations run on: newsletters recommending each other, subscribers gained on autopilot, everybody wins.

And for a while, it worked great.

Until creators started noticing that their open rates were sagging, their lists felt bloated with people who’d never engaged with a single email, and the subscriber numbers weren’t translating to revenue.

Eventually, after enough pushback from creators, Kit added a dedicated Subscribers quality tab to their Creator Network reporting. You can now see the average open rate, average click rate, and unsubscribe rate broken down by acquisition source — recommendations vs. everything else. They built that feature because enough paying customers asked loudly enough for it.

To their credit, they gave people the data.

But here’s what they didn’t do: teach anyone what to do with it.

Knowing that your recommendation-sourced subscribers open at half the rate of your direct subscribers is useful. Having a clear strategy for reactivating, segmenting, or letting go of those subscribers requires a whole other level of infrastructure that most email platforms — including Kit — make genuinely hard to build.

The energy here feels familiar.

And there’s a version of this question worth asking directly: does it serve Substack to show you that the subscribers you got through the recommendation engine are less engaged than the ones who found you through a Note you wrote? Or does it serve them to keep showing you the big number — because the more subscribers you have, the more “potential paid subscribers” they can point to when pitching their next funding round?

I’m not accusing them of anything.

I’m just saying the incentive structure is worth understanding.

What You Actually Have (Versus What You Think You Have)

Let me be really specific about what Substack gives you and what it doesn’t, because I think a lot of people are operating under assumptions that don’t match reality.

What Substack gives you: A subscriber count. Aggregate source categories (the buckets we talked about). Basic open and click rates at the post level. The ability to export your list as a CSV file. And a 5-star “activity” rating per subscriber that sounds useful until you understand what it actually measures.

Three stars out of five. What does that mean exactly? A rolling 30-day score based on email opens and web views — relative to your own list, not any external standard. And it doesn't count comments, likes, or restacks. So... not much.

That activity rating is a rolling 30-day score based on email opens and web views, ranked relative to your own list rather than against any external benchmark. So if your list is cold, your five-star subscribers might just be the least cold people in the room. It also doesn’t include comments, likes, or restacks — meaning the most visible engagement signals on the platform aren’t counted at all (at least not in the Substack metrics).

And here’s the kicker: because Apple Mail and most modern email clients block tracking pixels by default to protect privacy, a devoted reader who opens every single email through Apple Mail could show up as one star. You’re not measuring engagement. You’re measuring the narrow slice of behavior that Substack’s tracking can see.

What Substack doesn’t give you:

• Any way to tag subscribers by acquisition source

• Any behavioral segmentation

• Any automation based on what a subscriber has or hasn’t done

• Any way to know if the person who subscribed four months ago has opened a single email

• Any way to send a different welcome sequence to someone who came in through a Note versus someone who came in through a recommendation checkout screen

• Any way to identify who among your list is most likely to buy something

You can export your list. I want to be clear about that, you do own it, you can download it at any time, your subscribers aren’t held hostage. But export it and then what? If you’re not doing anything with that list in a real email marketing platform — one designed for segmentation, behavioral tagging, and automation — you don’t have a list.

You have a spreadsheet.

And this is where I want to say something that I think gets glossed over constantly in the “just use Substack” conversation: email marketing platforms have existed since before most Substack creators were in college (kidding, I don’t know if that’s true, lol). They exist for a reason that has nothing to do with sending blog posts. They exist because sending content and running a business through email are two completely different problems.

A real email platform lets you know that someone opened your last six emails but never clicked anything, and triggers a specific sequence to find out why. It lets you tag someone as “bought the $97 offer” and automatically exclude them from the pitch for the same thing. It lets you see that 3% of your list has opened every single email for a year and clicked every link — and that’s your group to make a higher-ticket offer to.

None of that lives on Substack.

It was never designed to.

Substack’s entire monetization model for creators is to get free subscribers and convert some of them to paid subscribers at $5-10/month.

That’s it. That’s the funnel.

There’s no native pathway to a $500 workshop, a $2,000 consulting package, or even a simple one-time digital product without pointing people off-platform. I know you can link to a store or paid products, which is great, but the price point expectations on Substack mean you have to go wide

Very, very, wide.

And while some writers have built impressive revenue through paid subscriptions alone — Heather Cox Richardson reportedly earns around $5M/year, meaning Substack collects roughly $500k of that — those are the outliers. They are not the model most people will replicate.

The honest math: if your business needs revenue beyond $10/month subscriptions from a percentage of your list, you need infrastructure that Substack wasn’t built to provide.

Again, I’m NOT slamming Substack or suggestions anyone leave (I’m not).

This is the full picture Substack gives you. Notes and Recommendations are in the same "Network" bucket. "Other" is a category. No further questions, apparently.

Here’s where to start taking back some control.

Any link you place in your footer, in a post, or on your own website can carry UTM parameters. A free tool like UTM.io lets you build and track those links without touching any code. You’ll know exactly which of your own touchpoints drove a subscription. What you cannot track is what happens inside Substack’s opt-in flow — that’s their server, their redirect, their black box.

But everything outside it? Yours to measure.

This is also why I’m in the middle of wiring up Bento (moved off of Kit, it’s half the cost and does a lot more), an actual email marketing platform — alongside my Substack presence. Not instead of it. Alongside it. Because discovery and relationship-building are different jobs, and they need different tools.

A tool worth knowing about for the research piece, that I’ve raved about already, is StackContacts by Finn Tropy. It’s a desktop app that pulls your Substack data into a local database and lets you query it using AI. It won’t solve the source attribution problem (that data doesn’t exist anywhere), but it gets meaningfully closer to understanding the behavioral patterns in the data you do have. It’s in beta and absolutely worth getting (for one publication, it’s only $19… once!).

The Reframe That Actually Helped Me

At some point in the last few weeks, something shifted in how I think about Substack. And I want to share it because it genuinely changed the feeling from anxious to clear.

Substack is a discovery engine with a great community.

That’s it. That’s the whole thing.

And that’s not a consolation prize — that’s actually quite valuable. The quality of the people I’ve connected with here, the conversations in the comments, the writers who’ve generously recommended my work to their readers, the Notes that go somewhere real... that’s something, and it’s actually kind of rare these days.

But a discovery engine is not a business foundation.

A business foundation is knowing who your people are, being able to reach them when you choose to, sending the right message to the right segment at the right moment, and having a clear pathway from “subscribed” to “bought something” to “bought something again and told a friend.” Substack doesn’t do that. It was never trying to.

The people who are going to get hurt — and some already are, based on the algorithm shift reports coming in from early 2026 — are the ones who built their entire business model on Substack subscriber counts the way a previous generation built it on Facebook Page likes or Instagram followers. When the platform adjusts its priorities (and it’s adjusting right now, in real time), they’ll discover that 40,000 subscribers and no owned infrastructure isn’t a business. It’s a number on a dashboard that someone else controls.

Customers vote with their wallets. Someone paying $7/month for your Substack is raising their hand and saying they’re interested in paying you. That signal is genuinely valuable. But it’s not the ceiling of what’s possible with that person. And without the data infrastructure to understand and act on that signal, you’re leaving most of the value on the table.

---

So What Do You Do With This?

I’m not going to tell you to leave Substack. I’m not leaving Substack.

But I am going to suggest three things worth thinking about right now, especially if Substack is a meaningful part of how you’re building your business:

Know where your real subscribers are coming from. Look at your source data with honest eyes. Notes, direct, Substack network, referral — they’re telling you something directional even if they’re not telling you everything. The people who found you on purpose are worth a different level of attention than the ones who clicked your name on a recommendation screen. Start paying attention to that distinction, even if the data is blurry.

Your list is only as valuable as what you do with it. You can export it right now. The question is what you do with it after that. Having a real email marketing platform running alongside your Substack — one that lets you segment, automate, and actually understand behavior — isn’t a “someday” thing anymore. It’s the infrastructure your business needs. (This topic deserves its own deep dive, and I’m going to give it one — including a full breakdown of what I’m building with Bento and why. Teaser: it’s not as complicated or expensive as you think.)

Build something you own. Not instead of Substack… in addition to it. Your website, your email list in a platform you control, your content that lives somewhere with no 10% cut and no algorithm deciding who sees it. Substack is where people discover you. The owned stuff is where the relationship actually lives.

Not sure where you stand? Start here. Copy this prompt and drop it into Claude:

I publish a Substack newsletter and I want to think through whether I need an email marketing platform alongside it. Here's my situation: [describe your subscriber count, whether you have paid subscribers, what you sell or want to sell, and what your business goals are]. Help me think through what I actually need and what would be overkill for where I am right now based on my business goals.

Where I Landed

Those 15 subscribers from Wednesday? Some of them might be incredible. Some of them might open every email I send for the next three years and eventually become clients, students, or collaborators. Some of them might never open anything. I genuinely don’t know yet… and that’s the point.

The story Substack tells — that growing your subscriber count is the same thing as growing your business — is the part worth examining. The number is real. What it represents is fuzzier than most people admit.

Platforms do what serves platforms.

That’s not a betrayal, it’s just business. Substack launched with idealism, and they may have meant every word of it. But they also have $190 million in venture funding, a $1.1 billion valuation, investors on their board, and a business model that only works if the platform keeps growing. Those facts don’t coexist with “we’ll always put creators first” forever. They never do.

The best thing you can do is know exactly what game you’re playing — and make sure it’s yours.

Where are you in this? Are you treating Substack as one layer of a larger system, or has it quietly become the whole foundation? I’d genuinely love to know. Drop it in the comments or reply to this email — either way, I’m reading everything.