The Future of Substack

The New York Times is obsessed with Substack. A few months ago, it published Why We’re Freaking Out About Substack. Recently, they published Substack’s Growth Spurt Brings Growing Pains.

They’re right to be obsessed. Substack might replace them.

Today I’ll explain why that might happen. And since the lessons are applicable to all creator economy companies, we will also get a glimpse of the future of video, audio, streaming, and other creator economy verticals.

To do all of that, we need to start with yet another competitor to Substack and The New York Times: Medium.

In Platforms and Aggregators, I explained what aggregators are: companies like Google, Facebook, Youtube, Twitch… They all have the same thing in common: they aggregate supply and demand in one place, which they own, and make it really easy for demand to find supply.

Medium is trying to become the aggregator of articles.

Why are they trying to do that? Because aggregators can become massive: tens of billions of dollars in value, sometimes hundreds of billions or even trillions. About 70% of the value created by tech comes from companies like aggregators, fueled by network effects. The more supply they host, the more demand comes to see it. The more demand is there, the more suppliers come to find new demand for what they have to offer.

So what has been Medium’s strategy to become an aggregator?

The first step was a tool to easily publish online. Before Medium, it was really hard to do. Your articles either looked like crap, or it took too long to set up something like Wordpress. That’s why I started publishing my articles on Medium.

This tool attracted writers like me, who would then publish their content on Medium’s site. Free supply of content for them.

These same writers would post their articles on other social media to garner attention, attracting readers to Medium—free demand for them. The company would try to accelerate that, for example by optimizing SEO—making the articles easily searchable on Google. That was the second step. More readers for their writers, but also for themselves.

But if readers go to an article on Medium and then leave, Medium can’t grow much. The magic happens if they can keep them. And that’s the third step: the reason why Medium was created as a destination to discover content from other authors.

What The Hammer and the Dance looks like today. Notice on the right the recommendations of other articles from other authors, but optimized to be relevant to what the reader is consuming. Interestingly, as you scroll down, you can follow the article down, but the right column remains static, making it extremely invasive: “I’m in your face until you click on me”.

This would keep readers on Medium, jumping from one article to another, and associating good quality content with Medium, not the original authors who attracted their audiences.

If you look at the screenshot above, there is no upside for me as a writer to have these recommendations on the right. Quite the opposite: they take attention away from the reading experience, and they send readers to other authors instead of to other articles written by me. The upside is participating in an aggregator where my articles send an audience to other authors, and I get some in return.

The same thing happens at the bottom of the article:

If, as a reader, you go to the home page of Medium.com, you will barely see anything from the authors you follow. Nearly all the content there is from authors you don’t know.

Because Medium doesn’t want you to stay with the authors who brought you in.

One crucial element to remain a powerful aggregator is to keep the relationship between the demand and the supply for themselves. For example, a creator (= supplier) on Facebook or Twitter might have followers (= demand), but Facebook and Twitter control these followers. They can decide to reduce the impressions they see from the creators they follow, or ban them altogether.

You don’t doomscroll Instragram to see content from specific sources, but to see what’s happening. That takes power away from creators. You don’t try to find any website directly; you ask Google. You don’t go looking for websites for apartments: you ask Airbnb. You go to Uber for a new ride, you don’t look for the card you received two years ago from a cab driver.

Medium is the same. They don’t want authors to own their audience; they want it for themselves.

For example, I have 85k Twitter followers, but a standard tweet might only get 10k-20k impressions. I don’t have control over whether my followers see my posts. Twitter does. The tradeoff is that it gives me millions of impressions for my most viral posts. I get more new user acquisition by giving away control of re-engagement.

Authors, of course, want both: they want new audiences from the aggregators, but they also want to capture their own audiences to re-engage them. So they will attempt to find ways to create a direct link with their customers.

The only way to independently reach customers is to get direct access, which usually means getting their email addresses. So creators might create email capture forms.

This is an email subscription form inside a Medium article.

When I did that, the team at Medium asked me to take them down. The incentive to oblige? My articles were only eligible for promotion if they didn’t have these forms. And that’s the year my articles became the 1st- and 4th -most read articles on the platform. If I had not contributed as much to the platform, who knows what they would have done?

The issue there is not as much what they would or would not do. The issue is what they could do. If you don’t behave the way they want to benefit Medium, at the very least they will take away your traffic. At most, what would they do?

What Medium is trying to do, in summary, is to create a walled garden where they own all the readers, and the authors get some money in exchange.

Medium owns the relationship with all the readers, so authors can’t build a reliable audience.

Users who land on Medium from Google will read the article they clicked on, but then they’ll notice another article recommended and click on that. Then another, and click on that, never to stick with the author that wrote that first article where the user landed.

The readers who sign in to read articles give their email address to Medium, who then can send them emails to come back to Medium. Those who download the app get push notifications to come read new stories on Medium. The entire relationship is between readers and Medium.

Without a direct relationship with their audience, authors can’t build a reliable business. They’re at the whim of the platform and depend on getting viral hits that will give them enough impressions to make some money.

Usually, 40%-70% of the readers of my Medium articles come from outside of Medium. If this is true across authors, our content brings Medium half of its readers from outside of their ecosystem every day, which Medium can then hook. Each one of these is an opportunity for more customers.

The last step is to monetize readers. They do that by putting some content behind a paywall. Medium’s paywall, not the author’s. The reader never pays the author, they pay Medium, who then gives a share of that to the author. How much?

My first 23 articles on Medium were read on average 5k times each. The 24th one, How to Become the Best in the World at Something, got nearly 300k views. It earned me less than $5k. That’s a dollar for every 60 readers.

Each Medium article I wrote after How to Become the Best in the World at Something but before COVID netted me 25k readers on average. They brought me below $300 each. That’s after a very lucky viral hit. An author writing twice a week would make about $35k a year, in a very irregular way. This is not enough for most writers. Even the lucky ones like me, or the ones who are very serious, don’t make a living from it.

These are the earnings for a publisher on Medium. Source

Even serious writers will only take home between a few hundred and a few thousand dollars, with no reliability to base a career on.

The incentive for authors to put the content behind the paywall is not just the money: it’s also the distribution. Medium tells you that you can publish free articles on Medium, but then you don’t get a revenue share for it, and Medium won’t promote your articles. So Medium authors put their content behind the paywall and contribute to Medium’s growth. If you don’t want to do that, the only use of Medium is as a tool to publish nice articles on a site that attracts eyeballs.

Every business in the world needs to get users, keep them, and monetize them. For the creator economy—whether we’re talking about articles (Medium), music (Spotify), video (YouTube), streaming (Twitch), or anything else—this translates into:

A way to easily create and publish beautiful content (the content).
Get users: a way to get new users to consume their content (user acquisition).
Keep users: a way for creators to engage the users they acquired (engagement).
Bring back users: a way to re-engage the readers who have already consumed their content (re-engagement).
Make money from users: a way to monetize the content (monetization).

Every creator needs this.

Given that, this is what Medium did to capture as much power in this market:

If you want, as a creator, you can just use the tool to publish your content for free, which gets automatically optimized for Google. But that’s the extent of your freedom.

Then there’s Medium premium’s value proposition:

They help you get new readers through internal distribution, recommending your articles on other authors’ articles. For that, though, they need to be able to also use your articles to recommend other ones.
They help you re-engage with your readers through mobile push notifications and emails. But they also want to use these channels to recommend other authors, so they keep control of that channel: authors were actively discouraged from getting their readers’ email addresses.
They then help you monetize. But again, to keep the power, they are the ones with the paid relationship with the audience. You pay your subscription to Medium the way you pay it to The New York Times.

The key to this deal is that all of it goes together: you only get distribution and re-engagement if you put your content under the paywall, so it monetizes for Medium. Their leverage comes from the fact that you can never leave Medium, or you lose everything you’ve built on the platform, including your audience.

To illustrate this for you, with 48k followers on the platform, these are the views my last three test articles got:

In other words, Medium tries to be the Spotify of articles: you can have an audience, but you can’t control it. Pool your content with the rest, and we’ll expose it to readers (listeners) we control, and give you a share of the money we make.

The main difference between Medium and Spotify is that Spotify is public about the share it gives to musicians (about 70%), because at least in the music industry, most of the rights are owned by a handful of labels who can pressure Spotify. Medium doesn’t face that pressure, because there are millions of writers who don’t coordinate. So Medium’s only limit to how much revenue they keep is how much they think they can get away with.

That works until somebody comes along and gives authors a better deal.

Substack doesn’t want to be an aggregator. They don’t want to be the intermediary between supply and demand. Authors on Substack own the email list of their audience and the payments relationship. When a subscriber pays me through Substack, Stripe processes the payment directly to me. If I want to move from Substack to any other platform, I can simply take my email list and my Stripe account and move them.

What Substack is doing is taking the power away from readers and themselves, and giving it to authors. What does that mean? They are betting that in the market of long-form articles, the player with the most power is the author, not the reader. It’s the supply that matters, not the demand.

Why is that? Why would the supply have more power than the demand in the world of long-form articles? What makes them different from Twitter, Facebook, or TikTok?

Disintermediation: Taking an intermediary out of a relationship and going directly.

Most marketplaces don’t realize that they are ruled by an obscure dynamic: the balance between matching supply and demand vs. their long-term relationship.

Take Uber for example. Every time you want to take a ride, you need to find a driver that’s close by. What are the odds that you’re going to have the same driver twice? Very low. As a result, every time you ride with a different driver. Every time you look for a ride, Uber will match you with someone new.

Now compare that to Thumbtack. This company helps you find local professionals: photographers, painters, handymen… Imagine you hire a photographer and she makes amazing pictures. You’re going to exchange emails so that she can send you pictures. And next time you need a photographer, you’re not going to Thumbtack. You’re going to call her directly.

That’s disintermediation: once supply (the photographer) and demand meet, they won’t go through Thumbtack to interact again. They will bypass it, disintermediate it. In fact, the only situation in which the user goes back to Thumbtack is if the photographer was bad in the first place. The more successful they are at matching (which is the core of their business), the less they will be able to make money from recurring relationships.

Has an Uber driver ever given you a card and asked you to call him directly? He was trying to disintermediate Uber. If that has happened, what have you done with the card? Most people just throw it away.

Every marketplace is in a different position on this graph.

For example, Tinder is just about matching people. If Tinder is really good, it won’t see its customers return (since they will meet, fall in love, and not come back, they will disintermediate Tinder). Tinder only works as long as it fails at matching you with your soulmate, or if you’re not interested in committing to one single person.

Conversely, most of the time you use Airbnb or hotels.com, it’s to go to a new location, so you go back to these sites to discover a new place to stay every time you’re looking for that. If, however, you had a great experience and want to return, you might decide to contact the host directly (and disintermediate Airbnb).

So where do long-form articles lie on this graph?

You probably want to keep discovering new articles from new authors all the time. You want new learning experiences by matching with new authors. But once you discover a good one, you want to read more—or maybe even everything—from them.

Since there are a lot of long-term relationships, there’s a big risk of disintermediation. This is the insight that Substack is using: they think Medium’s position is untenable because readers want an ongoing relationship with their favorite authors. So Substack is helping authors to break free, to disintermediate Medium. If an author has to choose between owning their audience or depending on an intermediary, they will always try to own their audience.

This was quite a contrarian bet when Substack launched. Nobody thought so many readers would be willing to pay to subscribe to an author’s content. If you had asked me before Substack launched, I would have guessed that only ~0.5%-1% of free subscribers would be willing to become premium subscribers.

If it takes somebody ten years to get to 100,000 free subscribers, and only 0.5% of these become premium subscribers, that’s 500 people. At $10 per month, that’s $5,000 per month or $60,000 per year, exactly the average reporter salary.

Salaries for the top 20 most common jobs in the news industry in the US

But here’s the crazy insight that Substack discovered: readers are much more willing to support their favorite authors than people originally thought. Instead of 0.5-1%, the actual conversion from free to paid turned out to be 5-10%. An order of magnitude higher.

With these numbers, the previous author who gathered a 100k audience would now be making $600k a year, an extraordinary amount of money.

The readers’ willingness to pay for subscriptions demonstrates that Substack was right: readers want long-term relationships with authors, or at least to financially support them, which means they’re very likely to disintermediate an aggregator that tries to stay in the middle.

Here’s the issue: that doesn’t make Substack into a platform. It makes it into a SaaS tool (Software as a Business).

In The Magic Formula of the Internet, I explain how SaaS works. It’s not like an aggregator, where supply brings demand which brings supply. SaaS businesses don’t tend to grow exponentially like that. They need to painstakingly sell to every new customer, hoping that every one of them becomes better over time, and makes it easier to hire the next customer.

This is the best case scenario for a SaaS business: word spreads and every customer is easier to acquire, and when they’re acquired, every year they spend more and more.

Substack is a SaaS in that it gives authors tools so they can write articles, send them to their audience, and monetize these audiences through subscriptions. Substack takes a 10% cut of all revenue in return (probably much less than Medium).

The missing piece is the network effects. Without them, it can’t be a platform. More authors should bring more readers, which should bring more authors.

Unfortunately, by giving the power over email lists to the authors, they can’t easily cross-promote from one author to the other. By putting authors first, they relinquish control over the look of the articles, which means they can’t cross-promote from one to the other. Look at this article: it doesn’t have ads on the side for other articles. And as a reader, your main relationship is with me, not with Substack.

This has the additional disadvantage for authors that they can’t easily reach new customers through Substack. They must invest time and money in growing their audiences in other ways: posting their content on social media, advertising, cross-promoting with other newsletters…

So this was Substack’s quandary:

It gave all the power to authors because it knew they had the power.
But by doing so, it gave away the power to easily cross-promote them, which meant Substack was just a SaaS, without network effects. It didn’t become more valuable to authors as it grew. And authors were not tied to the company.

The only way for Substack to break that is by figuring out some sort of network effect.

Originally, Substack focused on creating great SaaS tools, so that authors would come and use them to write newsletters. These authors would then bring their readers.

The hope was that, among these readers, there would be some authors who would then start their own newsletters.

This is why Substack had this strategy of paying famous reporters with an established audience to come to Substack. The hope was that one year writing for Substack would establish the habit enough for them to stick around afterwards. And in the meantime, they would bring lots of readers to the platform.

As Ben Thompson writes in Sovereign Writers and Substack:

As [Matt] Yglesias told me via Slack (he stopped working as a Vox writer last fall but still contributes to Vox’s Weeds podcast), the deal he took from Substack is actually costing him money, for now. Yglesias says he has around 9,800 paying subscribers, which might generate around $860,000 a year. Had he not taken the Substack payment, he would keep 90 percent of that, or $775,000, but under the current deal, where he’ll keep the $250,000 plus 15 percent of the gross subscription revenue, his take will be closer to $380,000. As Yglesias puts it:

‘Substack is making money off of Slow Boring [Matt Yglesias’s Substack]. More money in fact than they would have made without the deal.

_But that’s what made it a good deal. They needed upside growth more than I did and I needed to minimize downside risk more than they did.’”—_Matt Yglesias via Ben Thompson

This was a great deal, a bit like a book advance. But instead the author would be building an asset: a relationship with customers, who are willing to read their articles and pay on an ongoing basis.

Except that, for Substack, the risk is much smaller than a book deal, because they knew the audience that famous reporters already had, so they had a sense of how much money they could make back. So this deal was very good for both Substack and for the authors it attracted.

But it was not enough: authors were good at bringing their audience to their Substack, but few of these readers became authors. More importantly, these audiences were not very useful to other Substack authors, so without one of these Substack deals, there was no real advantage for big-name authors to join.

And without a way to grow their audiences, Substack is just a SaaS tool, and authors can just go chase the next tool if it gives them a slightly better service.

All of this brings us to the only way Substack can become a high-growth platform with network effects that attracts and retains all types of authors: it needs a way to cross-promote readers from one author to the other, while maintaining the authors’ independence, so that authors need to come to grow their audience.

This would create true network effects:

This is how we should interpret the features Substack has been releasing recently, and how we can predict the type of feature they will release in the future.

Substack has always tried to highlight to authors how much value the network brings them. They even call it Network Effect:

Readers are more likely to sign up for Substack if they already have an email in Substack (don’t need to input it, it’s already there), or if Substack already has their credit card. How beneficial that is to the author is unclear. But the one piece that authors can’t get anywhere else is the Substack Platform Features. This is where they will put all their weight. Here are some features pushing that:

Mobile App

Substack could have created a white-label app: one where each author could have built and released their own branded app on the app stores. For example, the Uncharted Territories mobile app, powered by Substack.

This is not what Substack has created. They’ve released the Substack mobile app: a centralized app where you can see all your Substack articles in one place.

Authors might balk at the fact that Substack is trying to gain power from them.

But most of them won’t complain too much, because a Substack app is better for them than no app for their newsletter. With this app, readers can be reminded to read their articles in another way: through their Substack feed and, most importantly, through mobile push notifications. This is how Substack gets authors on board.

For Substack, this is good: authors are happy and stick around more. But I don’t think it’s their primary goal with the app. Instead, it’s this:

You can tap on the Search and discover new authors.

And I’d bet that if the mobile app works well, Substack will start showing in your main feed “Articles recommended for you”, from new authors, so you can discover them, sign up for them, and Substack can offer new user acquisition to their authors, therefore keeping them on the platform.

For most readers, though, this app is not very relevant: they receive the newsletter in their email inbox. Why do they need to go read them on an app?

This is the part that fails on this bet so far. If you receive a ton of newsletters, maybe having them all in one place is helpful. For most readers, this is probably not the case.

So Substack will have to keep iterating on the mobile app to push readers to use it rather than their inbox. For example, today if you click on an article, you’ll go to the mobile web version. Maybe in the future Substack will send you to the app first? Another example: today, commenting works better on the mobile app than on mobile web. Substack will probably keep iterating on these small usability improvements on the app so that it’s more pleasurable to read articles there, and readers use it as a default, thereby spending more time on the app and discovering new authors.

In the meantime, it sure pushes authors to recommend the app to their readers. They do it on the main dashboard:

Or when writing a new article:

Although the mobile app is a good idea, it’s nowhere as valuable as its Recommended Authors.

Recommendations

Substack recently launched Recommendations, “a new way for writers on Substack to recommend each other and discover more great work”.

This is the reaction from one of the biggest authors on Substack:

So apparently this feature is working for what Substack wanted: create network effects by moving audiences from one newsletter to another.

The problem, as we discussed before, is that authors don’t want to lose control of their newsletters, and are fearful of a platform holding too much power over them. So how did they solve the conundrum? It’s the author making the recommendations. Substack relies on a tit-for-tat where authors will recommend each other for mutual benefits.

Personally, I think this is a good idea, but I’m not sure it will deliver the results they want. The best authors will tend to receive more recommendations than they will give. That will increase their audience, but the rest of the authors, unrequited, might stop recommending. In general, authors might be reluctant to “give” more readers than they receive, so they might fall back to recommending fewer authors.

This issue of fairness will be problematic for this feature. Most authors will only want to give as much as they get.

If I were Substack, I would end up doing a standard recommendation feed. I would use recommendations from authors, but I would also use recommendations based on AI: Substack has much more data on readers’ favorite types of authors and articles than writers do. If somebody reads the same five authors as you, it’s likely you’ll be interested in a 6th author they also subscribe to. Authors don’t know this information. Substack does. And it could make these recommendations in a way that transparently shows the subscribers sent and those received.

Another thing I would try is an exchange, which gives authors as much value as they bring to the network. Other companies have already figured this out. For example, in videogames. Companies like Chartboost appeared as exchanges between videogames. You’d plug Chartboost as an ad in your game that would promote other games. A good exchange shows how much value sent into the network (how many new players into other games, how much money they generate), and how much value captured.

Let’s summarize.

Medium tries to take all the power away from authors. But they’re not in a position to do that, because authors have a lot of power: they can disintermediate Medium because readers care who their favorite authors are, and are willing to follow them.

Knowing this, Substack created a SaaS tool for long-form articles giving all the power to the authors. The problem with that is that authors can go to any other tool as soon as it is marginally better or cheaper. Some do, for example, to Ghost or ConvertKit.

So Substack needs some network effects to keep authors tied to the platform, even if they can leave at any moment.

The best way to achieve that is through cross-promotion tools that allow authors to constantly gain new readers that they wouldn’t gain otherwise. That makes their presence on Substack truly valuable.

So Substack will keep building tools to cross-promote readers from one author to another, the way they already have emails to discover new authors, the mobile app, or the authors recommendations tool.

But what’s the best way to optimize so many authors and readers? Bundles.

Imagine 100 readers are willing to pay for a subscription to Newsletter A.
Another 100 readers are willing to pay for a subscription to Newsletter B.
10 readers already pay for both.
There are 50 people that don’t find it worthy to pay for only A or only B, but would be willing to pay a subscription for A+B.

So if A and B bundled their newsletters, they would get their original 200 readers, minus 10 who were subscribed to both, plus 50 who are only interested in the bundle. That’s a total of 40 additional readers for the bundle, or 20 more per writer: an increase in 20% to their subscriber base!

Now imagine that instead of A and B, I got A, B, C, and D for that same price. It would be an even better deal, and even more people would sign up for that. Follow this to its logical conclusion, and Substack can become the Comcast or Spotify of long-form articles.

Substack would have the data to distribute audience and revenue proportional to the value contributed to the bundle.

The issue with this is that authors would lose their direct relationship with their audience. So this is not something Substack can easily pull off. My guess is that it will take some time to get there. They will first continue pushing recommendations. Maybe they’ll facilitate collaborations. And as authors warm up to the value of collaboration, they might introduce bundles as an opt-in option.

You might have noticed that this is what Medium is already trying: a bundle of all Medium authors. Unfortunately, they don’t give enough power to authors, so the direct path to the bundle probably doesn’t work.

Lately, Medium noticed the disadvantage compared to Substack and has started giving authors a way to gather the emails of their readers. But I fear this is too little, too late. It’s not Medium’s core business, and it has already lost a lot of credibility among writers.

It sounds to me like Substack is the one company well-positioned to propose the bundle; as an opt-in, with a serious commitment to author independence if they wish to have it.

The day that Substack proposes these bundles, they will be one step closer to becoming a more successful version than The New York Times, because Substack will attract the best authors since it will pay them according to their worth.

Share

I hope you’ve enjoyed this article! I have a few more to continue this topic:

What will star journalists do vs. other journalists in traditional media?
What are other aspects of Substack’s strategy, and what other features will they release?
How can we use these topics to figure out what will happen to the Creator Economy in general?
What features should every Creator Economy company consider?

Subscribe to the premium newsletter for this!