What Medium needs to do to survive

“Medium is in trouble.”
“Medium is in trouble.”
Image: Medium
We may earn a commission from links on this page.

There is a need for platform such as Medium. Its only way up is monetizing quality using a two-sided, paid-for system.

I am a huge fan of Medium. Notwithstanding the recently announced layoffs, I remain optimistic about its survival. Here are my reasons for such hopefulness, and suggestions for the platform’s future.

First, for elegant text-based publishing, there is a need for a simple, easy-to-use, well-designed platform such as Medium. Wordpress was supposed to deliver just that, but it took a geeky turn, saturating its ecosystem with scores of third-party plugins—more than 48,000 at last count—whose quality can charitably be called uneven. Most Wordpress sites end up using dozens of plug-ins, each one bound to create its own set of problems: slow page-loads, crashes, random behaviors or update cycles that don’t match WP’s platform agenda. Unless you have sizable tech resources at your disposal, Wordpress is a nightmare. Switching to Medium gives you the impression of going from MS-DOS to iOS—even if some features are missing, I’ll come back to that very point in a moment.

Now, after less than five years of operation, Medium is in trouble. This is not a surprise. The platform went through a classic Silicon Valley chain of events: A prominent founder, Ev Williams, co-founder of Twitter, comes up with a compelling story; afraid of missing The Next Big Thing, investors buy en masse. Altogether, Medium backers coughed up $132 million in three rounds. All big names are on board: Andreessen Horowitz, Greylock Partners, Spark Capital, and Google Ventures. With this much money, the company could spend a lot now, and figure out the business model later.

The time of reckoning finally came last week. No one with a working knowledge of the publishing business was genuinely surprised by Ev Williams’ announcement. In short: laying off a the third of the team (50 people out of a staff of 150), with an unclear pivoting and soul-searching message.

What is wrong with Medium? Three things.

First, there is the naïve belief that the quest for editorial quality—which Medium achieved pretty well; it remains one of the richer publishing platforms for a variety of topics—will bring money through quality advertising. In reality, there is absolutely no correlation between editorial quality and the revenue it brings; I addressed the issue at length here.

And that is only one failure of the advertising model. Can it be remedied? Yes, I believe so. (Here, I’m preaching for my own church: reconciling editorial quality and economic value is at the core of my work here at Stanford during my time at the John S. Knight Fellowship, as discussed in previous Monday Notes).

The road to business model salvation might prove to be a long one, but I’m convinced the advertising community will inevitably come to the conclusion that quality breeds (a) higher engagement and, (b) better demographics, both bound to bring in premium advertisers.

Medium’s second mistake was the idea that its editorial (authors, publications) and notoriety were in themselves appealing enough to foster the kind of high yield advertising generated by The New York Times T-Brand Studio or The Wall Street Journal Customs Studios. It could have worked, but it didn’t. The key reason is that, to really work, native ads require large commercial operations as explained previously here.

The last—but not least—reason for Medium’s current difficulties is its publishing structure, the way the content is organized. To put it mildly, Medium is an elegant mess—unearthing interesting stuff requires assiduous digging. Since it is super easy to publish on Medium, the whole platform is plagued by a thick layer of noise. Admittedly, substantial improvements were made to both the site and the apps—even if the smartphone app misses some essential features like the ability to read articles offline.

I already looked at Medium’s peculiar publishing structure in two previous Monday Notes, here and here.

Last year, Medium’s main breakthrough was the launch of its publications system, which happens to be far more accessible via a smartphone than on the web. Looking at the numbers, it confirms a Pareto-like structure with a small number of items attracting the bulk of the audience. I used a Medium leaderboard, compiled by a gentleman named Mubashar Iqbal, that tracks 344 publications. For the 276 that have more than 1,000 followers, it goes like this:


Startups     1,582,984     23
Tech           703,723     26
People         516,567     12
Design         487,730     29
Politics       420,793       9
News           303,917       6
Business       273,336     12
Marketing       164,810       2
Development     97,564     13
Photography     59,123       3
Movies           56,252       3
Education       55,670       2
Travel           48,575     3
Venture Capital 39,599     4
Sports           36,845     3
Food             36,181     3
Social Media     32,932     1
Poetry           26,462     2
Bots             13,926     1
Fiction         13,744     3
Video Games       9,407     3
Nature           7,827     1
News-Satire       5,532     1
Advertising       2,723     1
Virtual Reality   2,578     1
A.I.             1,584     1

Uncategorized 1,694,383   101

Here are the keys findings:

  • The total reach for publications with more than a thousand followers is almost seven million readers; (these includes duplicates, i.e. readers who follow more than one publication).
  • 20% of publications account for 75% of the followers.
  • Only one publication has more than a million followers; it’s the publisher/incubator Matter.
  • Publications related to the tech sphere (including some in the “uncategorized” section), account for half of all followers.

This is both good and bad news. On the plus side, Medium mostly addresses the tech elite. This is a premium audience, attracted by quality, more likely to pay for information; on the supply side, publishers on Medium are more likely to go for a revenue sharing coming from subscriptions, as opposed to see their beloved publication infested with toe fungus ads.

On the negative side, as opposed to Buzzfeed, for instance, Medium is not a mass/general public destination, and will never be, even if it claims 60 millions monthly unique readers. This puts a ceiling on its potential for advertising growth—anyway a broken model, according to Ev Williams.

All this leaves the platform with a preferred option: increasing the dual ARPU of publishers hosted by the platform and creating ex-nihilo a revenue stream coming from the most solvent segment of its audience. This means encouraging publishers to sell their content, mostly through subscriptions (I don’t believe in the per-article revenue model). That’s the two-sided market mentioned above.

To justify a hefty fee, the platform needs to beef up the feature set directed at publishers. Missing functions include:

  • A full blown, friction-free, subscription system that can handle small fees.
  • A newsletter system as good as MailChimp (today, generating a newsletter right from Medium requires building a specific process to make Medium’s weird HTML acceptable for a mailing system).
  • Promotional tools for posts and publications, some tied to Facebook’s amazing promotional power.
  • A/B testing system and SEO tools required to increase traffic coming from search.
  • Better analytics to see what works and what doesn’t.
  • On the advertising side, an ad-serving system able to accommodate a whole range of formats, including sponsoring.
  • A “commercial cluster” system in which scattered audiences would be aggregated around themes and topics, making the platform more attractive to advertisers; Medium could easily reach the critical mass of content to do just that.
  • It should also provide more layouts, formatting options, and editing tools.

It is hard to say how much money each publication and each reader could bring. A full stack platform should probably shoot for tens of thousands of hosted publications with fees ranging from few hundreds to to several thousands dollars per year and per publication paid for Medium range of à la carte services. As for the reader’s ARPU, a modest ad effort could yield a couple of dollars per year and per user for a small part of its audience (Buzzfeed or Business Insider get about a dollar), in addition to Medium’s cut on subscriptions. The complete model is not easy to figure out.

I remain confident in Medium for several reasons: quality will inevitably monetize at some point; between fake stories and the failure of the ad model for news, winds are blowing in favor of good, paid-for, quality contents. Publishers, small and mid-size, will become more attracted to a clever, transparent, friction-free model, as opposed to Facebook which acts as an opaque and arbitrary editor capturing about 70% of the value created by others.