Twitter and Single Article Purchases

Published 07-05-2023 15:26

Recently, Elon tweeted that Twitter would soon allow media publishers to sell content on a per-article basis, i.e. “single article purchases”:

Rolling out next month, this platform will allow media publishers to charge users on a per article basis with one click. This enables users who would not sign up for a monthly subscription to pay a higher per article price for when they want to read an occasional article. Should be a major win-win for both media orgs & the public.

Having been the Director of Tech at a media publisher in Norway, I am very skeptical that many publishers will sign up for this plan, and I doubt that those that do will stick around for long. The economics of digital subscriptions at these companies just don’t work the way that people think they do. I explain that in the first part of this post. However, I also have opinions on how this feature should be built technically. Media publishers have certain “needs” on the tech side that might not be obvious from the outside, so I will discuss that point in the second part of this post.

The Economics

The idea that Elon expressed in his tweet sounds great to consumers, which is why Twitter reacted positively to the idea. We all know the feeling of hitting a paywall and thinking, “I would gladly pay to read this, but I don’t want a subscription.” It feels like there is an opportunity there: allow people who wouldn’t subscribe to pay for the article as a one-off payment.

The problem with this is that very few people really want a subscription when they first sign up. They want to read the damn article, and they need a subscription for that. Some subset will cancel immediately, others will cancel after a week or two, but a handful stick around. Whether they simply forget to unsubscribe or they actually fell in love with the product, it doesn’t really matter: they are paying. So, recurring subscriptions are really the result of people buying individual articles then sticking around for longer.

This is evident in the way the paywall is presented. It shows you a little preview of the specific article you are on, the title, the top image, and it then asks you to pay. It is begging you to spend money on this specific article, even though what you are buying is really so much more. That is because digital subscriptions are, for most companies, sold one article at a time. As a result, the ability to buy a single article and avoid a digital subscription is a huge hit to the subscription base.

Of course, proponents of single article purchases will argue that what is lost in subscription revenue is made up in the volume of sales. This, of course, is pure conjecture. Plenty of publications have offered single article purchases in the past and found that to not be the case, but of course in those cases the buyer still had to go through the hassle of creating an account, entering payment details, etc. For relatively inexpensive products, the inconvenience is a large part of the cost. Therefore, it is possible that a sufficiently smooth buying process will generate the volume needed to offset the lost subscription revenue.

However, that leads to another problem. Most media publishers are incredibly suspicious of any plan that involves building on top of someone else’s platform, especially a social media platform. Publishers have been burned so many times in the past this way. They relied on Facebook traffic, then Facebook changed the algorithm and left them with nothing overnight. They relied on Google ads, then discovered that Google was rigging the ad game against publishers. They went big on video because Facebook was artificially pumping video numbers. Traditional media publishers are tired of falling for these tricks. They want to own their own platform.

In short, Twitter’s plan is effectively to say, “Watch your recurring subscription revenue decline in exchange for a sales volume entirely controlled by someone else’s platform.” That is a very hard sale.

The Tech

But who cares if it will work, sometimes it is fun to build things even if they suck. So let’s say I worked at Twitter and I was tasked with leading a team to build this system. What would I do? (I would, of course, resign. Working at Twitter sounds horrible. But imagine I didn’t resign.)

The key is to avoid the temptation that all social media platforms have to render the content themselves. That is, do not build a Google AMP or Facebook Instant Articles system for Twitter. Why?

  1. Paid content uses custom designs far more often than free content does. Journalists often work on stories alongside digital designers to create a unique experience. Think Snowfall. Rendering this type of content anywhere other than on the publisher’s own site creates challenges for everyone.

  2. Publishers like having control. What about data tracking? What about ads? All of this becomes the responsibility of the platform that renders the content, meaning it will be different from their normal system. Publishers will not like that.

  3. Updating content in a timely manner is extremely important, but also a huge pain. If a journalist makes a change to a story, there is an expectation that the change will be reflected across all platforms immediately. If Twitter is creating a copy of the content, you can see how that can go wrong. Even getting the caching right is a challenge. The chance of inconsistency will worry publishers, and a few bad incidents might cause a publisher to drop the program altogether.

So, this basically means that Twitter needs a way of granting access to content on someone else’s platform. With that in mind, this isn’t hard. The way most paywalls work is that when you land on the page, the site checks “who is this user and does he or she have access to this content,” and if the answer is yes, the backend returns the content. If the answer is no, the backend returns a paywall page.

All Twitter has to do is to pass some single-use token as a query parameter, and provide an API for checking if a given token has access to a given article. Then, a publisher simply needs to add a step to its authentication/authorization check to request the Twitter API whenever it sees that query parameter is present. That should be pretty easy, and still allows the publisher to render the content in the “normal way”.


No matter how Twitter builds this feature, I do not expect it to see widespread adoption from the major media publishers that exist today. If I had to make an argument for trying this out as a publisher, it would be that Twitter’s contribution to driving digital subscriptions is so low that it won’t move the needle one way or the other no matter what. That is hardly the start of a media revolution.