Medium Changes Earnings Model
New method for calculating earnings and more transparent analytics will be welcome news for many authors
Medium announced this morning that it was making broad changes to the earnings model it uses to pay writers. The company also announced that more detailed statistics and analytics would be available for writers, helping to clarify how and why their stories earned money on the site.
For many writers, this new earnings model and more transparent analytics will be a welcome change. Thus far, Medium’s earnings model has been relatively opaque. Earnings are based broadly on “reader engagement” with “claps” from paying Medium members contributing in a substantial way to a story’s earnings.
There are several challenges with this model. One is that claps favor certain kinds of writers, and certain kinds of stories. People are much more likely to clap for an inspirational piece about a writer’s triumph over a workplace challenge, a positive change in one of their relationships, etc. than for a dry, technical piece about machine learning or artificial intelligence. That doesn’t make one piece more or less valuable, but with the “claps” model, it did make the more inspirational or celebratory pieces more likely to attract applause and thus to earn revenue.
Medium’s new model changes the game. Instead of basing earnings on claps — or other explicit markers of user engagement — it bases earnings on an implicit metric — total reading time. In this way, Medium’s change brings its revenue model more in line with other web-based content companies.
Netflix made a similar change early in their growth. One of Netflix’s main benefits is its ability to recommend new content based on content its viewers have enjoyed. At first, Netflix relied on users to explicitly rate films or TV shows, just as Medium relied on members to clap for stories. Netflix quickly realized, though, that ratings for movies were more inspirational than reality based. People tended to rate serious movies — classics like Citizen Kane — highly, but that didn’t mean those were the movies people actually wanted to watch.
As Netflixs’ Carlos Gomez-Uribe told Wired magazine, “People rate movies like Schindler’s List high, as opposed to one of the silly comedies I watch, like Hot Tub Time Machine. If you give users recommendations that are all four- or five-star videos, that doesn’t mean they’ll actually want to watch that video on a Wednesday night after a long day at work.”
How did Netflix change their model? They moved to looking at viewing time — what users actually spent their time watching — in order to recommend new content. As Gomez-Uribe said, “Viewing behavior is the most important data we have.” By changing to an earnings model based on total reading time, Medium is making a very similar change, which will likely be positive for both authors and readers.
A reading time based model does have its drawbacks, though. Many tech companies have been criticized for building services whose main goal is to keep users on their platforms at all costs. This risks creating “addictive” content which doesn’t deliver enjoyable experiences to users, but keeps them on the site, clicking and jumping around from listicle to extreme political opinion to disaster article to celebrity gossip piece. Just because we spend time reading something doesn’t mean we’re enjoying the experience, or gaining anything from it.
Ultimately, though, I think Medium will find ways to strike a balance between updating its model to rely on user behavior and erring too far in the direction of creating addictive content. Its recommendation algorithms are already excellent, and prioritize showing users content which appeals to their interests. And the curation system should help to weed out low-quality, addictive content, focusing instead on quality, long-form writing.
As with many Medium contributors, I look forward to seeing the new system roll out. This is a positive move towards updating Medium’s payments model, and one that I think will benefit both authors and members in the long term.