Boosting the boost
Updates for authors on Medium’s tools and Partner Program
When I joined Medium a few months ago, I made a point to say that I’d be focused on tools and quality writing. This is how I said it:
I’m coming to Medium to double down on our publishing tools on behalf of authors and to double down on the quality of our subscription on behalf of readers.
On the publishing tools front, we added tipping, custom domains, mobile editing, and syntax highlighting. Those are table stakes features for a blogging platform.
Clearly, Medium is a great place to host a blog just for the tools alone. Get your blog up and running quickly with all the tools you need to be successful. Write whatever you want*.
Then there is quality writing which is tied pretty closely to our Partner Program, which allows writers to earn money for the content they publish on Medium based on engagement from Medium members. In practice, this is primarily a bonus system for writers. Our business is to find the great stuff and give it to our paying members, and through this work, to give authors a bonus in both readers and money.
I’ve started referring to this as the boost, and maybe that should even be the name of the program. Our job is to boost the best ideas and best information on Medium. That raises a lot of questions. Which posts are we boosting and why? Is the boost big enough? Can other people boost?
For the most part, boosting comes down to which stories we recommend. Medium does a lot of behind-the-scenes work to find and recommend quality content that’s tailored to each reader’s interests. In total, we make several billion such recommendations each month.
To give readers the highest quality recommendations, we’ve seesawed between two primary heuristics. The original model was mostly human curation, with a bias towards well-written, well-constructed, and well-supported stories. Then we swung to a heuristic that was dominated by machine learning algorithms biased toward engagement.
As a result, the top-performing stories had the most engaging titles, most compelling intros, and the most entertaining writing.
Those all matter simply because writing can’t move you in any substantial way unless it can first move you to click and read.
But there are also lots of problems with that approach, including that we don’t want to keep contributing to the attention economy. Attention-grabbing is not the same as useful or valuable or entertaining. Often the tricks of engagement lead directly to disappointment.
In the last few months, we’ve made a number of changes to our recommendation system:
- Given readers more control of who they read — and which topics, authors, and publications they don’t want to read. This is through the true following feed and the show less like this feature.
- Increased weighting of recommendations for stories that people you follow clap for and highlight. This is a form of community recommendation.
- Decreased weighting for stories that have no signal. Previously we had reserved some recommendations for stories that had no engagement but appeared high quality to our algorithms. In practice, this was shifting a lot of recommendations away from proven stories to unproven and often lower quality stories. Authors who promote their article through social media, through email, to their followers, and on publications are now more likely to get their article picked up for a boost.
Have those changes affected you? Probably. I do hear from individual authors, see the trends in the publications that I used to run, and see the trends across Medium. I can tell you that revenue and visitor traffic is up since I joined. But that doesn’t mean that traffic was shared evenly across all authors and it’s definitely the case that what gets boosted changes as new readers and authors join.
I don’t think these changes are big compared to where we are going. In the future, the way we boost quality stories will be much more impactful. I’ll let you know when that change happens.
The near future
Before I get into this, I think it’s worth saying: write what you want to write. I saw a blog post the other day titled, “What does Tony want you to write?” and it was totally misguided. We want to reward you for writing your best stuff, but only you know what that is.
Here are our plans for boosting quality writing across Medium:
- We want to elevate publications, and not just the big ones, as important curators. The editors of the publications here know so much better than us what is an important read and why. Hopefully we can spark an influx of new publications as well. These publications will have a major say in what gets boosted. And yes, publication editors will get paid.
- We want boosting to lead to bigger rewards and bigger hits. That means more money and more page views for authors.
- We’ve started several projects to organize the best of Medium. This is another form of boosting. Evergreen writing should get boosted well beyond the day it publishes.
- I hate to say this, but also I hate not having said it: We will push back on bad faith behavior. I know the temptation for hacking a system is strong, but we want to boost great writing, not great growth hackers.
Now that I’ve shared where we are going, I expect we will be able to give more frequent updates about the partner program. Until then, go write what you want to write!