Tea about the future of Medium with Ariel & Buster
Tasty sips you may not have seen buried in the comments on a post I wrote on the Medium Blog
Hey friends! So this is a different kind of post because it’s as much about my colleague Buster Benson as it is about my own writing.
I just want to ensure that Medium obsessives who want the tea about the platform see the interesting intel buried in the comments on a story I wrote last week.
The backstory on this tea 🧋
Last week I wrote this story on the Medium Blog:
While the response to that post has been delightfully positive, a few important questions came up in the comments.
Reminder: I’ve only been working at Medium for three months. My job as the Director of Publisher Growth is to act as an ambassador between the publisher community and Medium’s product & content teams. This can be tough because guess what? I’m still learning! It’s a challenge to be the answer person when sometimes I still have my own questions.
When a deeply meaty question arises, I call on my colleagues to help me with answers. Recently, I’ve leaned especially hard on my colleague Buster Benson, Medium’s Principal Product Manager.
Some of Buster’s answers to reader questions have been completely gobsmacking, so I want to bubble them up here to make sure they get the attention they deserve.
Let’s start with our first cup of tea: why does Medium need a Boost tool or curation team when everyone else just uses algorithms? 🍵
Tom Kuegler asked this question:
I would love a section, Ariel, explaining WHY we need a boost. Substack doesn’t have a boost system. Youtube doesn’t have a boost system. Instagram doesn’t have a boost system. Facebook doesn’t have a boost system. LinkedIn doesn’t have a boost system. TikTok doesn’t have a boost system. They have algorithms that work and give people the content they naturally want and need. I don’t really understand why, when quality is so often objective and disagreed upon, why an internal team of boosters looking over every nomination and selecting which ones are the best is needed or even, according to you all, the best approach here?
I don’t understand why a platform that’s so awesome and advanced in so many ways has decided that they’re going to manually select stories to boost.
Why not let an algorithm do that automatically? Why is Medium doing manually what literally every other platform lets an algorithm do automatically?
I got out my Ouija board like any good medium would (…see what I did there?) and called in the great spirit of Buster Benson to answer this question.
Buster’s answer made my head explode:
A couple thoughts come up as I consider this question, which I think is a really good and meaty one that we perhaps haven’t shared our thoughts on very well yet. Probably deserves a longer answer but here are scrappy thoughts.
Essentially, these platforms you mention DO all have boost: ads. Ads and algorithms are the mission-aligned incentive of engagement-focused platforms. Human curators and boost are the mission-aligned incentive of quality-focused platforms.
Substack and traditional publishers that don’t rely on ads don’t have boost because they put the burden of distribution on the authors/creators. If you bring your own audience, they will take you on, but they make that calculated risk by only paying you if you can prove some amount of market value comes with you.
When the goal is story quality, and we sell a bundled membership that strives to be the best writing (ie not necessarily the most attention-grabbing writing), and we want to scale this across a diverse range of categories that each require special domain knowledge/experience to be a proper judge of, then this community-based human curation system feels like the right tool for the job.
Essentially, we’re doing it differently because we want a different outcome. Most of the information economics of the world are fighting over attention, which leads to outrage baiting, doom and gloom, misinformation, content mills, and derivative works. It doesn’t incentivize thoughtful, personal, original stories. We want to be an antidote to the engagement-obsessed information economy, so we have to do it differently.
It’s a bet, and not proven yet, but that’s at least the most honest answer I could give to that question at this time.
Do you see what Buster’s getting at here? Other platforms don’t need this kind of human touch or curation because their goal is not high-quality content — it’s engagement and time. Because they’re in the business of selling your eyeballs.
At Medium, we’re not in the business of selling your eyeballs. Here at Medium, you retain full ownership of your eyeballs. You pay us to ensure we bring you stories that delight your eyeballs (…and your mind).
(And hopefully your heart, too.)
Now, how about your second cup of tea: if Boost is so great, why might some writers see their views falling? ☕
Hope you’re still thirsty, because there’s more.
When Sandra Pawula asked this question…
I’ve had 4 posts boosted so far with views ranging from 1.3K to 3.7K. But strangely, my views on all other posts have declined radically. Posts like ones that previously received 300 to 800 views as well as a smaller selection receiving 1,000 to 4,000 views are now receiving less than a 100 views each.
As a result, the shift into the Boost Beta program has produced a net loss for this writer so far.
…I gave her the best answer I could, but again, I’ve been here three months. I’m still learning, so I pulled out the ol’ Ouija board again and invoked the ghost known as Buster Benson to give a more thorough response about how Boost plays into distribution on Medium.
Yet again, Buster dropped the knowledge:
I think what Ariel is referring to is that when Medium’s feed was entirely algorithmically generated, it tended to reward stories that fit a certain template/formula/strategy that is well-understood to work well in an engagement-centric platform (which we were, for a while).
And therefore it became easy to fit stories to that template/formula/strategy and to expect some consistent amount of views for those stories. We found that that system, over time, resulted in a lowering of quality across the board, since posting a lot of stories seemed to get better results than posting fewer more crafted stories.
Readers and writers have both been quite vocal about how this has not been good for the overall site. So we’re now starting to do it differently.
As we’re continuing to roll out a community-based boost nomination program, the criteria is changing so that stories that succeed can be less reliant on the template/formula, and more about being truly constructive and original and moving in some way. Which is harder to write, but much more rewarding to read. This ends up meaning there are less predictable outcomes for stories, because writing is quite difficult.
The only way to really write for this new system is to spend more time on fewer stories, because the investment of making a story a bit more personal, original, memorable, etc will result in more distribution and earnings than a bunch of stories that don’t.
It’s not a perfect solution, and we’re still iterating on it and keeping an eye on feedback like this to make sure we’re not over-correcting or creating new counter-productive incentives. And in the meantime trying to be as transparent as we can about our thoughts on this, even when they are still being formed. Hope that’s helpful!
To summarize Buster’s summarization of what he said I was saying (…still with me?), if you’re looking to game the system at Medium, the new game is not about writing as much as possible. It’s about writing less and more thoughtfully, using Medium’s Distribution Guidelines as your beacon.
Yes, this shift may mean that some styles that have previously worked to generate story views won’t work in the same ways. As Buster says, we’re still iterating, and we’re trying to be transparent as we go. Posts like this are part of that.
Alright, folks. Wipe your chins, because that was some delicious tea! Thank you so much for being a part of this journey with me.
We’re all learning together, and staying hydrated while we do it.
