Zulie Writes

View Original

Breaking Down the Medium Algorithm by Analyzing the Medium Homepage

Screenshot of author’s Medium homepage

It’s no secret that Medium habitually changes its all-mysterious algorithm. They tweak how they recommend stories to your followers, how they decide to curate stories, and where your views come from. It’s hard to guess what they’ll change next, and it means it’s hard to predict how successful (or unsuccessful) you’ll be.

But by examining their homepage, you can get some hints for what’s happening. For example, for the last few days, I noticed my recent stories weren’t getting hardly any traction. Seriously: all curated, and all under 150 views, despite the curation and alleged 5.7k followers I have.

Screenshot from latest stories

So I wandered over to the homepage and had a look at what was being recommended to me. What I noticed was that there were a lot of older stories doing the rounds — things that were months or even years old. 

Medium only has so many places to recommend stories, so if they’re recommending predominantly older ones, it means your newer blog posts might not have the upper hand. This explained why my latest stories weren’t getting a lot of traction.

That’s just one example of the information you can glean from the Medium homepage. Here’s what else you should look for.

1. Their featured stories are all in their in-house publications.

Screenshot of Medium’s homepage

If you mosey on over to your own Medium homepage, you’ll notice all those stories right up at the top? They’re published in Medium’s own in-house pubs. Forge, GEN, Zora, Human Parts. Personally, I couldn’t find any stories in there that were self-published only.

OK, I won’t lie: When I saw this shift, I groaned. It’s hard enough to get stories that do well on Medium. Now I’d have to get them in exclusive publications, too? 

But what that represented was a need to pivot. Previously, I’d focused on writing the best stories I could. Now, I regularly submit to those publications. The sting of rejection sucks, yes, but the two stories I have had accepted there made it worth my while.

Spend time reading those stories, learning what those publications are looking for. And submit. Honestly, the vast majority of my stories submitted there aren’t accepted, but the two (so far!) that were have done incredibly well. Medium paid me a flat fee of $200 for each, with the guarantee that if they earned more through engagement, I’d continue earning. 

Love it or hate it, Medium has shifted to promoting their own publications. You can rail against the injustice of it all, or you can do like I did and accept you’ll need to change your priorities, too.

2. How are stories recommended to you?

Screenshot from Medium’s homepage.

If you scroll down and look to the left, you’ll see Medium gives a small justification for recommended stories to you.

I scrolled through 50 of the stories recommended to me. Fully 34 (68%) of them were from my “reading history,” 12 (24%) were from my network, and 4 (8%) were from topics I followed. 

READING HISTORY

It seems the vast majority are from my “reading history,” so I delved deeper into what that meant. Dates varied — as you can see above, one was published only hours before, one was published in May. All of the ones outside of publications were curated. I followed either the author or the publication for all stories here.

It’s interesting: I would have assumed that “reading history” is based on the topics I’m interested in — the topics that Medium curators assign — but it seems those are shown to me under the “topic” section of recommended stories. I know, from speaking with the support team, that tags (the self-assigned tags writers can add up to five of when publishing stories) don’t have anything to do with how Medium recommends topics. 

In the meantime, we can assume that this is the “algorithm” at work: it may be suggesting topics I don’t follow but have read and clapped for stories of. It may be some combination of writers/publications I follow + topics I’ve clapped for. Perhaps this “reading history” section is for folks who haven’t followed any topics yet.

I have emailed Medium’s support team for more information on this and will update this article when I’m able!

TOPICS

This section seems straightforward — based on topics I follow, I am shown stories. The interesting thing here is that often I don’t follow the author or the publication this story has been curated in.

Screenshots from author

This is super important: it points to the fact that curation is vital for reaching new followers. The more topics your story is curated in, the wider the potential audience.

FROM YOUR NETWORK

Several of the “from your network” stories were uncurated which was great to see as it means you are able to reach your followers without curation. However, the vast majority of them were inside publications.

I personally don’t follow many people (around 100) so potentially I’m being shown more publication-based stories.

If you’re targeting people like me, who are more likely to follow publications versus individuals, this indicates you may want to shift your publishing strategy to publications instead of self-publishing. In the past, I’ve advised against publication-publishing unless you have a very strong case for a niche fit, but this suggests that Medium is once again pivoting to showing readers publication-based stories.

3. New from your network

Screenshot #1

Screenshot #2, taken five minutes later

This is the only section purely dedicated to writers or publications you follow.

The important thing to note here is that the timing differs. If you check out the stories, you’ll see one was published half an hour ago, two were published yesterday, and one was published a week ago.

This means it is *not* chronological. Human Parts, the publication in there I actually follow, has published several posts since the 1st of September but they're recommending this story to me now. It might have more to do with how much I engage with Human Parts, how often they publish, or even something completely different.

A few things here to keep in mind:

  1. The “new” from your network is relative to how new their work is, not how long it’s been published. If someone you follow has only published one thing in the last month, Medium will probably display it on this tab for you at some point. 

  2. The more you publish, the more you show up on this tab. Shannon Ashley, for example, is a prodigious writer so I often see her work here. This is especially important for newer writers who may not have many people following you. If you want to earn money, you want people to clap for your articles as much as possible. 

  3. This area is super dynamic. I left the homepage for five minutes, and when I returned all the articles were different. You might only get one chance to grab readers when your work appears in this tab so you want to be absolutely certain your headline is a corker. 

The very last thing you need to keep in mind is that the Medium homepage is not static. Even when I began to write this article versus now, publishing it, there’s been a change. Look at the title image for here, and you’ll see “Startups” used to be one of their featured tabs, in between “Self” and “Human Parts.” Now? It’s been replaced by yet another Medium publication, Marker:

Another Medium pub - Marker.

What that means for you is that you have to keep working at it. Things will change, the strategies will change. Not from year to year, but month to month or even week on week. Last week, a lot of my older stories picked up steam. This week, newer stories seem to be doing a little better. Previously, I recommended against submitting to publications. Now? It might be one of the better ways to crack into the recommended stories.

Unfortunately, there’s no single correct way to write. This article is really a reflection on the need to stay up to date with how Medium recommends stories to you, so that you know how your stories might be recommended to others. Whether you’re getting curated stories or not, whether you have 100 or 10,000 followers, the recommendation strategies are sure to shift. You should shift with them.