Updated Guidelines for Boost
How Medium curators assess story quality, and how to think about it in your writing
We’ve updated Medium’s Quality Guidelines based on direct feedback from the Medium community, and I wanted to say more about how we hope it does a better job of explaining our approach in a useful way.
Our underlying philosophy about story quality hasn’t changed, but we learned that our previous guidelines reflected biases toward nonfiction content. Though our curation team used a very wide interpretation of the guidelines, we found that they conveyed something entirely too narrow. It was difficult for writers and editors to extrapolate them into genres like poetry, fiction, humor, and memoir.
These are not new guidelines so much as they are refinements to be more clear and to ensure that they reflect the vibrant diversity of Medium’s writers and publications. There is also some freshening up of our stance on AI-generated content; we need to make it absolutely clear that Medium is a place for human writing. And we’re trying to align how we talk about this process internally and here, in public, to help us be as transparent as possible. So we’re asking your thoughts and feedback on this update, too — we’ll just keep making it better.
How Quality Guidelines Are Used to Review Articles
First, a quick refresher on how distribution works. All stories on Medium fall into one of three categories:
- Boosted: Stories selected by the Medium curation team to get priority in being matched to readers
- General Distribution: Stories that get “regular” priority in being matched to readers. This is the default category for any new story on Medium.
- Network Only: Stories that are not Boosted nor in General Distribution, but do get to readers who are following that writer (and/or the publication, if the article is in a publication). The writer’s and publication’s followers are the “network” for these stories.
This post focuses on our new guidelines and how the Medium curation team selects stories for Boost. For more about the entire process of story distribution on Medium, see what happens to your post when you publish on Medium.
What Boost is, and what it is not
The best overall way to think about stories we Boost is to think of them as if they are cover stories for a magazine. We’re selecting the stories that shine.
Boost is not an assessment about whether an article is “good” or not. There are hundreds of good—even great—un-Boosted stories written on Medium every day. Medium’s recommendations engine is there to make sure that these stories get distributed to the right readers. To extend the analogy further, those stories are still in the magazine. And many of them are “on the cover” in publications on Medium, which feature their own story picks.
The qualifications to be Boosted are not easy; it requires that writers and publication editors spend time polishing and crafting the work. Boosted stories are something extra special.
Writers and editors are justifiably proud of their work—and we’re proud of them, too. But it would be misguided for any writer to expect all of their stories to be Boosted. Boost is a bonus, giving a story a short-term push in our system, not a judgement. (We also don’t “auto-Boost” any writers or publications—a common myth about Medium curation.)
The updated Boost guidelines
In deciding which stories to Boost, we look for those that help readers deepen their understanding of themselves, others, their endeavors, or the world.
These stories can come in almost any form: nonfiction, fiction, poetry, personal essay, memoir, humor, photo essay, tutorial, and a myriad of other possibilities too numerous to mention. (When we say “story”, we mean a work of any form created by a writer on Medium.)
A story need not exemplify every one of these elements to be Boosted. We interpret these elements in a broad, nuanced sense — not as a checklist of requirements, but as a target to shoot for. We’re looking for stories that we’re especially proud to put in front of Medium readers. Here’s a more detailed look at what we look for:
- Is the reader’s life enriched by reading the story? Stories selected for Boost might elicit emotions like laughter, tears, excitement, wonder, and more. Or they might help a reader learn something that will help them do their job better, master a new skill, lend words and understanding to feelings, navigate relationships with greater ease, better understand issues of the day, reconsider their own perspectives, or even feel less alone in the world. Whatever the genre, there’s a relationship between readers and writers, and Boost-worthy stories exemplify respect for that relationship; the reader is not left with a sense that their time has been used to read a sales pitch or to indulge a writer’s bid for attention. The story shows genuine regard for the reader (as opposed to being written mainly for the benefit of the writer).
- Is it original, human-created content? The story is original in that it explores something previously unknown or not frequently examined, or it re-examines something we think we know a lot about to shed new light, voice, or perspective on the topic. The story does not appear to be generated by a large language model or other text-generating technology. (The story does not need to be originally published on Medium — see more here about republishing from your blog or other site.)
- Does the author speak from relevant knowledge and experience? The story demonstrates that the author has credible, first-hand experience with or knowledge about this subject, and they care deeply about communicating it effectively. That may be because they’re a subject matter expert or because of their lived experience. One way or another, there’s a clear and compelling reason why this particular author is writing about this particular topic — their insights (whether revealed in the context of fiction or nonfiction) reflect relevant experience.
- Is the story well-crafted? The story is well-written, free of errors, appropriately sourced, and narratively strong. The content, editorial choices, and central point of the post are clear and compelling. The length of the story is appropriate and serves the purpose and context of the story itself (well-crafted stories can be short or long). The writer has worked to craft an interesting, non-clickbait title and the story delivers on what that title promises. Images, if any, add value to the story, include ALT text to make them more accessible, and are appropriately credited. Formatting used correctly is one of many indicators of a writer taking care to craft their work (see formatting tips here).
- Does the story have impact? We’re looking for the kind of stories that you’re still thinking about days later. Are you compelled to share this with your network? Did this story get you thinking? Move you? Make you feel good about the value of your Medium subscription? Are you glad you read it? This type of impact is another facet of craftsmanship and applies whether the topic is universally appealing or totally niche, for stories of any genre or that defy genre.
What about genres like fiction, poetry, humor, and memoir?
Our previous quality guidelines focused on some qualities that writers and editors found difficult to parse for all genres. Previous guidelines included questions like “Is it constructive?” and “Does it feel memorable?”—we heard legitimate complaints like “This is a humor piece, it isn’t meant to be constructive!” and “This is a factual case-study with take-aways, not an inspiring personal story—how is it supposed to be memorable?” (Here’s a link to of those previous guidelines on the Internet Archive if you’d like to compare)
In those previous guidelines, we failed to convey that we gave these elements a very wide interpretation across genres. That perspective is what’s motivating this new version. For example, rather than asking if something’s “constructive,” we focus on if “the reader’s life enriched by reading the story.” Instead of “Does it feel memorable?”, we ask if it has impact on the reader.
The foundation of how we assess quality is still there, much of it unchanged. However, we’re trying to be more specific and do a better job of reflecting the actual spirit of how we work. We drew inspiration especially from the feedback we received from the participants in our Boost Nomination Pilot. They offered valuable insights into specific problem points in our documentation and shared their frustrations with a process — and decisions — that were confusing.
We especially appreciated the way writers reminded us that the way we’d described “quality” may have been too biased toward “how-to” nonfiction. In looking for ways to better serve more literary genres, we also took inspiration from the literary journal jmww:
We aim to publish poetry that contains solid, riveting images, is emotionally complex, and that yields meaning beyond what is merely personal, thus having implications for others and/or society.
- from the jmww submissions guidelines
What a wonderful description of the type of poetry we’d like to put in front of readers!
We’re sometimes asked about our guideline for “Does the author speak from relevant knowledge and experience?” In the context of genres like poetry and fiction, the relevant experience of the writer is their ability to imagine something that hadn’t previously existed. Their craftsmanship is exemplified in their ability to draw the reader into that world.
Who are these curators? What about bias?
So who are Medium curators, and how do they make decisions on individual articles, and how can writers and editors trust the decisions they make?
Curators contract with Medium to apply Medium’s standards to articles. They come from a variety of content backgrounds, including technical/programming, DEI, science, film, health and fitness, lifestyle, and more. Many of them have writing degrees and/or deep experience in writing, editing, and content curation. They work in countries and time zones around the world. They are paid a flat hourly rate — that is to say, they’re not paid on the basis of what they Boost or don’t Boost, or on the number of articles they review.
They care deeply about Medium’s readers and writers, the Medium platform, and about doing excellent work. Evaluating writing for quality is an obviously subjective undertaking, and part of the job is to consistently apply Medium’s standards rather than their own as much as possible. Curators routinely initiate discussions around decisions, noting that they’re concerned they may be bringing their own bias to the table and seeking additional perspectives to overcome that bias.
I’m fond of saying that “I don’t like it” and “I don’t believe in it” are not reasons for not Boosting an article — and that’s an expression of how we approach our work. For example, I personally don’t believe in astrology and I loathe diet articles, but in my work, I have Boosted articles about both. All of our curators operate in the same way.
We recognize that the task of curation is inherently subject to bias and personal preference, and we work hard to come back to our guidelines, time and time again, as the basis for our decisions.
We use the same guidelines to quality that we’ve just updated on our web site, but we also have a manual that collects helpful details for more complex decision-making. You can read this curation manual for How to Assess Story Quality if you like, to get an insider view of how we operate.
And when we encounter a difficult example, we discuss it and try to learn from it as much as possible. We sometimes even change our minds about decisions. We remind ourselves often that we’re not here to be gatekeepers — we’re here to ensure that readers see the very best of what Medium has to offer.
Our work here isn’t done by any means. We understand that writers, editors, and Boost Nomination Pilot participants want to have a clear understanding of our stance on quality and Boost. We want to be trustworthy partners in bringing the best stories forward to promote a place of intellectual wellness for readers and writers alike.
We’re also doing more work to match stories to curators with specialty knowledge in particular genres or subject matter. We’re doing work to make sure we do a better job of listening and learning from feedback. And we’re constantly honing our internal tools and processes, which are very much in a constant state of improvement—like the Boost Guidelines themselves.
This work we’re doing in curation doesn’t happen in a silo. Our work supports and is supported by the work being done to update the Partner Program, improve how we harness the power of Boost in all its contexts, and of course all of the continual work being done to improve recommendations for Medium readers. And of course, we’re working closely with Boost Nomination Pilot editors (and adding many more of them) to expand the number, variety, and sources of great stories for Boost.
It’s exciting work and we’re grateful that you’re on this journey with us. We’d love your thoughts and feedback on the new guidelines and our approach in the responses.