There is an online platform whose main objective is to get people to start reading again. Actually reading. When Medium launched in 2012, the minimalist layout offered up just one font for any text published on its site in an attempt to bring the focus back to the words and less to all the extras that surround articles on a page today. When Twitter co-founder Evan Williams realized how much Twitter had changed the way people communicate on the internet, he also found that it was limiting, to 140 characters exactly. In an attempt to lengthen the conversations, he went back to his original incubator Obvious Corporation to launch a site that could change the way people communicate, yet again.
Medium’s platform is a writer’s heaven and they hope it will become their haven. The white space that presents itself when posting an article is open and inviting. While Twitter allows people to exchange words, Medium allows them to have conversations. A response to an article becomes a new post and can be just as long, if not longer, than the original post. This freedom to not just reply, but to develop and build upon thoughts and ideas is what Medium thinks has been missing from the readers’ experience online. Williams has expressed his hope that the platform will become a place where “ideas and people build off each other,” a thought that is paralleled in Twitter’s origins and traces all the way back to Williams’ first big tech project, Blogger.
But unlike these, Medium’s founder sought to create not just a different product but a different company. Since its early days, Medium has been structured as a company run by self-organizing teams rather than a traditional management hierarchy, very similar to the theory of Holcracy. Although there are managers at Medium, the flat, open structure is intended to allow roles to be distributed evenly in order to instill autonomy within individuals. The Medium offices have classes in yoga and meditation to encourage a culture of mindfulness and to encourage teams to stay nimble and work efficiently. Each Tuesday night, employees gather for “jank and drank” sessions, where they can have a drink and work in a relaxed setting with each other, often tackling bigger projects that may have been languishing in the day-to-day. Although Medium does not subscribe to Holcracy in the traditional sense, all employees are still given freedom to speak and resolve issues without any traditional corporate constraints.
One of these employees is Joy Chen, an engineer at Medium’s headquarters in San Francisco. With a degree from MIT and two internships at Google under her belt, Joy joined the company in 2014 and has watched the company, and her love of writing, grow. Even though most of her work is based on servers, web-clients, Android, and data pipelines for user-facing stats, Chen is deeply impacted by the writing she finds on Medium, stating that her interest in “mythology, religion, and how the language we use and stories we tell influence how we model the world and make meaning out of our lives.” This is Joy’s day at Medium.
Can you describe the culture at Medium?
We’re approximately what you’d expect of folks building a platform for everyone to read, write and share ideas. We work hard, care about our words, and create with a mind for timelessness and meaningful interaction. You can roughly bucket us into computer-people, product-people, media-people, and people-people, but we’re very collaborative and share similar values. We move fast and highly value thoughtfulness and empathy in the product and each other.
How is it different than other company/companies you’ve worked for?
My favorite thing is that there are few enough of us that I know everyone by face and name! As an “individual contributor”--- what we call people who aren’t responsible for other people --- I interact daily with a team of about 3 to 10, and regularly discuss (or just show off…) stuff I’m working on with the whole company. We’re at a great stage where we have a good core product, but we’re always looking for the next most valuable thing for us to work on, and we’re nimble about moving people around internally to reach our goals. Over time, I get to work with lots of different people on different things.
We are also unusually open about our ideas and expressive with our words. We are all active on an internal version of Medium (different url and different database backing it) that we use to communicate with each other. We write tech specs, product musings, product complaints (ours or our users), analysis, and basically our hopes and dreams for each other to read and riff off of.
As an engineer, what drew you to Medium?
Besides the relatively small company size and millions of readers/writers on the good-looking service, mentorship! Medium was my first full time job, after some great internships at Cisco and Google. I wanted to be with people who were really, really good at their jobs, and had empathetic, kind personalities, because those were people who could teach me how to be a better engineer. Medium highly values on-the-job mentorship. We have an official Guild program for direct 1:1 mentorship, but we also encourage other indirect mentorship behaviors like internal technical spec writing. People are patient, open-minded, and curious about new problems, even if they aren’t their own. I love that while working as an Android engineer, I can pop over to Data Science to ask about building a data pipeline I need for a feature, or pop over to Web-Client Guild to ask how to rearrange the styling of a webpage in a way that doesn’t confuse a future engineer.
Can you talk a little about what you do day-to-day as an engineer?
I work closely with designers, product managers, and usually a tech lead to build product. At the beginning of a product cycle, I help figure out a strategy for improving various aspects of the Medium product: allowing a new interaction, showing information in a new way, or making the experience of being a Medium denizen better. Then, after we put together a rough plan, hopefully with pictures attached, I’m responsible for building it. While I’m building and iterating on the initial idea, we dynamically steer based on pain points/new ideas that emerge. When the feature is good to launch, I’m responsible for making sure it’s safe, verifying that it meets our engineering, design, and product quality bar (checking in with appropriate experts on these subjects), before flipping that switch and watching actual people use it. Post-launch, I become the de-facto expert on that feature, so people will ask me about various bugs. I then interpret, triage, and likely propose fixes for bugs. If it feels important compared to other things I’m working on, I may act almost immediately on the proposed fix. Otherwise, I leave it for someone else (including potentially my future self!) to act on.
What is the best part of your job?
Working with and learning from wonderful people. Right now, I’m working with a Product Manager I really admire, a designer who I’ve learned a lot from, and a tech lead who geeks out about data and tells me all about it.
As to stuff I do, my favorite part of the project development cycle is the “Build it as fast as possible!” stage, where we’ve figured out what we want, and it’s just me, the code, and my code reviewers for many hours as a run. It feels pretty awesome to make an idea a reality. A close second favorite is a few weeks after launching a feature, looking at graphs and reading user reviews. I get the best warm-fuzzies when someone, using something I built, tells me they love it, right as the feature is fading out of my working memory.
What is the worst part?
Not being able to build everything. Prioritization is my best friend and worst enemy. When presented with a problem, my (our) imagination runs amok with ideas of what a solution could be and what it could do, but we have to prioritize. There’s a range of engineering personalities at Medium, but I’m particularly vicious about sticking to my priority list. I work on it if I think it’s important, regardless of whether or not I’d rather be doing something else more fun. When there are many important things to work on, I find myself sometimes longingly looking at a designer’s pretty, next-level mock of a feature, wishing that I could afford a week to make that a reality.
In terms of engineering, what is the hardest aspect of working on Medium’s platform?
It’s different for every engineer and the problems they’re working on, but I can broadly say that the sheer variety of problems we solve is a challenge. Most people will gravitate towards work that fits their particular skill set, e.g. an iOS engineer will do primarily iOS client-work, and their primary skill-development path is in iOS expertise. However, in the course of product development, we often want to build things end to end, and the little team in charge might be missing a certain type of expertise (server, API, database, ops, etc). If you’re a mostly-mobile-client engineer, there’s no “API-team” that you call on when some API-endpoint isn’t doing the thing you expect. You dive into server code and see what’s up. While working in codebases we’re not familiar with, other people are happy to guide and direct you to a good solution, but we believe in the permeability of ideas and engineers between our codebases.
Are there any aspects of the layout, platform, or design that will be worked upon in the future?
All of it! I suppose from a distance, Medium looks rather polished and complete, but in our minds (and our code), it’s constantly growing and evolving. On pure design, a few months ago, we chose a new font and a new Medium-green. On pure layout, we recently redid the Medium profile page to better represent Medium people. (Check it out on Android! I’m extra proud of that one.) On platform interaction, we just launched a change that allows us to show conversations many-levels-deep.
Is there a post that captures for you what Medium is attempting to do with publishing?
So many that this question is nearly impossible for me to answer! I actually started at Medium post-launch, but of course it wasn’t as well known a year and a half ago.
My favorite category of stories might be nerdy, human ones, particularly ones that are self-deprecatingly funny. Some examples include Nick Santos’ “Why ContentEditable is Terrible” and Jamie Talbot’s “What are Bloom filters?”
In terms of broader influence, and things that felt Important with a capital I, the New York Times and Amazon having an intelligent, meaningful (albeit rather heated...) conversation on Medium about Amazon’s work culture was super cool to see play out.
What makes a post successful on Medium?
We’re actually very interested in that question ourselves! Our user research has shown that success means different things to every writer and organization who represents themselves on Medium. For example, some writers just want as many eyeballs as possible (views), and some care about a real-world action (buy this product I’m selling), but most normal people like me just want to find our audience. We want our writing to reach people who care about the subject, and to be able to have our ideas be heard, understood, and built upon. For example, I personally don’t care if 1000 people see my story, but I like that 100 actually read all of it, and love that 5 people responded with ideas of their own.
Speaking to your actual question though: Given the assumption that “success” is measured purely quantitatively, successful Medium posts are usually good, honest stories, statements, or analyses that introduce a new, useful idea into the world, written by anyone who has something important to say. If they gather enough steam in the Medium machine, through recommends and other interactions, they are seen by more and more people.
Medium is first and foremost a long-form publisher, yet on average, the “most recommended” posts average less than 10 minutes to read. How do you think this will change as the site continues to grow?
First, we actually don’t care that much about what the most recommended posts are. Metric-wise, we care about reading time, and we actually know, as exactly as possible, the post length that gets the most reading time, and yes, it does fall under 10 minutes.
I don’t suspect that to change very much. It seems like a property inherent to the people who are reading on the site. 7 minutes is about enough to convey an idea completely, without being too wordy, so that’s the “Medium-length” article that works well on our platform. I suppose if we suddenly get a huge influx of attention-deficit teenagers reading on Medium, that number might drop, but I’d be very surprised. People reading on Medium are here to read, and that takes a few minutes at least.
How do you think Medium will change the way long form content is shared?
Good ideas will come from more places than they ever have before, and we’ll come to expect more from “stuff that’s written on the internet.” The diverse, collective wisdom of the internet will find an extra inviting habitat where there are efficient feedback cycles that breed meaningful discourse.
Is there a disadvantage to the fact that almost all content has not been edited by anyone from the company’s editorial team?
Nope! Medium exists to allow for all voices, no matter their editorial polish, to coexist and build off of each other. I don’t really have the expertise to speak to this, but I can quote our CEO, who thinks about this type of stuff all day.
Ev Williams said: “We know that length is not a measure of thoughtfulness. The quality of an idea is not determined by the polish of the writing. And production value does not determine worthiness of time investment on the web any more than it does at the movie theater.”
Do you think Medium will increase its editorial oversight of content? If no, why not?
Nah. Compared to everything else we want to do, helping people edit their stories is just not that valuable. People should say what they want to say, and we want people to own their writing. We definitely don’t expect everyone who publishes on Medium to generally have that level of expertise. I mean, I write on Medium….
What advice would you give to prospective job applicants?
Use Medium. Read, write, interact with our platform, because that will give you a much better idea of what we do, why, and how, than any 3rd party description will be able to explain to you. When, at the end of every interview, we ask if you have any questions, maybe you’ll have more questions that challenge *us*. :)
Can you talk a little about the interview process?
On structure, we run a phone screen and maybe a take home, followed by a much more interesting day of interviews. We’re constantly working to making our process better. Our technical interviews try to be relevant, and we screen thoughtfully to better understand our candidates. We care a lot about 4 qualities even over technical competence: Curiosity, Awareness, Resoluteness and Empathy. We believe that looking for these character traits allows for trust, distributed authority and autonomy at Medium. Also, we’re working with CODE2040 and other partner companies to refine our interview process, mitigate bias, and build a rich, diverse company culture.
Can you give us any insight as to what changes readers and writers might see in the future?
Only a little bit. Not for secrecy, but because we’re just discovering them ourselves! We like to propose questions, not solutions, and it’s hard to tell you exactly what to expect because fairly soon after we have conceived a solution, engineers like me will have built and launched it. The major questions we’re answering now are:
1. How might we help casual Medium users, who don’t follow many active people/publications, find great stuff to read?
2. How might we expand the capabilities of our publications so that they’re attractive to professional bloggers/editors/writers who already have a fully-featured website?