What the world really needs is a new place for people to express themselves on the Web.
If this is a thought you have frequently, chances are you are Medium CEO Evan Williams, in which case you can stop reading this interview, because you already know what you said when we talked the other day.
To the rest of us, that proposition was far from self-evident when Williams -- whose previous startups, Twitter and Blogger, went a long way toward democratizing access to mass-communications tools -- founded Medium in 2012. Yet Medium has been quick to catch on, first among the digital intelligentsia, select members of whom it courted to be its early users, and more recently in the general population and with brands like BMW, which sponsors a Medium-based publication called re:Form.
Earlier this week, I sat down with Williams at Medium's offices on Market Street for a long talk about Medium's future plans, its ambition to generate more "casual" content and Williams's feelings about the state of the media business. (He's no fan of television news.) Following is a lightly edited transcript of our chat.
INC.: What's the latest with Medium?
EV WILLIAMS: There's a lot going on right now. It feels like a bad time to go on vacation, which I'm doing on Saturday. Things are growing really well. Medium's growing. Lots of good stuff is getting published. We're changing the product quite a bit.
What's the idea behind Medium?
I started Blogger 15 years ago and in many ways Medium is a continuation of that project. There are ideas, literally, that I was working on at Blogger that never saw the light of day and now I'm trying to figure them out for Medium. One of those is the idea that not everybody who has things to publish or ideas and stories to tell wants to be a publisher. If you're a blogger, even if you can do it in a lightweight, casual way, there's still a cognitive commitment to being a publisher that's separate from having something you want to put out there in the world. Those things have usually been separated. Online, it's great that you can be [a publisher], but I think for the vast majority of people, even among the subset of people who want to occasionally put out ideas or give their perspective on things, the idea of doing it on a regular basis or doing it in a way where they have to promote or get visitors to their stories -- that part is unappealling.
So that was an insight I had while working on Blogger, but Blogger basically was a tool for making websites and updating them, and every other tool or CMS since then essentially started with the same premise: What you want is a website, and that's where you put your thoughts. Eventually, that notion will seem somewhat antiquated. I sometimes compare it to at one point everyone wanted their own land with a fence around it, and over time, more and more people have moved into the city, because, while you have less control and less space, you have more convenience and the ability to bump into other people. For those people who really think about ideas and make culture, bumping into other people is critical. We're trying to create essentially the same thing on the web.
So it's a little in line with that post-ownership society, sharing economy idea of 'Why do I need to own my own power drill if I only use it twice a year?'
Yeah. I've never really compared it to that, but in a sense I guess you could.
There's a certain kind of piece that I now expect to read on Medium first, rather than on an op-ed page or on the Huffington Post or somewhere else. It seems like that happened very quickly. How did you accomplish that so fast? Is it just building a better mousetrap, or leveraging relationships?
There's primarily three things. One is that we built a place where we give, in my opinion, the proper frame for a person and what they have to say that is focused. It's all about the person and the words they have to say. It's not in the context of a social network or something that is surrounded by ads and tons of links. It's a good frame or stage to present on. Two is we did a lot of work to try to get great examples. We did use connections. We have a team that's dedicated to helping people get on Medium and we've had that from early on. And, three, once it starts happening, it self-reinforces.
You wrote "It's hard to understand what Medium is, who it's for and what should be published on it." What did you mean by that?
Tens of millions of people have come to Medium and read stories and they don't know they're on Medium. If they did know it, they're a long way from knowing what Medium is or that they could publish on Medium. I think we can do a much better job of that. Also, even for people who are relatively insider in the media or tech world, we still get questions: What is Medium? I read it all the time, but what are you doing? This idea that anyone can publish is not that well known. There's a bit of a disconnect because people show up and they read a lot of professional content, so it doesn't look like the type of place where anyone can publish. Then, from a company perspective, this whole question of "Are you a publisher or a platform?" is one that confuses people.
Medium's working really well, it's growing and people are publishing great stuff despite the fact that we make it kind of hard to wrap your head around in basic ways, like when you show up on your homepage, we don't tell you where you are or what's going on. We just launched a new homepage, and we're starting to do more onboarding-type experiences. We just are in the process or rolling out sections on the homepage so you can browse a bit more and get a better view of the breadth of stuff on the platform. We're doing a few more things to make it easier to understand what you're looking at.
Those are the product changes you were talking about?
Those are some of them. Another thing we're trying to do is broaden the scope of what kind of content feels natural on Medium. It's never been the intention to create a place where everybody publishes. Most people in the world are not going to take the time, nor do they have the inclination, to write something longer than a tweet on a regular basis. Which is totally fine. I started this with a different goal than Twitter and Blogger. Those were really created with this ethos of "More people in the world should have a voice, and let's fulfill the promise of the Internet to really democratize idea-sharing and publishing." Blogging certainly did that to a greater extent than the Web had enabled before. With Twitter and other forms of social media, obviously the barrier was lowered to the floor, essentially, to the point where anyone who's connected to the Internet can put out thoughts and in aggregate that has amazing effects.
With Medium, that was no longer the goal. The goal was, with people who will publish more substantial things, we wanted to give the best environment to them. We wanted to both lower the bar -- not of the writing process itself but of the publishing process, gaining audience, gaining feedback. That said, it's not our goal to make Medium just a long-form publishing site, or a serious or professional platform. It should be for ideas even if they're small or unpolished ideas.
You've voiced some critiques of the online news ecosystem -- the types of stories it rewards or calls into existence. It seems like part of what you're saying with Medium is that the world doesn't need more news -- it needs more writing that's not news.
Yeah. Twitter is essentially a news service, and news is important for society, but personally I don't want to spend most of my reading time reading news. I get a lot more value out of other types of content and I think most people do as well. There's a natural urge to know what's going on. It's biological and it's social: What's new? What's everyone talking about? What's this shiny object? The vast majority of it isn't valuable to people and it may be even harmful, and that's well-reported, accurate news.
Harmful because it's misleading people or because it's taking up mindshare that could be used for better purposes?
It's creating anxieties that have no outlet. It's psychologically harmful. TV news is the epitome of this. Its whole goal is the raise enough anxiety that you tune in. It's not to inform or make society smarter. That causes us to have irrational fears about things that are extremely unlikely and highlights those instead of normal things that are actually harmful, like what you eat or watching TV news.
So how do you create a system that directs large amounts of audience to content that actually deserves it without creating a system that can be gamed -- whether by stoking irrational fears or by tugging on heart strings, in the manner of all the feel-good stuff that does well on Facebook?
There is news on a Medium. There will probably be a lot more, and there will be stuff that's designed to attract maximum audience. We're not going to stop that. It's good for Medium as a whole if people are publishing things to get clicks on headlines. But to the extent we are creating distribution internally, we're trying to design a system that gives thoughtfulness a more reasonable chance.
How do you do that?
One thing is we never promote anything because it's popular, which is basically how most of the Web works. By popular, I mean pageviews or unique visitors. Any of our lists that show what's popular on the site is based on time spent reading or recommendations, not just views. Recommendations are based on what people say is good. Can that be gamed? Of course it can be gamed, and we'll try to figure out when people are gaming it and try to lessen the chance of it happening. But looking at the data, you get very different lists of things when you look at what's popular by page views versus by what people read and recommend internally at Medium.
We're just scratching the surface on all these things, but the feedback loops we have within Medium are designed to align the creators with the consumers. It's not that we're trying to make people consume a certain type of content. It's just that, in a world of infinite media, if someone is spending time reading something and someone explicitly says it's good, we think that's a really important signal, a more important signal than "Did someone find this headline compelling enough to click on?" So we want to reflect that signal back to the writers and to other readers. Those feedback loops are what drive behavior on any network. So, while the system can be gamed, what feedback loops networks choose to build in has an effect on behavior on that network.
I know you've talked about the role of professional content on Medium as being sort of a sourdough starter that seeds all the fermentation.
I've never used that term, but I'm going to start.
You're welcome to it. And you've also said it's there to make sure the system is up to the needs of the most demanding users. Is there more to it? Is part of the mission here to create an ecosystem that can support more professional journalism?
Yes, it's that as well. The publications that we publish, which we commission stories for and pay for, are meant to fulfill all those purposes. As for the latter, we do think eventually Medium can be a better place to support professional journalism and other types of content than going and starting an island on the Web somewhere, for the same reasons it makes sense for someone doing it for pure self-expression: If you can tap into an audience, if there are great tools and you get more feedback more efficiently, that will make sense for all types of content. And in fact they're complementary.
There are really interesting things our publications are able to do because they're on this platform -- for instance, publish something and get responses from readers that are full-on posts, or pull in content from the community into the publication. We're still in experimentation mode, but we think Medium will be a viable platform for commercial publications with professional content that needs paying for.
Something like the BMW-sponsored publication re:Form -- is that the prototype of what you're trying to do with monetization?
It is for now. Really the goal of that is to help fund great content more than it is to monetize the platform as a whole, and there seems to be a lot of interest from brands to get their names beside great content, some of which is about them but a majority of which is not.
What types of content are you hoping to see more of on Medium?
It will become more diverse in a couple ways. Rather than short or long, I think about the spectrum of casual to serious. Different forms of media by nature land on different places on that spectrum. Instagram is one of the most casual forms of media, at least the way most people do it. Twitter and Instagram for the most part of really casual. Movies and books are very serious forms. Blogging is somewhere in between.
For Medium, we want things that are fairly serious -- the 5,000-word professional piece. Those aren't the types of things that happen on their own. They often need to be paid for, they need a team that's doing the editing as well as the productionization of it. That's a pretty serious endeavor. Then there's the individual who just has the idea and spends a couple hours writing it. We want to stretch more into the spectrum of "I have an idea and maybe it's a couple paragraphs, but I want to put it out there." We want Medium to be a fertile ground where you can put out these ideas and maybe they grow because someone riffs off your idea or gives you feedback to encourage you to expand it.
Part of that is a perception thing because a lot of the professional content has gotten attention so that reinforces "Oh, that's what this thing is for," so if I have something casual to publish I'm not going to publish it on Medium, even if I know I could. Part of it's also a design thing. We have some design changes coming that will make it so many types of things live together in a very natural way.
The perception thing is interesting. As as a writer, the prospect of writing for Medium is appealing but confusing. It feels like there are multiple doors I could be entering through and it causes some hesitation.
What we're trying to do has multiple layers to it. That's a conscious decision that's important in the long term but has challenges in the short term. For example, the fact that Matter or Cuepoint or Backchannel are their own brands that live on top of Medium, which is its own brand, and then there'll be an author in there -- so you land on a story and then you have to understand the relationship between Steven Levy, Backchannel and Medium -- that's kind of complicated. Over time, it gets clearer, just as if you show up in a Vice video on YouTube you're not confused about the relationship between Vice and YouTube. The individual-versus-professional thing gets solved over time as well because you see more things. Just because there are professional music videos on YouTube, if you want to put a video up on YouTube for your friends, you still do it because you understand YouTube's for everything. I think we can get there over time. It's just that when everything's new you're trying to understand it from a few data points.
It seems like some of the problems you're dealing with are similar to the ones Twitter had and has.
A lot of them are similar. Twitter had the problem forever, and still does to a certain extent, that people didn't understand what it was at all or why they would use it because it was a weird idea. Tweets didn't exist before. Medium is less weird because blog posts and articles exist and most people read them, so a place where you read and write this type of content is easier to understand at its core but there's a similar thing in that everybody has a different Twitter. People can have very, very different experiences on Twitter. They can see it as a utility for news and information, they can see it as a pure social environment, they can see it as a place for self-expression and self-promotion.
I remember seeing some celebrity on TV fairly early on talking about Twitter and how great it was because it helped them connect with their fans, and I thought, "Oh, that's why if you're a regular person and you don't care about that celebrity, you'll think that's why Twitter's not for you -- because it's a way for celebrities to connect with their fans." So they write it off -- it's a celebrity thing. Over time, they get more data points -- "Osama bin Laden got killed and it broke on Twitter, and I care about that." And eventually you realize it's very general.
There was a big outcry this year among some Twitter users when its executives started talking about using algorithmic curation to improve the content of their timelines. Is there a reason algorithmic curation would be more suited to one platform than another?
I think a lot of it has to be with where you start. I think Twitter needs to get more algorithmic over time. It's very clear, it's been clear for many years, that any given Twitter user is not getting the ideal timeline or ideal tweets for them based on their interest in the 500 million tweets published today. Those who do do a tremendous amount of work curating their own timelines and followings, which is a lot of work and something most people aren't prepared to do or really know how to do. So I've been a fan of moving more toward having some component where you use the computers to help the people on Twitter for some time. I think the outcry about it misunderstands what that necessarily means and they point to Facebook as the inevitable result and they don't like how Facebook works and they also don't like the idea of losing control. I don't think it necessarily means you lose control. First of all, all algorithms are not created equal, obviously, and they reflect certain values. What are they trying to optimize for? You can create an algorithm that optimizes for serendipity, or an algorithm that optimizes purely for popularity or for clicks on commercial messages. It's completely up to the system designer to decide what they're optimizing for, A. And B, Google is driven algorithmically and you have no control over what it shows you but it certainly does a hell of a lot better job showing you what you care about for any given query than some manual thing, whatever that would be.
I think Medium will be the same. We actually have moved from being more algorithmic to being less over time.
More human curated?
Yes, more human curated. If you're logged out, the homepage is all human curated. If you're logged in, you can get the human-curated stuff as well as recommendations from anyone you follow. That's basically crowdsourced curation. Recommendations on Medium are essentially like retweets. That will certainly evolve over time. What we've realized is algorithms work better the more data you have and we're still small.
How much urgency is there around building a big business here?
As in it'll happen when it happens?
Yeah. Well, no, we have to make it happen. But it's not our focus right now, for a couple reasons. One is that we're well funded and we have access to more capital and this is a long-term play. We have a lot of work to do to make the product better and to build the momentum and the really critical mass of publishing and engagement which is going to set the ground work for a great business.
That said, we are generating some revenue today and we're going down that path much, much earlier than I did at Twitter because we want that professional content. With Twitter, tweets just happened and our problem was building a more robust system to support the network and the activity. We were confident there was a good business there. Contrary to external rhetoric, it was very, very clear to me that if you had a unique real-time information service that had commercial usage from day one, there was a business to be built there. Similarly, there are multiple businesses to be built here if we build a great platform for sharing stories and ideas because companies want to do that, and in fact many organizations are using Medium for publishing today. That's commercial usage we're not charging for, nor would we want to anytime soon because it helps the platform. People like the content. But for the professional stuff, it would be helpful to get that stuff paid for. So that's how we're looking at revenue today: How can we get money into the platform that helps support great content creation.