After reading about our experiment with slow live blogging, Vox Media chief product officer Trei Brundrett reached out to us to talk about his own experience building Storystreams, a similar idea he launched in 2009 on SBNation.com. (Here are two good examples from SB Nation and The Verge).
We asked Brundrett about the impetus behind Storystreams and—after pulling on the thread—discovered there’s a rebellious philosophy behind them. His argument: Publishing companies need to come together and tell tech companies how to index news—not the other way around.
Why were Storystreams first created, and how have they impacted your traffic?
The user data validates some of the things that we learned early on and then iterated on. It's actually an interesting story. The reason we created them was that when we first launched SBNation.com, we only had team sites, and the big challenge that we were trying to solve for was that we had really great writing on all of those sites, but they were writing for their audience with a lot of assumptions about what they knew. If you're on Purple Row [Ed. note—my favorite SB Nation site, go Rockies!], they're going to write with a lot of assumptions about what you know as a Rockies fan. Not just this year, but over the past few years—what the nicknames are, what you call the coach, how bad a certain player is, whatever. Sometimes they do really great writing like that, but it's hard to feature that story. We were really thinking about how can we put those stories into context from the network? Especially when we had things like a trade between two teams, we'd have their different takes, and then we wanted to write a summary and then follow that trade rumor as it was going along, even before it got to their reactions. So that's how we came about with the idea, to provide that context.
The other problem we were trying to solve is that our blogs weren't always super fast to publish the stories. But the dedicated editorial staff that we hired for [SBnation.com] could do that. So, we wanted a format that actually worked with how quickly the narrative was evolving, and all the different sources that came into it.
Immediately when we launched Storystreams, we realized that it was also really valuable for us as bloggers. What we'd do before to update a story is we'd put a little update at the bottom of the post. But the problem we found with that was that you didn't have a way to tell the user that there was something new there, and it didn't have its own URL, so it didn't go out to Google or ping Twitter and Facebook. We iterated pretty quickly on that because what we found from the data is that they're really powerful for search and for social. It's improved our pages per visit because when you're on a story that's in a Storystream, we show the other Storystream entries, and we get really good second click through that because it's super relevant content to promote, and it's on a timeline so people can see where they are on a timeline and navigate around that.
The thing that's interesting to me about Storystreams is how little writers are using the update feature. I loved the original Storystreams, where you had a lot of self-contained updates. That experience is great. But so many stream entries today are big, long articles. It seems like that forces the user to do a lot of reading of 400-500 word articles to get the gist of the story.
It sounds like you're more familiar with how we had this working before. It used to be that a much more granular piece of content could be pushed into the stream, rather than all of these articles. The reason that we got away from this is that we kind of fell a little too in love with how much Google was driving traffic into the streams. We started paying a lot of attention to it trying to figure out exactly how that worked. What we found was that Google News has a 150-word minimum to be included, or it will give you an "article too short" error. So we started writing longer updates, and because we brought on a lot of new editorial staff during that time, they were less familiar with the Storystream format we were doing. They thought, "Oh, I just write articles and organize them into this stream" rather than creating the stream. We got away from this, but we used to have this inline editor at the top of the stream where you just pop it open and write an update and it would push it in and update everybody who was following it.
Now we're shifting back to that. We've actually had feedback from users who say, "your streams used to be easier to interact with because they weren't just topic pages, they were an actual stream of updates." What we found is that it's fine to punctuate with articles that are more summary. But a lot of these things don't need to be as long as they are, and we can be willing to forgo some of the Google traffic to make the value of the stream continue to work. What's happened is social is becoming a lot more important. Social doesn't care about the length of it.
So you're basically saying you don't care about the Google News traffic as much for these individual updates?
I think what we're going to do is balance it. That's a better way to think about it. As a digital news organization, Google News is important to us. But when we get somebody from Google News, oftentimes they're not regular readers. We're focusing more on trying to introduce them to our higher quality content and keeping them around, rather than trying to bring them in for these quick hits. We want them to understand the value of it. If they come here and just find a bunch of articles, they're not going to find the value of the stream. To be frank, I think we have been and always will be much more socially oriented than search engine optimized.
One question I have, because we have this same issue—is Google, both from a search and a News perspective—indexing the stream pages themselves?
Yes, they do. Actually, I was out at Google I/O this year and I cornered the head of Google News [Richard Gingras]. I've met him a couple of times and he's tired of me talking about this, but I'm obsessed with this idea because I really think we need to evolve our formats. You know how sometimes in Google Search results they show subheadings underneath the main heading? One day one of our Storystreams showed up like that where the sub-updates were little sub-links beneath the main result. It was the coolest thing. I thought, "If they start doing this, this is going to be amazing!" You'd be able to see what the current major headline is and see all of the sub-updates underneath it as a Google searcher, and dive right down to the update that matters to you. I keep asking him "how did that happen? Make that happen again!" He says he doesn't have any control over Google Search results. But for Google News, they could start to honor an idea like that. As more news starts getting streamified, what if you had a format for it? One of the things we do is publish RSS for Storystreams. The format actually speaks really well to this—you have a major title, and then subtitles.
So I can subscribe to the RSS for your Storystreams?
We don't expose it anymore because what's the point? But it's what you would expect it to be: The title of the stream is the title of the RSS feed, and then each item is contained within that. There's even a date updated on the feed. Actually, RSS has everything in it that you would need. But readers aren't designed for a feed that you would subscribe to in that way. So what I moved to instead is I worked really hard to try to talk Facebook into doing this. They actually kind of went along with it. If you "like" the stream, then we push the major updates into your Newsfeed. We do that through Open Graph. They said, "look, if you just 'like' an article, you can't publish through Open Graph. But if you say this is an object, then you can publish through the Open Graph."
Okay, I see it in the source code. Og:type is SBNation:Stream. Interesting.
Right. And they basically whitelisted us to do this, so as long as we don't spam users, they've allowed us to continue to do this. The way that we don't spam users is that on the backend some of these are major updates and some of them are minor updates, and we don't push major ones through. It's actually been pretty successful in that way.
It sounds like what you're talking about is the need for some kind of universal standard for search engines and social networks to index this type of story.
Right, exactly. We're getting really good at publishing, but what's on the other side of it? We're trying really hard with Facebook, trying really hard with Twitter. How do we actually evolve the publishing experience, and not just how it's built into our sites, but generally? I'd love for Google to be parsing things differently. I'd love it if there was more of a format for this. It doesn't seem like a complex one to me. All of the things that we're talking about from a product or technology perspective are pretty straightforward. It's really about the editorial production process and making it easy to understand how you get the value out of this from a reader's perspective.
That's also why I'm excited to see everybody experimenting with this. One of the biggest problems we had, and the reason we gave it the name Storystream, is that we were trying to explain to users what it was and how it worked. We were trying to put a label on it to signal to them "hey, this is a different thing." But I wonder how much longer this is necessary if more people start publishing this way.
[Image: Flickr user Sethoscope]