Despite starting Blogger, Evan Williams rarely blogs. But yesterday, for the first time in several months, he decided to put the digital pen to the digital paper in order to lay out his thoughts for Twitter’s new Retweet functionality. It’s a great view into the mindset behind what is already becoming a controversial change.
Why is there so much controversy? The answer is simple — literally. When Twitter began, you could do one thing on it: Send a blurb about what you were doing in 140 characters or less. This led to an immediate outcry from a wide range of people who thought that it was just about the dumbest service in the world. Others saw the potential behind such a simple service, precisely because it was so simple, and history has proven time and time again that sometimes simple ideas can explode into the biggest ones.
As Twitter grew in size, its simplicity remained largely intact. While just about everyone had ideas for what features Twitter should add, Twitter stayed the course in its core simple vision. Instead, it decided to rely on both its user base (@replies, RTs, etc) and third-party developers to add functionality. In fact, at points, Twitter began removing features (auto-refreshing, IMing) because it simply could not scale with so much load on its servers.
While some might view this as a failure to innovate. I would argue that this adherence to simplicity is what brought Twitter to where it is today. We live in an age where feature-bloat reigns supreme. Far too many startups replace the word “better” with “more.” That is to say, rather than perfecting the product they have and maintaining a singular focus on what they want to accomplish, they keep adding new features either because rivals are doing them, or because users are suggesting them. This is rarely a good idea. One great feature beats a dozen half-assed ones any day of the week. Keep it simple, stupid.
That’s why the past several weeks have been so interesting for Twitter. With its scaling problems seemingly now solved and with enough funding in the bank to buy a small European country, Twitter finally gained the flexibility to address a terrifying question: What’s next?
The answer actually started coming several months ago when Twitter finally wrapped Twitter Search into its core service. By all accounts, that has been a rousing success. Twitter Search is now perhaps the key way to get a pulse of what the web is thinking about at any given time. Microsoft, Google, and Yahoo have all recognized this, and all now either incorporate tweets into their results, or soon will.
The job is hardly done for Twitter in this regard. The success of search has led to a massive influx of spam, which Twitter now must combat in both the search results and the Trending Topic area. The solution for both, it seems, will eventually be a move to highlight tweet relevancy. Part of this will undoubtedly be based on user reputation (which will be controversial), another part is likely to be based on location (more on that in a bit).
Then there is the question of links. Sharing links is one of the key uses for Twitter, and the company has said in the past that it would like to create a way to better search those links. Twitter has run tests on and off to track clicks on links, but it remains unclear if their intention is still to incorporate links into search results in a way more similar to what the big search engines do. That could be a further off goal now since a few of its bigger third-party developers, like bit.ly and Tweetmeme are also working in that space.
But search was just the first major change to be implemented. These past several weeks have seen a second, and much larger round of changes. It’s been a phase that I’ll call Operation: Features.
The first of these new additions was Lists. Despite fears that spam would overrun it, or that it would simply be a new vanity benchmark, Lists seem to be a hit. They are proving to be great both used as filters, and for user/content discovery.
There is already a robust developer community swarming around Lists, to both add them to existing services, and create new ones around the functionality. That’s great news for Twitter because it reinvigorates their already strong developer ecosystem.
And, if you’ll allow me to go against my “simple” mantra for a second, it seems like Lists are a natural addition to Twitter just for the filtering capabilities alone. Any community that gets big enough absolutely needs a way to filter. Sure, you could argue that you can just do that by not following anyone that you absolutely do not want to see in your stream at all times, but if Twitter continues to grow and more people you want to follow continue to sign up, that’s simply not really a realistic proposition. Filters, in this case, Lists, are.
Another wave in Operation: Features will come shortly in the form of Geolocation. While the ability to add location information to your tweets will undoubtedly raise privacy concerns, it’s important to note that this feature is entirely opt-in. And while the usefulness of such a feature may not be immediately apparent to a lot of users, there are a ton of things Twitter and third-party developers can use this for to make the service more useful.
For example, as I alluded to above, imagine if you want to do a Twitter search query but want to tailor the results to a specific area. You might say that was possible before because of the location information in users’ profiles, but that is hardly real location information. I could say I’m in China, for example, and no one would know that I’m not. With geotagged tweets, a whole new layer of information about happenings in particular places is possible. And Twitter clearly knows this, as it has just released an API for it.
We’re also on the verge of seeing an explosion in interest in location-based services. Foursquare, Gowalla, Brightkite, Loopt, and Google are all already in the game. Facebook is likely to enter soon as well. With its Geolocation API, Twitter has the opportunity to become the go-to place to send your location information when you want to broadcast it widely (well, probably not from Facebook, given their little rivalry). Or, to put it a different way, this feature could allow Twitter to become the de-facto location social graph.
In terms of your social information, like it or not, location will eventually play an important role. Twitter has positioned itself well here.
But here’s where things get interesting. So far, none of these stated new features have or will significantly alter Twitter’s core simplicity. Each of these is an add-on. They are either shoved to the side (literally, the right-hand side of the page with Search and Lists), or are opt-in (Geolocation). And with third-party apps, these also do not alter the core experience in any meaningful way. But Retweets are an entirely different story.
As we’ve seen with a limited roll-out over the past few weeks, Retweets alter not only Twitter’s main stream, but also users’ actions. That is why Evan Williams wrote his blog post. And that is why we’re already seeing backlash.
As a general rule, users dislike change to the services they’ve grown accustomed to. But that dislike could be amplified on Twitter for two reasons. 1) They haven’t made such a significant change to the core functionality before. 2) Twitter is an extremely simple service that just got a little more complicated.
As Williams notes in his post, users have been asking for a Retweet button ever since it entered the lexicon of Twitter organically, rising out of the collective. But rather than give everyone a simple RT button that might work like the current Reply button (pre-populating the tweet field), Twitter decided it was time to use this new function to alter the user experience.
There are two key things that people don’t like about this: 1) It potentially inserts a user you don’t follow into your tweet stream. 2) You can no longer comment on the retweet.
Strangers In My Stream
To the first point, Twitter is doing more or less what Yammer CEO David Sacks laid out in a guest post on TechCrunch several months ago (which Williams acknowledges). On one level it’s actually not really different at all from what the current retweet mechanism is doing — which is, putting the words of someone you may or may not follow into your stream. The difference, of course, is the icon. Previously, it was the icon of the user doing the retweeting, now it’s of the original author. This can be unsettling since it alters a previous constant: Twitter’s main stream.
Already, we’ve seen users completely and utterly confused as to what is going on with these new retweets. In one high-profile example, Mallory from Family Ties believed users were actually paying to insert themselves in her stream.
As I wrote at the time, this could well be a harbinger of what’s to come. It doesn’t matter if the actual tweet information is fundamentally the same, what matters is the user perception. Users were used to an unwavering constant: That their stream would contain only tweets (and again, their pictures) from those people that they follow. By breaking that, Twitter will have users questioning the integrity of their stream.
Eventually, most users will get used to this. And remember, from Twitter’s perspective, they believe that there are still many more users who have yet to sign up for the service than those who already have and now have to learn something new. And there is also a way to turn off these new-style Retweets from showing up in your stream, but only on a user-by-user basis. It may be wise for Twitter to introduce a universal off-switch, even if that’s silly and against what it’s trying to do, just to give users comfort.
I Want My Say
The second point may actually be even more problematic for Twitter: Users want a way to include their own statements in Retweets. The new way of doing this does not allow for that. The fundamental principle behind this should be obvious: If you share something, that’s a natural desire to explain why you’re sharing it. That’s what a lot of people do with current retweets. Even if they just add “LOL,” it shows that they think the tweet they’re sharing is funny.
We’re also vain. Sometimes retweeting something is more about getting your say in rather than simply highlighting what someone else has said. Or, maybe you’re even retweeting something because you disagree with it. With the new Retweets, you can’t let that be known.
But Williams says that they’ve thought about this commenting functionality for the new Retweets, and notes it could come down the line as an add-on. I suspect Twitter may have to do something like that if it’s serious about getting everyone to use the new Retweets because otherwise, people are just going to keep doing it the old way.
Two Roads Diverge
And that’s another important thing to note in all of this: Nothing is stopping you from retweeting the old way. But that raises a different problem. Because of that, we’re likely to see a bifurcation of the Retweet. Some will do it the new way, some will do it the old way. In fact, it seems perfectly plausible that people may start to use the two types of Retweets for different reasons.
There’s nothing inherently wrong with that except that it introduces confusion. And again, when you’ve been sold as a simple service, that’s the last thing you want. Perhaps it would have been a better idea to call this something other than “Retweet,” even though that’s what it is. Maybe you call it “Highlight” and see if users (who again, invented the retweet) switch to using it more than manual RTs.
This would seem to go against my belief of tacking-on new features, but as Williams lays out, clearly Twitter believes this new Retweet structure will fundamentally change Twitter for the better. In that case, a new feature is absolutely worth it because it’s extending core concepts (discovery, trackable data), while cleaning up problems (noise, attribution).
Box Of Chocolates
In fact, none of this is to say that the Retweet change is the wrong call to make. When I first read about it, before even seeing it in action, it made perfect sense to me as a concept. But what I am pointing out is that this is a risky call to make. Actually, I would say that it’s the riskiest call Twitter has made yet.
Remember, this is Twitter taking a component that was invented by the users, and altering it. While Twitter has adopted user-created elements in the past such as @replies and hashtags, they didn’t actually alter anything about what the users had created, they simply enhanced them. (And when they did sort of alter things, by removing the option to see @replies for people you don’t follow, there was a mini-shitstorm.) They’re enhancing Retweets too, to an extent, but they’re also changing the idea and making it more complicated. It might ultimately pay off, but again, it’s risky.
But it would seem that Twitter is doing this at an opportune time. As has been discussed a lot in the past few months, Twitter’s growth has been slowing. If everything were still growing like gangbusters, it would probably be ill-advised to change core functionality. But in a slow period, it’s easier to justify risks, and all of these changes are things that could actually help kickstart growth again.
Since Twitter has started Operation: Features, I find myself visiting the actual site all the time now, whereas before I would often use a third-party client. More importantly, I find my overall usage has kicked up a notch. That’s the key. Twitter has known about the problem where people would sign up and then not continue using it, all of these things should help. (Maybe none more than a small new feature: That damn new tweet notification box, which has me refreshing every few seconds.)
But here’s the key question that will play out over the next few months: If a service is predicated on being as simple as possible, and then moves away from that, even just a little, will it still be the same service? Will it maintain its appeal with its rabid user base? Will it pick up new users?
Facebook changes all the time, but it’s ultimately okay because Facebook has always been a mess of things going on. They’ve improved some messes, made others worse, but they didn’t go from simple to that mess. Facebook was not built for one purpose, Twitter was.
Is Twitter still Twitter when the features roll in? You never know what you’re gonna get.
[pictures: Paramount Pictures]