At O’Reilly, we’ve long believed that one criterion for long-term success is creating more value than you capture. When you do so, you create the possibility for an ecosystem that’s larger than you are. You create a healthy environment in which you, your partners, and even your competitors can thrive.
Twitter, long one of my favorite companies, has turned the corner. After creating a very healthy ecosystem of third-party apps (standalone, web, and mobile), they’ve decided they need to shut down the market. In a post to Twitter’s developer forum, Ryan Sarver (@rsarver) clarifies Twitter’s terms of service and says new Twitter clients are unwelcome and existing clients ought to watch their backs. Further down in the thread, Raffi Krikorian (@raffi) steps in and does some damage control. While Raffi does a good job of moderating Ryan’s heavy-handed statements, I don’t think the position changes much. “No” becomes “it’s a bad idea to create a business where you would have to bend at the whims of another organization.” Exactly. When you create an ecosystem, you have a responsibility to that ecosystem. Perhaps it’s an ethical responsibility rather than a business responsibility. But that ecosystem is as much responsible for your growth as you are for its, and subjecting it to your whims isn’t prudent.
I’m personally saddened, at least in part because I’ve used Twitter as an example of how to build a successful data ecosystem. It’s a system that has contributed substantially to Twitter’s success, in a way that can, and should, be emulated. If you build the ecosystem successfully, the opportunities will come.
But now, Twitter has decided that it would rather not compete with its many children. Why? The reason given is that consumers are confused, and Twitter needs a consistent user experience. This reasoning strikes me as wrong-headed. Consistent user experience is a red herring. Is anyone confused by the diversity of user experience that’s out there now? It’s a throwback to the time when Word files only worked with Microsoft Word and WordPerfect files only worked with WordPerfect. But that was a pre-web world, and that’s not how the web works — never has been. If that’s what the future holds, we’ve got much bigger problems to worry about than net neutrality.
Are users stupid, silly, weak creatures who are easily misled? I don’t think so. As I’ve argued elsewhere, the proliferation of bad software guarantees that we’ll get some good software and occasional brilliant software. Twitter is ultimately a data publisher — and the many ways that data can be used is what makes the data publishing business interesting.
Twitter states that 90% of their users use an “official” Twitter client at least monthly. I’m sure that’s true, but it’s also misleading. I easily use a dozen Twitter clients, including clients built into other apps. While I use an “official” client daily, it’s probably under 10% of my total use. And that’s certainly part of the problem. I’m sure that Twitter is feeling pressure to find a viable business model to justify billions of dollars of valuation and hundreds of millions of investor capital. Promoted trends, tweets, and follower recommendations might be viable, though my reaction is often “You’re kidding; why would you even think I’d want to follow X?” Offered a choice, I’d rather not see them — and the clients I prefer don’t force me to see them. Is that the real issue? Other clients aren’t displaying “promoted” topics or users? Other clients give users choices that Twitter would rather they don’t have?
Are other business models for Twitter conceivable? Of course. In Twitter’s early days, I suggested to Blaine Cook (@blaine) that supporting private corporate Twitter networks was a possibility — a route that was latter taken by Yammer, and where there’s probably still a significant opportunity. Selling the “firehose” is also a possibility, and something that’s already happening (with Gnip the preferred reseller). Analytics services, facilitating customer service — these are all legitimate opportunities for Twitter. I’ve been surprised that they haven’t gone after them.
I am not saying that Twitter has no right to enforce some rules on the playing field it has created. Specifically:
- I have no objection to Twitter enforcing rules protecting the privacy of users, or restricting the proliferation of spam and malicious content. A client that edits tweets, changes profiles, or does anything else without the user’s permission should expect to be barred from the service.
- I also have no particular problem with Twitter competing with its ecosystem. There was some angst about Twitter “eating its children” when it rolled out the “new Twitter.” Tough. For a long time, Twitter’s own clients were sub-standard. Twitter has every right to build the best client they can, and to get as many users as possible to use it, as long as they maintain a level playing field. They only cross the line when they say “we won’t tolerate any more competition.”
- And I have no problem with Twitter saying that there are perhaps too many clients (I wouldn’t disagree), and suggesting areas in which developers who want to do something unique can experiment. Sarver offers a few such areas: publisher tools, tweet curation, realtime data analysis, etc. But here’s the rub: if you’re building publisher tools or selling analytics data, what’s to prevent Twitter from deciding they want that business three months from now, and shutting down the competition? Given the way Twitter has behaved toward its client ecosystem, I would certainly be nervous about starting an analytics business based on Twitter data, or funding such a business. Will the rug be pulled out from under me a few months down the road? I don’t know. Quora analytics, anyone?
But while Twitter can enforce rules on their playing field, they shouldn’t say it’s our field, it’s our ball, everyone else go home. And while there are plenty of uninteresting, dull clients, there is plenty of possibility for innovation. A year or so ago, I saw a Twitter client that displayed all incoming tweets on a map. Last night, I came up with several ideas for interesting, useful clients: threaded clients, clients that display tweets in the context of the user’s social graph, clients that use artificial intelligence techniques to find tweets that will be relevant and interesting. (Sort of like my my6sense, but for Twitter.) There’s no shortage of ideas. As Raffi says, developers ought to think big, not just come up with dull clones that simply display timelines. But he and Twitter are missing a fundamental law of innovation: you can’t tell people where (or how) to innovate, and where not to. Innovation just doesn’t work that way. The best way to prevent “think big” innovation from happening is to cut off the small ideas.
Twitter’s limitations on research (specifically, whitelisting and redistribution of raw data) are also troubling. Using social sites as a primary source for research is an important trend that we’ve been watching and publishing on, particularly in 21 Recipes for Mining Twitter and Mining the Social Web. While I understand Twitter’s restrictions on redistributing data for commercial purposes, and perhaps a desire to prune the number of users who can pull a large datafeed, it seems — as Inside Higher Ed argues — that these changes hurt academic research more than anything else. Twitter has been an enormously useful channel for finding out how history unfolds. Gephi.org has a wonderful visualization of the progress of the Egyptian revolution through Twitter tweets, and I’ve seen much similar work. Is this kind of research still possible, without buying a datafeed that a student or an underfunded department might find prohibitively expensive?
Screenshot from Gephi.org’s “The Egyptian Revolution on Twitter.”
Again, I see no problem with Twitter restricting commercial access. If you’re making money from the Twitter stream, you shouldn’t expect the stream to be provided as a free service. It isn’t difficult to write terms of service that distinguish between research and commercial use. Pure research is symbiotic, not parasitic: that research is one reason that Twitter has its multi-billion dollar valuation. In addition to informing us, it demonstrates what you can do with the Twitter timeline, and shows opportunities that Twitter (and others) can cash in on. We’re only now realizing what Twitter can teach us. Is it time to shut down further experimentation?
We have seen many companies thrive by creating more value than they capture. Twitter was one of the best. But once you’ve given something away, it’s hard to take it back, particularly if it’s something as fundamental as the right to create innovative software or to do academic research. Has Twitter said that developers can’t innovate? No, but they’ve demonstrated that they might tell you your innovation is out of bounds somewhere down the road. It’s disappointing to see Twitter undermining the ecology that made it valuable in the first place. Twitter would not be the company it is today without Tweetie, TweetDeck, TweetGrid, and many, many others. Likewise, Twitter won’t become the company it could be if it cuts off scholars’ ability to learn from its timeline. It remains to be seen whether a company that captures more value than it creates will inevitably stagnate. But if so, Twitter will be an unfortunate lesson, and we’ll all be the poorer.
- 21 Recipes for Mining Twitter
- Mining the Social Web
- Programming Collective Intelligence
- Restful Web Services
- Rest in Practice
- Web 2.0 Architectures