"clearly we definitely could be clearer if there is so much confusion" (How Twitter backed away from a platform strategy)

That was a tweet on August 16 by Ryan Saver, Twitter’s platform leader. It was in response to the kerfuffle that emerged regarding Twitter’s proposed changes to its API. I saw the discussion and debate emerge on various tech blogs and it was hard to know what to make of it. Marco Arment, who runs Instapaper and has been integral in other ventures such as Tumblr, was particularly distressed. I wondered if it was misplaced. Twitter relies on third party developers too much to seriously jeopardise that, right? But then last night he tweeted:

If Twitter cuts off Instapaper’s “find friends” access, I’ll probably remove all Twitter functionality. It’ll hurt, but it’ll feel right.

That worried me. I often tweet from Instapaper but don’t use the ‘find friends’ feature. Arment was willing to crimp the product as a protest and one that would give a leg up to his competitors. I figured that it was worth spending a little more effort to look into the economics of what was going on.

Let me report my conclusions first. Twitter are removing commitments they seemingly made at their founding and are putting themselves in a position to exert much more control over developers who interact with the core Twitter assets. It is a dramatic shift away from the pure platform strategy that Twitter was previously pursuing. More to the point, it indicates that they made a set of errors in formulating that earlier strategy but also that, sometimes, you can’t just correct errors and expect everything to remain the same. In an effort to lock down their own core assets from future competitive intrusions, Twitter are creating an environment that could accelerate those competitors.

I’ll try and explain this by first reminding everyone what it is that Twitter does. Twitter is a communications network. Users can send messages (tweets), they can read messages and, now (and this evolved) they can converse (replies) and share (re-tweets). Twitter is unique in that individual messages are restricted to 140 characters. That means that it has been positioned as the most real time of the social networks.

So far so good but then comes the issue of monetisation. For a communications network there are several options. First, they can charge for participation. The problem with this is that they get less participation and for start-up networks that can mean death. App.net is now trying that path but it arguably wasn’t open to Twitter at their founding and the whole idea had yet to be proven. Second, they can charge for a certain type of participation; for instance, the ability to send messages more broadly than a user’s followers. This is something we could term advertising. For Twitter that might mean embedding sponsored tweets in people’s feeds. Finally, they could charge for access to deeper information. Twitter knows lots about messages, trends, and social graphs. This is valuable information and someone may be willing to pay for it.

All of the Twitter’s moves are best seen in light of preserving options for monetisation. First, Twitter moved to acquire the most popular third party apps that people use to read and share messages on Twitter. I emphasise read rather than send because Twitter have, at the same time, been encouraging apps, and indeed operating systems, that allow people to tweet. So Twitter acquired Tweetie and TweetDeck and made them their own. At the same time, it has sent strong signals to others that it will not be giving those apps free reign in the future. That may effect some 30% of users. And the moves last week turned a signal into reality. Third party apps will have to comply with strict rules as to how tweets are displayed (presumably to allow Twitter to place ads in feeds and collect information about them). They will also be limited in use before Twitter exerts control. Get more than 100,000 users and you have to have a quiet word with Twitter. In other words, when you get big Twitter is likely to play close attention.

What’s going on here is fairly obvious. Third party apps are a threat to charging for participation (i.e., sponsorship and advertising). Controlling them defends the platform. To be sure, companies do this all the time. However, in Twitter’s case, it is an apparent reversal from what they have encouraged previously. Indeed, developers are rightly concerned. Independent developers were responsible for lots of successful innovations that helped Twitter develop. For starters, they saw apps rather than Twitter’s website as more valuable to consumers. And they were the first to integrated other services (including the ability to read content later, host pictures and shorten urls) that Twitter later adopted. They also helped give the ‘re-tweet’ function more power. Interestingly, ‘Retweet’ is not an entry in Wikipedia (it redirects you to the Twitter entry that contains no information on its origins) but I found this entry that does do this. It conformed mostly with my recollection. Users invented the ‘RT’ function in 2008 and app developers put it in as a button. Twitter then adopted it officially and, right now, are asserting control over anyone’s ability to use it.

This means that should Twitter start earning money from ads, it can ensure that third parties do not siphon off that revenue either by grabbing it themselves or developing apps that remove ads or sponsored tweets; e.g., Tweetbot has a function that can mute tweets. But the cost to this is that it puts a check on third party development in general. Here is what Twitter is saying: “go develop but, really, you should consider Twitter as your ultimate customer.” Invent something great and Twitter has pretty much the exclusive option on purchasing it. Why? Because if they don’t like it, they can shut you out. No app developers might want to compete to have Twitter as a customer but that competition is much less powerful when app developers do not really have outside recourse. It also means that it is highly unlikely app developers for Twitter will be seeing Instagram or Zygna type rewards. And you can be pretty well sure that apps that allow multiple social network interactions (say including Twitter, Facebook and Google+) will be off the agenda. LinkedIn, for example, can no longer display user tweets to others on its network. And apps like Storify that curated Tweets in an innovative fashion panicked. Twitter moved to assure them but, frankly, given past behavior it is not surprising that those assurances are less than fully accepted.

Second, with regard to its ability to earn revenue from its information, Twitter, will be concerned about giving that information out too freely. That is why it has moved to block others from accessing information regarding user’s social or communication graph. This week that meant that Instagram and Tumblr both had to drop features that allowed users to find their friends using Twitter. They can still do so for Facebook and Google+ but Twitter is basically saying that it does not want entities to know social graphs. What is going on here is that Twitter are likely worried that a competitor may come in and be able to easily replicate their social graph and migrate users to some alternative. Do that and Twitter loses its informational advantage. The same issue exists for Facebook and Google but they allow these activities and have opted to compete on other matters. Twitter appear to be engaging in a more extensive lock-down strategy.

These moves are moves away from openness. They dramatically change the incentives of developers who interact with Twitter and who, in the past, had added value to it. Does Twitter need these things? It is hard to say. As they were arguably the source of Twitter functionality in the past, the smart money would suggest that this makes them vulnerable. But then again, by locking down the platform in this way they can secure what they have including any network effects that exist. That can help them build a sustainable business model even if, as a user, the future does not look as rosy as the past.

But there is a stronger point here. When you pursue a platform strategy that relies on investments by independent parties, you have to make commitments that leaves money and opportunity on the table for them. Facebook could have extended its credits to help it in allowing it to earn money from apps but has instead retreated from that. That forecloses a monetization opportunity on the belief that the network will grow and they will make money elsewhere. Once you make that commitment, reverse it is extremely costly as it prevents you from changing where you make commitments in the future. The point is that if Twitter wants ‘help’ it is going to have a hard time getting it now. Of course, at this point, it is not at all clear that Twitter’s strategic pivot involves others despite their two dimensional chart that seems to indicate otherwise. I, for one, can see how developers in all four of those quadrants can be a threat to some future aspect of Twitter monetisation.

In a sense, when you found a company like Twitter you have to make a choice at the outset. One option is to start off smaller and make fewer commitments. In many respects that is what Facebook did. It did not rush to open its API and when it did it controlled the terms. As it turned out, there was enough user value to build the network despite this. Twitter started off with a bigger bang on openness. As it turned out, they perceive that today as limiting future revenue opportunities and are now reversing course. But at the same time, there are others waiting in the wings to see if they can pursue a Twitter like idea but with more openness. Twitter’s latest efforts have taken a group that once gave it their full attention and caused them to start looking elsewhere. How this plays out will be interesting to see.

Oh yes, and it may be that simply quoting a tweet as I have done in the title of this post could be blocked. That should disturb those who write tweets.