Twitter's Policy Changes Don't Add Up

A few days ago Twitter sent what has been widely interpreted as a "fuck you" announcement to third party application developers who create applications which allow Twitter users to talk with their friends. They claim it's about a unified user experience. They also claim that it only applies to new applications, not existing ones. The thing is, there are logical errors here which don't add up.

The user experience argument is that application developers can change things which confuse end users, such as I could create an app which calls Tweets "Bills", and the Retweet button "Rebill". End users would be confused by this Bill concept and be looking for the Tweets. This argument falls apart with the Darwin school of thought. If it's so badly or confusingly designed to use, people won't use it, so it becomes irrelevant. There is no need to protect users from it, as they will choose to protect themselves. Not only that, but they will seek to protect their friends, family, co-workers etc. In short, these applications will fail spectacularly because they're not hitting the right notes. If on the other hand these applications succeed, then they are obviously not confusing the end user, as the end user has worked out that Bill is Tweet etc. In which case they don't need protecting either.

Different clients have different features so there is some inconsistency there however they are all competing with the primary client ie the official Twitter client, which presumably has every feature Twitter offers in all wrapped up in a code filled miracle bundle. If third party applications can't compete in terms of features or usability, or even compatibility then people will choose the official client. This also means there is no legit reason to protect users. By "compatibility" I mean "Twitter have deemed that only Windows will be supported, OSX and Linux users can fuck off", then there is a niche third parties will seek to fill. You can't protect users who you've turned your back on, it's also disingenuous to suggest it's in their interests too.

If there is such a need to protect users from UI / feature confusion, then why continue to allow existing clients? Surely they are the reason Twitter decided to create this change of policy in the first place, that those existing clients were different enough to cause user confusion. By allowing those to continue they have undermined their whole argument.

They also say that around 90% of their user base tweet from the website, which seems like an incredibly large number to me, but it could be true, and could also explain why it's too busy to respond a lot of the time. They say that the trend is increasing, that people are choosing more and more to use the website rather than clients. This is supposed to emphasise the point that "people want consistency, since clients don't offer that, they head for the website, which does". The problem is that if the percentage is so high and rising naturally, it again undermines the need to act, since it's self repairing.

Being someone who has to use Windows from time to time, I tried to find a decent Twitter / Identi.ca client, and have to say I didn't find anything even remotely decent. Maybe I have high standards, maybe the "Identi.ca" part of my needs narrowed the choice down too much, or maybe the Windows clients are just terrible. If the clients are so bad, it's a large part of why people just use the website. In the end it's like voting, nobody knows why you vote for the party you did, just the numbers who voted for them. Plenty of party officials will claim God-like knowledge of why people voted for or against something, but unless the voter tells them, they're just guessing.

One of the reasons given was about "not being able to modify the stream" which reads to me as a euphemism for "we're gonna sell sponsored tweets, feed them into people's streams without saying they're sponsored tweets and we sure as hell don't want third party apps filtering them out, or letting their users filter them out". That seems like the primary reason for this sudden change, it has nothing to do with a consistent user experience, that's just the PR spin to distract from the real reason.

As much as Twitter is ubiquitous, it's also a company with a need to make money. To my knowledge they've struggled to find a business model for what is a funky service. Yes they have a LOT of users, but they have no real business model to monetize them. People who visit web site and tweet from there could see advertising, but they could also block adverts as I do. Plenty more people use a client, meaning they only ever visit the site when they need to change something in their account like their avatar or email settings. They don't see adverts at all, so there has to be some other way. Sponsored tweets is something they've been pondering for a while.

How many people will put up with sponsored tweets in their streams? Microblogging is a kinda personal social thing, many people will regard tweets offering goods or services arriving unsolicited into their streams as spam. If this is Twitter's business model, then those advertisers are Twitters real customers and the users are their product. Twitter will be responsive to their customers, since it's their customers who pay them money. Twitter will not treat them as spam, no matter how many users complain. If their client of choice can't filter those out how many people will abandon Twitter?

It's not as if the microblogging concept is unique to Twitter, arguably Identi.ca is much better, it's open source and it's very responsive in dealing with spammers. It's also decentralised, so if Status.net (the company behind the Lanconica software and the public instance called Identi.ca) decide to inject sponsored Dents into your streams, you can install it on your own server and choose not to connect to the main Identi.ca instance, giving you all the benefits without the drawbacks.

The gamble Twitter feel they're big enough to make and win is that people have built enough of their online networks around Twitter that they will either not notice / care whatever Twitter decide to inject into their streams, or they will have little choice but to grit their teeth and accept it, locked in because all their friends and family are on Twitter. Time will tell. It does seem like a very arrogant and disingenuous move though.

As for me? I use both Twitter and Identi.ca. I interact with Identi.ca which cross posts to Twitter. If you want to talk with me you're gonna have to use Identi.ca as the replies and DM's don't cross post back to Identi.ca. If Identi.ca goes away I'll simply install my own instance of Laconica and carry on as usual. If Twitter goes away overnight it's no biggie, I use it as an extra one way service, and those who use Twitter will be scrambling around for something similar to replace it, which leads us back to Identi.ca. It has a sort of symmetry to it doesn't it?