Guidelines for competing with Twitter

Thursday, February 28, 2008 by Dave Winer.

Every time Twitter goes down I think of how can we create something to use when Twitter is down. 

There's a difference these days, because there's serious talk among developers about whether or how to compete. Earlier this week at lunch with one of them, I said I didn't know how great the opportunity was, given that Twitter has been staying up, through heavy use during the debates and primaries. Then, as if on call, it started going down. In the last 24 hours it's been down more than up, or so it seems.  

And of course that has re-kindled the back-channel.  

Some guidelines for potential Twitter-competitors. 

1. I don't like names with the word "killer" in it -- even in private. It's the wrong idea. No one will kill anyone. And names you use in private have a way of leaking out in public.  

2. Don't use the term crowdsourcing -- it betrays a perspective that's arrogant and wrong. I am not part of a crowd, I am a creative important person. Most Silicon Valley companies have this attitude. It's a good vector for competing. Our users are sentient human beings, individuals. Important not just as a collection of people. 

3. Most important, any service meant to compete with Twitter must be 100 percent compatible with the Twitter API. Porting of apps that build on Twitter means making the domain an option, where you call twitter.com it must be possible to change it to mytwitterclone.com, for example. 

4. It must be possible to use your clone when Twitter goes down and then switch back to Twitter when it comes back up with no loss of data. If you want smoothe entry into the market you must serve as a backup, earn your place with the users. Everyone will love you because it gives Twitter a very real concrete incentive to become more reliable. 

5. To everyone, twitter.com included -- this is a utility like email or IM. Reliability is key. If it's going to be used in business (very powerful idea) then auditability is essential. To assume that users love the product is not a good idea, any emotional connection becomes a negative if you can't keep the system up.