mikepk.com about current projects contact

Twitter clones

May 24, 2008

I thought I'd post why I've a sudden interest in Twitter and its problems. The truth is, we (Grazr) have some really good technology (IMHO) that could easily be repurposed to do Twitter-like things. As I mentioned before, we really over-thought the scaling issues so that happens to be one of our strengths (handling massive distributed data).

Adam and I discussed possibly building a twitter clone using the foundational pieces of our technology and open-sourcing the results. It wouldn't take us long and we figured we could do something good for the community and bring Grazr back into the technology conversation (since we get very little attention). The other reason is that we have a lot of open source roots, the team is made up of the guys who helped scale slashdot and release slashcode, the open source version as well as having some ties to MySQL.

The problem is that we were guilty of looking at Twitter through the technology lens. Twitter is not about its technology, it's about their oddly individualized experience and the people already in the system. There are already open source twitter clones, and closed source twitter-like services, yet none gain traction.

Duncan Riley has a post about killing twitter that I think hits on some of the main problems.

Dave Winer and a few others have previously discussed a distributed, open source Twitter that doesn’t rely on a centralized database. It’s a nice theory, but it ignores two realities: business model and people. Centralization is a business model that works as one company builds something they own, sure open source has its place but no serious startup (note startup, not big players) is going to build a system that doesn’t in some way provide them direct benefits going forward. The second problem is people: Twitter has them, every one who has tried to compete with Twitter so far (Jaiku, Pownce) doesn’t, and people only end up using services where their friends are. The bonus mix is competency: the dev team has to be competent at scaling, developing and running a Twitter like service.
(emphasis mine)

I think the open source arguments are assuming you're building a company that does exactly what Twitter does, and somehow expects to extract value from that service. I don't think that's an argument against open source, there are lots of ways to extract value from a product whose codebase is open sourced. In fact, you could argue even Twitter themselves have yet to find ways to derive value from their own service with a closed model.

I highly agree, though, with the second point and that's the primary reason why we decided it wasn't worth it for us to put together a clone. Replicating the technology, that's easy. Replicating the je ne sais quoi of Twitter and getting people to leave the established twitter audience, that's hard.

We decided that we would rather target some of our technologies to play in the twitter eco-system that's emerging rather than replace the twitter service. We actually think Dave Winer's suggestion that Twitter is a reef, is more apt now than ever before.

The noise regarding "killing twitter" is getting louder but I don't think Twitter is in too much danger yet. Having said that, there is a downtime-breaking-point at which Twitter could self-destruct. This would be a point where the Twitter experience of a majority of the users is irreparably compromised but I don't think we're there yet. Clearly the irritation of the "super users" is a bright red warning flag though.