I’ve been fascinated by the blogosphere’s musings on Twitter’s downtime. I, for one, am relatively sympathetic toward their plight – there’s nothing tougher than trying to scale a system that’s growing quickly and has a very unpredictable, bursty traffic profile. And as much as I like and enjoy Twitter, it’s not a life-or-death kind of service for me.
While most people seem to be making off-hand remarks along the lines of “aside from uptime, they need to figure out the whole business model thing” I think it’s more complex than that.
*Twitter users have an expectation that the service will be free. My favorite article on this point is this one from Dan Farber at CNet.
*According to most estimates, a large volume of Twitter traffic comes from 3rd party services (I’ve heard estimates as high as 90%)
Combining these two points, I think the real challenge for Twitter will be that whatever business model they look to deploy will have to accept these two realities. Once you’ve set a service expectation with users, it’s hard to change it. For a non-technology example, look at the hue and cry over American Airlines assessing a baggage fee to fliers who want to check a bag. People who travel view the ability to check a bag as part of the service. Regardless of whether the $15 fee is fair, it feels like a violation of the contract between fliers and airlines. Similarly, I think Twitter can’t do too much to impoverish the free service. I can think of a laundry list of possible business models for Twitter and I’ve listed them below:
Freemium / Paid Subscriptions – I’m not sure what premium services Twitter could provide. I don’t believe that you can make guaranteed uptime or higher uptime a premium service until the free service has acceptable levels of uptime. This isn’t web hosting, where there is a willingness to pay for more uptime. If you’re going to provide a service, free or paid, users have some expectation about regular uptime. Also, having a service where “free” users don’t get good quality service doesn’t help Twitter grow.
Advertising – The challenge here is that people Tweet from all places – mobile phones, IM clients, 3rd party clients, browser plug-ins, etc. If the stats on the amount of twitter traffic that happens off the core website, a real revenue opportunity around advertising would require that the ads live with the tweets. I’m not sure how you would accomplish such a thing. Perhaps Twitter will come up with some clever way to attach an ad to the end of each tweet and 3rd party services will be required to serve / show those ads to continue getting access to the API. There are a million difficulties with this approach, so I’m not sure how they’d actually make such a thing work.
“Tweetmining” – For lack of a better term, I’ll lump all approaches that focus on evaluating the content found in tweets to provide some kind of intelligence (market research, trends, brand feedback, etc) into a meta-term that I’ll call tweetmining. I don’t think there’s much of a business model here because the very same data that Twitter would want to mine is generally available to anyone else who’d want to search it. There are already Twitter search engines and other Twitter analysis services that take advantage of this information to give people insight into what’s happening on Twitter. Unless Twitter has access to relevant and valuable information that the general Internet does not, I’m not sure there’s much of a business here.
Licensing – I could see a model where Twitter partners with other companies, social networks in particular, to provide the ability for those services to have some kind of mobile update capability. They already have a plug-in that works for Facebook. But with most of the major social networks opening up their platforms, that negates the opportunity to do a revenue-producing licensing deal for the most part. This would also boost the need for increased attention to uptime.
Those are, as far as I can tell, the major business model opportunities. And, as far as I can tell, none of them are terribly attractive if Twitter’s goal is to build a large, revenue generating service. What do you think?