CNET también está disponible en español.

Ir a español

Don't show this again

Christmas Gift Guide
Tech Industry

Twitter business models in the fast and the long

Trends in real-time information and the desire to archive offer a window into business opportunities for Twitter.

Although this post is nominally about ways that Twitter could potentially make money, that's really just the springboard to discuss a pair of trends that will lead--and, indeed, are already leading--to markets for many products and companies.

The first of these is real-time--the increasing velocity of information if you would. Twitter as we know it today both reflects and reinforces this trend. For a lot of people, Twitter has already become the go-to source for breaking news. The accuracy and the depth of that news may be a matter of debate, but it's hard to dispute Twitter's emergence as a window into crowd reaction and eyewitness reports to events as they happen.

However, when it comes to certain activities like trading stocks, real-time is not a matter of a vaguely defined "pretty fast." Rather, it can be measured in milliseconds. Both the desirability and effectiveness of "flash trading" is controversial. Nonetheless, it's hard to dispute that there's commercial advantage to acquiring information even just a bit faster in many cases.

In the Twitter context, one could imagine a premium real-time feed and a free feed that was delayed by a minute. I've no doubt such a change would be hugely controversial although I'd note that it mirrors the way stock quotes were handled in the early days of the Web; free quotes were delayed by 15 minutes. Perhaps less controversial would be making an augmented infrastructure with a better and/or more predictable service level available to paying customers.

The other trend is archiving. The recent demise of the tr.im link shortening service, caused a fair bit of consternation over the fact that tweets which had used tr.im links would effectively be rendered worthless going forward. (Because of Twitter's 140 character limit, most tweets that link to Web pages make use of a service that encodes those links into a shortened format. However, the service has to remain operational to provide the conversion.)

Many then responded that Twitter as it exists today isn't really intended to be archival. Scott Rosenberg puts it well: "Twitter is great at 'now.' But as far as I can tell, it's lousy at 'then.'"

One can mitigate this lack on a personal level to greater or lesser degrees, but any such fix is limited. And, while much that takes place on Twitter is banal, much the same charge could have been leveled at Usenet in its heyday. And most would agree that losing those archives--as they almost were--would have been a loss to history.

The ephemerality of the Web may serve to obscure youthful indiscretions but, in general, it's something to be bemoaned rather than celebrated. And, services such as The Wayback Machine notwithstanding, it's not really being addressed in a systematic way today.

Truly long-term storage is another and thornier issue. However, I can imagine Twitter offering a paid service that did things like archive all the activity feeding through its service, thread replies, expand links, and perhaps even include a snapshot of a link at the time it was sent. None of these strike me as especially difficult to do (although implementing at scale would doubtless have its engineering challenges). And it could be extremely valuable for researchers looking at trends or otherwise mining historical data.

The big challenge with "freemium" business models is coming up with premium services that are valuable to some users but whose lack doesn't make the service unusable or uninteresting for those unwilling to pay. Striking the right balance is especially important with a service such as Twitter that depends on network effects to be useful. However, providing subscriptions that let users get information faster and to access it longer and more usefully would seem to be examples of value that Twitter could charge for.