A couple of days ago, Twitter announced some improvements to its API and Terms of Service (TOS). Effectively, it’s raising the limit on the number of calls per hour that each application can make and opening up its Firehose API (already available to Google and Microsoft) to other developers as a paid service.
This certainly opens the door to even more innovation within the Twitter ecosystem (currently 50,000 apps strong). Nevertheless, I can’t help wondering if the additonal operational demands of handling the increased traffic might not best be placed in the hands of a company with that type of focus so that Twitter itself can focus on “higher layer” innovation.
Though Twitter has made some strides over the last several months in improving the performance of its infrastructure, changes to its TOS and API limits will increase traffic substantially. I’ve always been impressed by the number of things Twitter needs to get right day in and day out. It not only provides search and other functions to give users insight, just as Google does, but unlike Google, it also has to host all the data it draws from (tweets). That’s no trivial challenge for such a small company, as is evidenced by the multiple instances of traffic overflow messages over the last years.
Leaked documents earlier this year showed us that several big players — Google, Microsoft and Facebook — have shown an interest in taking this problem off Twitter’s hands by acquiring the company and integrating its information into Google Search, Bing or Facebook Search. But the Twitter executive team clearly hasn’t been interested in an acquisition.
Still, I’d posit there’s a good chance Twitter execs might be willing to form a tight infrastructure alliance in 2010 with one of the industry heavyweights. And while Google, Microsoft and Facebook would all be good bets, I’d actually wager on someone like Amazon and, more specifically, Amazon Web Services (more on that in a minute).
Twitter’s founders have expressed their desire to invest the time required to build out an independent company. To do this and maintain its momentum, the company needs to execute on a number of different levels as it continues to look for ways to offer users added value (e.g., search, etc.) from its platform. All of this is a tall order for any management team and becomes even more daunting when one adds to the mix the operational demands associated with this vertically integrated strategy.
Probably the scarcest resource Twitter has right now is its management’s time and bandwidth. So help with operations would free some of that to focus on higher-level value.
From Twitter’s perspective a deal with Amazon Web Services would offer a number of opportunities. One possibility would be to use Amazon Web Services as a hosting environment for part or all tweets as well as a partner for managing a large portion of the thousands of developers in the Twitter ecosystem. Twitter would require the same kind of substantial operations prowess that Google has developed to cull insight from a growing ocean of data. Nevertheless, hosting tweets and reliably making them available to a user population that increasingly depends on them in realtime, might be better left to another.
From Amazon’s perspective, integrating the Twitter API into one or more Amazon Web Services offerings would further strengthen the company’s position in the nascent Cloud Computing space. Applications that integrate insight from personal tweets and, possibly, leverage Amazon’s recommendation engine technology, could provide more relevant information and advertising to users. More importantly for Amazon, the power of aligning the Twitter brand with its own might be enough to recruit a good portion of future Cloud Computing developers who might have sat on the fence for a while longer.
Whichever way you cut it, 2010 will certainly be a year of critical strategic decisions for Twitter. It might choose to tackle all fronts alone. Or it could look for options to reach its objectives more rapidly and change the industry’s competitive dynamics in the process.