In the past -- in my typical skeptical manner -- I suggested that relying on the mere existence of an API was a bit risky when it comes to integrations. What has happened in recent times with Twitter puts a different spin on this topic, namely: what happens when that API just goes away?

Back in the olden-days (say 2008) Twitter was a lawless place. Almost completely open access to the Twitter API meant that just about anyone could build a service off the back of it -- such as a client application -- and consume Tweets. And as a result, just about everybody did. Some were good, so good in fact that Twitter ended up buying them and (partially) rebranding them as their own (a' la Tweetie and Tweetdeck). Recently, one of the investors behind the latter suggested that they sold up when they did because they could seen the inevitable way in which the chaos was going to play out.

Gradually, Twitter has turned lawman, initially turning its guns on the rest of the 3rd party client market. Since then, it's taken to legislating its way around those services that it believes -- correctly or not -- hurt its revenues. All the "Promoted Tweets" in the world were never going to provide enough cash to keep its own investors happy, the value is in (always was in) its data, and it's very difficult to sell something if you're giving it away for free.

As we noted in our recently released Digital Marketing research, when it comes to Social Media Monitoring tools, one of the critical areas to examine is how it goes about "Data Collection and Processing." What sources does the tool purport to use and to what depth is it able to examine those sources? A couple of recent announcements from Twitter have made that analysis even more prescient.

Firstly v1.1 of the Twitter API will mean that all access to it will have to be authenticated. No more anonymous access for applications. If you want access, you'll have to hope Twitter will give you a token. That same API update will now restrict the number of endpoints that can be called by said authenticated users to 60 per hour (from a current 350).

That of course doesn't mean that you cannot get access to much larger portions of that data, but in order to do so you'll have to buy access from one of Twitter's data resellers: Gnip, Topsy or Datasift. Take a look at their sample customer lists and you'll see many familiar names both from our research and -- in the other key recent announcement -- the "Twitter Certified Products Program." In order to get your very own sheriff's badge, you can apply to join (as long of course as you promise to remain a law abiding citizen and don't find some way to monetize the service that doesn't involve Twitter getting their cut somewhere in the chain).

If you are looking at Social Media Monitoring tools right now this should provide you with some interesting questions to ask of potential vendors as to how they populate their products with data from Twitter today and how they might be changing that in the very near future.