I'm a customer of Gnip's - they provider several enrichments and data normalizations on top of Twitter's stream. Could Twitter have hired X engineers and duplicated what Gnip does? Perhaps - but Gnip also has an existing customer base, existing revenue flow, infrastructure setup to handle customers creating their own matching rules etc.
There's more to Gnip than just re-selling the Twitter firehose.
Will these sites want Twitter to have firehose access to their data? For example Tumblr, Foursquare, Wordpress? They also offer Facebook data, but it is a managed public API connection.
Capital is raised by companies to accelerate growth. Growth acceleration can manifest itself in many ways. Specifically, in this case, as you rightfully point out, Twitter could easily build this service themselves. They also, at the time of Gnip's inception weren't totally sure there was a business model here.* So they let another organization take that risk for them, instead of using their own capital. They end up paying more capital in the long run but the risk is reduced.
*I recognize that most likely the certainty of Gnip's business model is (1) much higher due to the "duh, of course people want this" nature of their business and (2) that the people investing here are all "in this together". It, however, doesn't negate the fact that there is indeed risk.
Twitter provides their raw data to multiple outside companies. These companies compete with each other to add value to the raw data, while Twitter focuses on solving its core problems. Everyone wins. (Except Twitter's users, who make these companies millions of dollars without a penny in return.)
Interesting: It seems twitter now does see distributing their data as a valid businessmodel (before they 'outsourced' this to DataSift and Gnip). Wonder what it will mean for the other sources Gnip has, and for the customers of Datasift.
"To that end, we have agreed to acquire Gnip" - I'm wondering for example if the "agreed to acquire Gnip" means that Twitter doesn't see this as a high-value acquisition on the same level as Crashlytics or Mopub.
I found that phrasing weird in the blog as well: "Today I’m pleased to announce that Twitter has agreed to acquire Gnip".
Makes it sound like they almost forced the acquisition on Twitter.
I'm not sure that Gnip is a 'high value acquisition' for Twitter though - AFAIK they're basically just re-streaming Twitter fire hose data which Twitter could do themselves. This is probably just a way for Twitter to lock down/get even more control over their ecosystem.
I guess that makes Twitter data cheaper than before? Last time when I contacted them for one week of worth tweets (with some filters on top) they wanted me to pay tens of thousands dollars. I ended up crawling the data myself.
Interesting, I had gotten the impression Twitter was aligning itself more with DataSift in recent years. I used Gnip a few years ago and thought it was a great service, although the high price point drove us to ultimately roll our own.
Yes, Twitter streaming API plus several other data sources (Facebook, Reddit, forums, etc.). It's enough for us since we're always filtering by keywords or users. We were also granted elevated access.
It was pretty straightforward in our case. We had to show them that we had use cases that weren't feasible with the default access level and also demonstrate that we weren't competing with their user experience at all. Not sure how much it mattered but we had been using their streaming API for almost as long as it existed.
I had the same question when I read this. The streaming API is only a fraction of the data offered by the firehose, but that's not as much of a problem in every use case.
Not surprising to see Twitter finally acquire Gnip after working with them for four years. A great achievement for the team at Gnip, and will be interesting to watch the data model they intend on developing with Gnip as time goes on.
Question: is Twitter's free "garden hose" sampled data stream still available?
I used to use the small garden hose sample of near real time tweets. I just tried running an old script to fetch the garden hose and it didn't work.
On topic: I thought that the Gnip business model was interesting, and I can't help but think that the other data sources that Gnip processes and resells may not like this aquisition.
Are you sure the Garden Hose has free access? I can't find any public access to it in the docs, everything directs me to their "data partners", which charge for this information.
He is being downvoted because it is crazy to expect websites to work around the bugs in various 3rd party plugins. This is an issue for the plugins to fix.
I would argue that relying on Javascript to display what ought to be a simple static site is the problem here, but I also agree that perhaps the plugins ought to recognize what's going on and allow that JS to run.
