What value was there for DataSift to resell Twitter data and take a cut? What could possibly be in it for Twitter? In fact it's worse on two dimensions. First, financial (DataSift apparently gets an 80% cut). Second, DS gets the lock-in and "commoditizes" Twitter's offering, by putting it right beside a dozen other "sources". It's obvious that this deal was just just market evaluation or laziness on Twitter's part, or at best, wanting to keep a barrier between them and the data seller in case things went wrong. DataSift had a time-limited contract and it ran out. Their competitor (Gnip) apparently delivered a useful product by spending far less money.

(Of note is that FB's new sentiment analysis data product is being trialed via DataSift. This feels like a similar thing, with the added benefit that if it goes sour ("omg FB is literally selling our data/feelings"), they've got a slight degree of separation. Plus faster time to market etc. (Note that this didn't help DatSift's sudden loss of FB connectivity the other month, as FB dropped an API when they felt like it.)

On Meerkat, they were flagrantly disobeying the rules Twitter had set forth. They knew they were going to get in trouble, but "growth hacking". Related: Maybe my awesome search engine that just frontends Google isn't a great idea either.

That there's any sympathy or surprise here is confusing to me.

I strongly suspect that at the time they signed the deal (5 years ago), Twitter didn't have the technical capability to resell data (the site itself was going down lots!), so they had to do deal with DataSift and others.

DataSift previously ran TweetMeme, and did have good technical / scaling capabilities.

(I think I read that there was some other scaled technical thing DataSift did for Twitter too, but I can't remember what - so take this comment as a hint rather than facts)

I was thinking about Meerkat and DataSift last night when I read about the seemingly idiotic idea of "no-stack startups": http://blog.aweissman.com/2015/05/no-stack-startups.html. Might as well call them "no-moat startups".

Edit: already submitted here: https://news.ycombinator.com/item?id=9500424

There's a solid number of companies out there that have been contractors for bigger companies for decades. Nobody wants to talk about them because half of the tech industry press is all about high growth trendy companies when there's a lot of silent tech businesses that churn out cash reliably and for probably pretty boring stuff.

"Never, ever trust another company for your business model."

Or, to quote the '90s, "Where does [Microsoft] want to take you today?"