That's may be true today, but one of features that distinguishes modern forms of surveillance from what was possible in the past is that truly massive amounts of data can now be stored "forever"[1].
Google itself may change what it wants to do with its data in the future. Worse, the fact that data was captured at all means it will be available for a subpoena or "national security letter" and Google will have little say in what happens to the data.
[1] the value of "forever" generally varies by the size of the available budget, which is "large enough" in the case of nation-states and companies the size of Google (or Apple)
The original question posed, though, was "Why do people at Google/Facebook/NSA continue to collect all this data? What's their purpose?", which I think I answered. While it may not be true in 10 years, it's true now.
Also, I think that the sheer pace of change today may insulate a lot of people from the consequences of having data available on them. The data may be out there - but by the time it gets into the wrong hands, it no longer reflects reality, which makes it largely useless. That guy who talks about having his dick sucked by sororisluts [1] may, in a couple years, be the guy with 200 million users whose company you desperately want to invest in.
[1] http://techcrunch.com/2014/05/28/confirmed-snapchats-evan-sp...
It's not about why you collect the data. It's that a massive database of such information exists. The intent in mind when building a system has no jurisdiction over the potential uses later.
If the NSA wants the data and it's inside of a US-jurisdiction entity, they will get it. Period. Maybe they NSL some people with legitimate access and make sure they know they go to federal if they let their boss find out. Maybe they recruit some SREs. Maybe they blackmail the night watchman in the building that holds the equipment that has the dark fiber L2 encryption keys so that they can get the inter-datacenter link monitoring going again.
It doesn't matter what people intended when they built it. The fact is that it exists and the men with guns and a money printing press want it. Game over.
And then I recall about Prism. And I think about how a secret program like that costs money and significant technical involvement from the host to run. Google took public money to help run Prism, and Google techs actively supported the dragnet. Pretty shameful, IMO.
As an ex-Googler who apparently can speak out, I ask you: if Google has competitors that are privacy focused, isn't it in everyone's best interest to select the competitors' products over Google's despite Google's excellence? And, can you report if anything has been done at Google to lift this stink of surveillance from its products?
Is Google still a willing participant in Prism-style data collection of its users?
Then the news about MUSCULAR [3] broke, and people were very, very pissed [4]. I saw the internal reports where people put together what must have happened; I'm not sure what sort of details I'm allowed to say, but it basically amounted to the U.S. (well, technically the UK, acting on behalf of the NSA) conducting military operations against the property of its own citizens. Google ended up encrypting all inter-datacenter traffic soon after. [5]
[1] http://en.wikipedia.org/wiki/PRISM_(surveillance_program)
[2] http://en.wikipedia.org/wiki/Foreign_Intelligence_Surveillan...
[3] http://en.wikipedia.org/wiki/MUSCULAR_%28surveillance_progra...
[4] https://plus.google.com/108799184931623330498/posts/SfYy8xbD...
[5] http://www.washingtonpost.com/business/technology/google-enc...
This news item you reminded me of - "Google ended up encrypting all inter-datacenter traffic soon after" - makes sense as a response that the government can't argue with (yet).
Looking over the MUSCULAR wikipedia article, I wonder if the NSA could still capture the user's and enterprise's unencrypted data at the Google Front End Servers like they had been before the inter-datacenter traffic was encrypted?
http://www.pcworld.com/article/2903192/safari-users-in-the-u...
http://www.reuters.com/article/2014/12/15/us-privacy-google-...
http://www.nytimes.com/2013/03/13/technology/google-pays-fin...
http://techcrunch.com/2014/10/01/hamburg-google/
Please understand, I'm not saying your wrong...I'm just saying from the outside [even as someone who works in IT and relies on data to make decisions] that I've never worked at a company that had any of these issues. It could be a question of scale and the fact Google is an easy target to make a headline with.
When they'd found out what evil thing they have collectively assembled, piece by piece, each of them blinded by his inexperience and enthusiasm for pure science, it would be too late.
And I'd finally have to replace them with expensive hired guns, or more clueless people, like some SV 20-something startup kids who would kill for 20 minutes of techcrunch fame.
Within Google, data's collected primarily to build better products. Or at least that's what the majority of the rank & file believe, and personally I believe the executives believe it too. This has the effect of widening the moat between them and competitors, but it also has real, tangible convenience benefits for users. I can guarantee that the majority of folks are thinking of those benefits, not the additional ad revenue, when building new products. I worked in Search for 5 years and was explicitly told by executives to ignore Ads when making decisions; they have plenty of their own engineers whose job is to optimize revenue.
This may seem hard to believe for someone who's never worked at a place that improved a product based on data, but it's a really fundamental part of the culture. You see the headlines about the NSA or Google's threat to privacy; you rarely see any news story about the A/B test they're running or the new feature in Android/Gmail/Chrome/Search that you get to use because usage logs showed that it took you too long to accomplish your task and some engineer decided to fix it.