But, more recently, in conversations with non-programmers, I see that 'big data' to them, means 'broad data' - it means trying to track everything possible and make sense of it. The average business user is really excited to be able to cross-relate disparate types of data - in an effort to make things better. 'Big data' enables the breaking down silos and enabling of cross references. It's about making empirical decisions based on data rather than opinion or intuition. That's really good, in my opinion.
So, 'big data' in that way is more amorphous than just the size of the data. With services and networks, the question becomes where does the data begin and end? Big data is potentially everything.
That's how language has always worked. There are people who are still uptight by the current "misuse" of words like "awesome" and "hopeful", but those of us who grew up with different meanings in common usage mostly just shrug.
For technical terms, usually I can live with words having domain-specific meaning that differs from common usage, but "theory" is one I still can't get over. It causes too much miscommunication.
eBay has a 9 petabyte DW that cuts across all of the types of data on their whole site: listings, bids, feedback, categories, clicks, etc.
Sometimes big data is actually big data, both in terms of raw size as well as complexity.
http://gigaom.com/2013/03/27/why-apple-ebay-and-walmart-have...
I realize that reporters like to throw buzzwords into anything to cater to the "simpler" readers. But come one, this is outright silly.