Have a problem, Big Data will solve it. The problem still is in data architecture and appropriate analytics. And most importantly, understanding the business reason for this Big Data (solve a problem, discover new insights, etc.)
Today's tools are cheap and powerful. For instance you can download the open source edition of Pentaho to your desktop. It will connect to numerous data sources including Hadoop. You now have a very large haystack to find a needle.
It's like have the world's largest junkyard and you want to buy a used 2001 Ford Focus with a broken water pump. You can design a data architecture that links to all the junkyards in three states, NAPA for new pumps and car dealers. You then develop the analytics to determine where to buy the water pump. And you find one 2 states over that can pull and ship you just what you need.
Problem solved, right?
What you missed was the fact that the water pump was very scarce. Why is that? Perhaps that model year had massive water pump failures. By investigating further you may have seen that model year had above average repairs and perhaps buying it in the first place was not a good idea.
Wrong needle, right haystack. And that takes planning with insight.