Data exhaust, dark data and IOT
Not to start with the data -- going against Weber's advice -- but there's the dark data problem as well as the data exhaust problem. Dark data is the stuff that you're already collecting but not using to its fullest potential -- customer data and product data that's unused often falls into this category. Data exhaust is the stuff that you're simply throwing away either before or soon after collection, because the cost and trouble of storing it was too high by old stands. But the cost and trouble of storing data has changed with new platforms, like Hadoop and NoSQL databases, that make it easier and cheaper to work with such data.
I totally agree with Weber's start-with-the-business-need advice. But keep in mind that we're thinking up new possibilities because new platform possibilities have made it possible for everyone to dream big dreams. I also agree with Weber on IOT. I've seen a lot of examples of predictive maintenance and use of vehicle telematics at recent events including SAP Sapphire and MongoDB World.