Re: vending machines
Ordering history is very basic and dumb. What time of day are various products selling? What are the weather conditions when certain products sell? What's selling by which location? Are items in danger of running out such that the machine can be restocked in time? If customers are paying by credit card or smart phone, what can you say about customer segmentation by product?
As for equipment maintenance, is the product being properly chilled, heated, dehumidified or otherwise protected and served? If you need to go to the trouble of sending out a repair person to fix X, what other services should be completed at the same time in anticipation/prevention of other service problems?
In short, when Ruh talked about "deep" analytics, these were a few of the possibilities he had in mind.