Retail analytics has always been a combination of art and science; even before the data scientist buzzword. The main premise is the combination of quantitative (science) and qualitative (arts) data. If the collection process has been thorough you would have many data points for both, going back in time. Customers rarely ever do what they say they are going to do. Hence the importance of POS, transaction data to see what they actually did.
In todays vernacular this equates to a "journey." The 360 degree cycle of single or multiple shopping trips. This data is critical to define customer preference segments; communication channel preference and transactional channel preference. With the recent advent of digital channels for both shopping and communication the preference segments are not just nice to have, but a requirement to stay in business.
Add to this the X,Y,Z and Millennial generation that will far surpass the baby boomer in shopping disposable income and their preference for anytime/anywhere shopping. If you get it wrong, they will delete your app without a second thought. There are too many alternatives.
Understanding the customers' "why before the buy" is not a nice to have, it is a matter of retail survival.
Chris said it well: "All sources should feed a personalization for the customer." One step further, this data needs to be held in a centralized database.