When we talk about a multi-dimensional view of the customer - we refer to view of the customer purely from the business perspective. Usually, these “business perspectives” are related to profitability, risk, responsiveness, loyalty, behavior and preferences.
Therefore, to be able to see a single customer along these dimensions would undoubtedly give an organization an incredible competitive advantage in being able to serve customers better and in return being awarded by customer’s larger share of his wallet.
And this can certainly be achieved by using predictive analytics. Since output of predictive analytics are often probabilities (in the case of binary or ordinary target variables) we would be able to see probabilities that customer will respond, buy, default on credit, leaves us for competition, become valuable customer, or anything else that is in interest to us. If we add to this customer purchasing affinity and purchasing behavior – we have all ingredients to devise the best specific action toward that customer.
Let's say - if a customer is of low value to you, who cares about his loyalty and preferences? You don’t want to waste a cent of marketing budget on him. In fact, you want to open the door as wide as you can and let him go. On the contrary - the customer in the highest value segment whose loyalty scores are dwindling deserves to be phoned by your top account managers to see how you can improve your service to him. And if you know his preferences and buying habits - you know what to offer him to have him re-think his intention to leave you.
This assumes that his initial grievances that lead to his decision to leave you – are addressed. If your product is too costly for the real or perceived value of that customer is getting – analytics will not help you change his mind. Improving quality for the better price will – but at least analytics have brought you that customer, and how you going to respond – is up to you.
So we can see that - at this level, analytics are no longer used for decision support – they are used for decision making. This is what I call “holy grail” of big-data analytics. Unfortunately, only few are doing it in this way, and for those companies who use analytics at that level and for those purposes – pay-offs are huge, but you may never know about it.