The ‘What If’ era is predicated on an ability to modify the future in our aforementioned reality time. We have learned what has happened in the past, we have understood why it happened, and now we want to use this information to change what is likely to happen in the future. Let’s use a simple example:
I find out that fewer people than expected flew from London to Paris on a specific flight.
Why did it happen?
I analyse competitive information and discover that a competitor has just started the same flight but with a much lower introductory price to grab tourists.
How will I change things? (What if?)
Now surely the obvious strategy is simply to compete on cost, and this illustrates the beauty of BI and illustrates why BI is the cornerstone enabling the move to the ‘What If’ era. Let’s suppose we drop our price, we can assume that this will also drop our profitability, and in doing so we start a downward spiral that no-one can win except the customer. Before we know it, the competitor drops his price, we respond and soon this once lucrative route is costing money. This is where many companies are today because they simply don’t have sophisticated BI capabilities.
However, by deep analysis and predictive capabilities my chosen strategy is to replace several rows of economy seats with business class seats and market an offer of triple air miles to all business users and anyone else flying business class.
You might be surprised at the complexity of the response when you were thinking that the obvious response would simply be to lower the price to beat the competition, but the actual response that I proposed used deep predictive modelling to come up with a hypothesis that would not only maintain profit but might even increase it, and in a way that the competitor could not copy. Basically, the analysis enables the identification of a set of passengers that did not care about the price of the flight (within reason), probably because they did not pay themselves, but did value air miles for the twice-a-year private holiday. If all else was even, then they would fly the cheaper airline – but hey! – if there’s something in it for me privately, then let’s go back to the higher price and greater comfort.
Why can’t the competitor respond? Because they have nothing to offer, no air miles programme, no greater comfort, in fact all they can offer is something that not everyone evidently values.
Now I’ve made this example up, but it illustrates one thing: ‘what if’ is very difficult because the obvious actions to take to change customer behaviour are often the most damaging.
‘What if’ analysis is usually based on predictive modelling, data mining, and we will come back to it later. One thing that data mining depends on, however, is a rich source of historical and detailed data.