You know by now that I am fond of examples so let’s use another one at this stage. Again imagine that you are that marketing executive at a mobile telephone company and you get a report on your desk every week telling you how many of your customers (they like to call them ‘subscribers’ for some strange reason) the company has lost to competitors over the previous seven days. Now of course this number is very important to the executives in the company, but it is a classic number from the ‘What Era’, often a classic report from a classic, traditional, dull old Data Warehouse.
So, we should ask ourselves, what is missing? Well there is one thing obviously missing, and that is any reason why these customers have left (churned). Wouldn’t you value the investment spent in your BI infrastructure if not only could it tell you who churned but just maybe it could have a go at telling you why they churned?
I will not lie to you here. What’- type reports are usually quite easy to create but moving into the ‘Why Era’ can be more difficult. Again, there are several reasons.
Firstly, we do not expect BI systems to help us with the ‘why’ issue, so tools aren’t too good at it. There is too much of this ‘just give me the facts and I’ll work out the rest (because that’s why I’m paid so much)’, and secondly, to understand the ‘why’ bit needs data that can be difficult to get. Standard reports just will not do, someone has to show some initiative here.
Just returning to our example, what might we do to enter the ‘Why Era’? Well, as a marketing guy I would guess that there are two things that could cause people to churn in large numbers: firstly, people churn because someone else offers a better product/price plan; secondly, people churn because of poor quality of service – too many disconnects etc.
Armed with this intuition, if I was the marketing guy I would ask for more data to be included in the Data Warehouse to better characterise the churners. I might organise some quality-based questionnaires to be commissioned to see what people think of our service, I might start collecting all competitor product/pricing information. I might start to gather all information on failed calls in my Data Warehouse and above all, I might ask someone with good technical knowledge to help me write the telling reports which would combine all relevant data to provide me with the ‘why’ point of view.
Armed with all of this I use my BI environment to advise me that the most likely reason why ‘Jon Page’ churned is because of a better price plan being offered by a competitor who has better coverage of Mr Page’s house. Get the idea?