products
The principle is straightforward enough; find out which customers are more likely to buy your stuff and concentrate on them. The method and data used to build the model will be key when it comes to how powerful the model is, and therefore how much it costs you to make those sales.
But propensity modelling and scorecards are much commoditised these days, which is something of a shame as there is art and skill involved in creating powerful models.
The analyst who understands your data, your objectives and how to build a model from scratch using the most powerful variables is your best friend here. They need to be aligned intellectually to the commercial objectives, so even if they do go all gooey eyed over some extreme maths they call elegant and beautiful they know that you really only care about the performance - think BMW engineers and BMW drivers!
