Segmentation

I always knew the most difficult and the most important step in marketing once you are clear of your business objective is to determine criteria for segmenting the target customer base.  From a classical economics point of view, the utility functions should decide the dimensions; but most marketers have been too dumb or their study of marketing had started with Kotler and ended at Kotler.  Psyhographic or behavorial segmentation based on parameters derived from difference in utility functions tends to be the ideal way of segmenting customers.  Most marketers however on account of lack of data and conceptual clarity revert to mundane and more tangible parameters like demographic segmentation.  Even that is perfectly fine, if their intent to use demographic or other variables is to create a surrogate of behavioral or psychographic segmentation model and try to get as close as possible to the ideal situation.

Another major fallacy that Kotler hasn't clearly addressed in his writings is whether your require a single segmentation model or multiple.  Well, you require a segmentation to be done initially for determining which segment you are going to target, may be another segmentation for defining which subsegment of your overall segment of customers would you target a particular promotion campain ( and the promotion campaing may be targetted at achieving sales or just simple improving brand health).  Infact the truth is that for each and every activity that the marketer intends to do ( whether come up with a new SKU, a new discount scheme, a new advert., a new promotion campaign), he again requires to go through the rituals of segmentation all over again.  Creating micro segments is something on which Kotler and party sold out.  But the busy guys that the marketers are trust their guts most of the times rather than getting into the technicalities of trying out further segmentation.  Also lack of a quality MR agency and their exorbitant rates makes it difficult to repeat such exercises time and again.

And today, I am in a peculiar scenario where I lose one fifth of my customers every month, gain back one third of the customers whom I lost this month over the next 3 months, and gain 20 % new customers every month.   I end up in a scenario wherein how a long a customer has been with me doesn't bear any probability of how long is he going to be.  I have absolutely no demographic data for any of my customers.  And it doesn't help my customer base is in millions.

So what ????  Put on your thinking cap.  Get back to basics.    Lets segment and segment further and segment further...buss...itna hi...jyada segment karne ki bhi zarurat nahi hai