01-25-2015 08:07 AM
Can can someone help me please?
I'd like to segment my customers for marketing communications, some customer purchase in a single business whilst others have purchased across multiple businesses. The problem I have is the transaction amounts vary significantly from one business to another. One business is fast moving consumer goods (food retail) and another is high priced insurance, for example. Is there a function or process to create an overall score by applying a weighting to value (£) and then aggregating each score for each business to create one overall score?
Any help much appreciated.
01-25-2015 02:08 PM
Thank you to all for responding.
Reeza, I shall add the question to the other forums. You make a valid point, it is a broad subject ..
Kiwi, Market basket analysis is not what I'm looking for. I have an RFM scoring technique in place already for food retail consumers which ranks my customers well, however, I can't apply RFM on insurance products as they tend to be less regular (yearly in most cases). Perhaps the technique should be one score for each business and then weight those scores to an overall score? If that is possible to do?
01-25-2015 03:33 PM
bob you are saying your RFM scoring technique in place for retail. See a relative short time-frame for the events.
Your remark that you cannot use that for insurance because the events are based yearly is nonsense. Your remark is telling you you should adjust the time-frame. 5 years 10 years, 30 ? what is present and applicable for that "insurance product".
Customers value and customer segmentation will follow the business goals.
- No interest in long-year planning? Drop those "insurance products"
- Just cross-selling insurance as extended product life support. Put those in retail
The customer value and customer segmentation should reflect those business goals. How much do you expect to be the profits on some customer.
01-25-2015 12:03 PM
Your question is too broad. There are books written on this subject.
You should also post this in the Statistical or Data Mining forum instead of general discussions.