I have 3 discussion questions that I'd like to throw out to the group.
1) How do you leverage your inbound channel and customer analytics to drive intelligent cross-sell and up-sell initiatives, ensuring the best opportunity is presented to each customer? 2) How much of this process is being done in a real-time environment and how do you derive the business case to justify the IT investment for real-time? 3) Can you share some examples specific to your company/industry?
I work in telecommunications. We use customer telephony usage data and customer product holdings for analysis of churn/attrition, up-sell, cross-sell, loyality rewards/freebies, etc. Multiple product holdings usually create more 'stickyness' and lower churn/attrition rates. Some simple numbers relating to lower churn might prove a sufficent business case.
We use a Teradata warehouse (and data mining tool as a UI) to analyse data usually somewhere between 1 day and 6 months old (depending upon the analysis). Much of our analysis is run monthly, so in some cases our product holdings info might be a little out of date by the time it is communictade to the customer.
My team don't often examine history of product purchase because the products change so frequently. Instead we try to examine each customer's telephony usage and geo-demographic data to ascertain suitable current services, or similarities with other customers that themselves purchased a specific product (clustering).
For the us the greatest source of information is the transactional detail (thier telephony usage). Teradata is great in that it can actually store and process many millions of rows in a practical timeframe. This enables us to use a lot of detailed behaviour information about the customer and ensure product recommendations that honestly fit the customer's needs. Only where we are very confident of a need and benefit to the customer do we contact them in any outbound activity. Our criteria for contacting the customer are quite strict as to prevent needless harrasment.
For inbound, we score customers (in a simple flat table in Teradata) with a product recommendation where suitable. The CRM system simply displays this product cross-sell and up-sell recommendation when a customer contacts us.
-- my suggestions for you -- Use detailed data where possible. Examine which customers buy a product or service (or multiple) on their own 'free will' (or without as much effort by marketing) and then use data analysis to find other customers that have similar behaviour but have not purchased the product.