Each variable is expressed as a percentage of weighted historical spend. For instance, a customer that only ever purchases via the telephone will have 0% values for email, search, mobile, social, and your website, and 100% for phone.
A customer that purchased via search last year and via mobile this year might have 66% values for mobile, 34% for search, and 0% values for all other channel-based variables.
Finally, I add two 1/0 variables into the mix ... a variable that indicates whether the customer is a new customer in the past year (1 = new, 0 = otherwise), and a variable that indicates whether the customer is a loyal customer (1 = loyal, 0 = otherwise).
Up next ... the math that helps us map the customer life cycle.