How to Compute the Next Best Product
Suppose your catalog has 40,000 products. Which do you push? Until the advent of database marketing, the answer depended on intuition, observation and luck. To these three, we add one more: next best product analysis.
This tells you what to sell to whom. It divides customers into similar buying groups, examines what each group buys now, then scores each customer by their relationship to the sales of other members of their group. It can be a powerful guide to the sales force.
For a BTB application, how do you determine the next best product for each of your customers? Let us assume that you already have a database with your customers and their purchases for the past two years, and that you have overlaid your file with SIC codes, number of employees and annual sales. This data can be obtained from D&B.
Your first job is devising meaningful SIC Code categories. For each SIC group, determine your total sales. You may have to do some grouping to get something to work with. You will end up with a small group of meaningful SIC codes. Then categorize the final SIC codes by number of employees or annual sales, depending on your preference.
What do they buy? From all the SKUs sold to customers, group them into categories. For example, a company with 40,000 products might group them into hoses, valves, pipe and switches.
Now comes the time to use a sophisticated model. We take each of our several hundred customers and determine their sales in each of our major product groups. We look at their SIC and employee size category. The model will create a next best product for each company based on its product purchases ranked against the other companies in their SIC and company size group.