How to Determine the Next Best Product

One of the most practical and useful applications of database marketing for a business-to-business company is determining its customers’ next best product (NBP). The NBP is a guide for the sales force. Typically, the sales force tries to get companies to buy more of what they are already buying. That makes sense, and it is logical and easy to do. What many salespeople have found, however, is that their best customers are “maxed out” in certain categories. They are already placing all their orders for a category with this one supplier, and so pressures to buy more may not be effective. So what else do you sell to this company? Suppose you have 40,000 different products in your catalog. Which ones do you push? Until the advent of database marketing, the answer depended on intuition, observation, and luck. To these three, we can now add one more: next best product analysis. Next best product analysis will tell you what to sell to whom. It divides customers into similar buying groups, examines what each group is buying now, and then scores each customer on the basis of the relationship between sales to the customer and sales to other members of its group. It can be a very powerful guide for the sales force.

For a business-to-business application, how do you go about determining the next best product for each of your customers? Let us assume that you already have a database with information on your customers and their purchases for the past 2 years and that you have overlaid your file with SIC codes, number of employees, and annual sales. These data can be obtained from D&B.

Your first job is coming up with meaningful SIC categories. For each SIC group, determine your total sales. You may have to do some grouping to get something to work with. For example, Table 9-1 gives just a few out of perhaps a hundred SICs.

Table 9-1 : Sales by SIC Groups

SIC	Your sales	Number of companies	Average sales
A	$1,889,000	14	$134,929
B	$23,000	1	$23,000
C	$4,599,000	18	$255,500
D	$102,000	3	$34,000
E	$56,000	15	$3,733
F	$124	1	$124
G	$3,445,220	178	$19,355

I would not waste time on SICs B and F. They should be combined with others. You might want to collapse SIC E as well. You will end up with perhaps 40 meaningful SICs or even fewer.

Then I would divide the final SICs into categories on the basis of either number of employees or annual sales, depending on your preference. We can divide them into three categories as shown in Table 9-2.

Table 9-2 : Customer Companies by Number of Employees

			Number of employees
SIC	Your sales	Number of companies	Under 400	401–1000	1001+
A & B	$ 1,912,000	15	$ 552,001	$ 200,000	$1,159,999
C	$ 4,599,000	18	$ 45,000	$ 456,000	$4,098,000
D, E, F & G	$ 3,603,344	197	$2,001,000	$ 500,000	$1,102,344
H	$ 1,002,000	41	$ 200,000	$ 62,000	$ 740,000
I	$ 4,567,000	101	$1,002,000	$ 890,000	$2,675,000
J	$ 2,089,000	43	$ 45,001	$ 1,998,000	$ 45,999
Total	$17,772,344	415	$3,845,002	$ 4,106,000	$9,821,342

We now have divided our customer base into 21 different groups on the basis of SIC group and size.

What Do They Buy?

Next, we turn our attention to the products sold to these companies. We have to do some consolidation here as well. Some companies are buying 100,000 different products, and others are buying only 4 or 5. To make sense out of these numbers, we have to break these products down into meaningful categories. By meaningful, I mean that not only do they have to have sufficient sales to be worth bothering with, but they also have to have some relevance to one another. For example, we might have categories like those shown in Table 9-3.

Table 9-3 : Average Sales by Product Type

	Total sales	Total buyers	Average per buyer
Hoses	$ 6,001,000	220	$27,277
Fasteners	$ 778,001	198	$ 3,929
Valves	$ 1,002,000	18	$55,667
Pipe	$ 9,001,000	144	$62,507
Wire	$ 339,000	201	$ 1,687
Switches	$ 603,000	22	$27,409
Connectors	$ 46,001	32	$ 1,438
Total	$17,770,002	415	$42,819

Looking at these categories, I might concentrate on the four product categories in which the average sale per buyer is $27,000 or more. Depending on the situation, you might also combine some of the categories.

Now it is time to use a sophisticated model. We take each of our 415 current customers and determine that customer’s sales in each of our major product groups. We look at the SIC group and the employee size category. The model will create a next best product for each of the 415 different customer companies based on that company’s product purchases relative to those of the other companies in its SIC and company size group. Each of the 415 different company scorecards will look like Table 9-4.

Table 9-4 : Next Best Product for XYZ Company

				Total average annual sales
	Hoses	Valves	Pipe	Switches	Total
SIC group average	$38,002	$60,003	$42,001	$2,002	$142,008
This company	$22,001	$78,003	$ 1,001	$8,990	$109,995
Next best product			$42,001

What Table 9-4 says to the sales force is that the next best product to sell to XYZ company is pipe and that the annual sales could amount to $42,000 per year, based on the purchases made by other similar companies.

These numbers can be produced for every company and given to the sales force. Each salesperson can be measured not only on total sales but on the ability to increase sales in these conquest categories.