17.1. Creating a Customer Profile
When we select customers to receive the project offer, we are hoping to effectively predict what the customer's response will be to that offer. We define a profile based on characteristics we know about the customer. Frankly, we could segment and target by any characteristic, but not all are useful. Customers' first names, for example, are almost universally available. But a customer's first name is extremely unlikely to predict anything about behavior. Using customers' first names would not produce actionable segments.
Of course, no one would really choose “first name” as one of their segmentation criteria, but many companies (especially those that are product-centered) do something similar. They try to segment their customers based on what product the customer purchased. It's not that the purchased product isn't extremely valuable information; it's just that it doesn't really tell us much about who the customer is. For example, XYZ's consumer customers didn't buy large-scale products, but those darn business guys were always buying products designed for consumers. “Purchased product” isn't great for segmentation, but it can be very useful for targeting (such as for an upgrade offer sent to all those who own a certain product). We need to look at the characteristics that do help us understand customers.
17.1.1. Using Marketplace Information
Marketplace information about the customer is easily obtainable, often from third-party sources such as list compilers and brokers. This demographic data is available for both business and consumer customers. The data is generally leased or purchased for a one-time use as a contact list or to add/validate profile characteristics to your own file(s). Remember that using a brokered list as a source of contact names, though certainly a widely used direct marketing tactic, is not CRM. Of course prospects can become leads who can then become customers and then you can begin building and managing their relationships.
The work to append additional profile data to your customer information can be accomplished by either your internal database team or by the data providers themselves, who are often well equipped to provide you with this kind of service. Table 17-1 gives examples of some of the demographic data that is available for purchase.
The advantage of using marketplace facts about customers is that they are fairly easy to obtain because these facts are generally observable. This data is available for purchase, so why would any company bother to collect this data directly from customers? Just buy it, and save your company and your customers much valuable time. The major disadvantage to marketplace data is that the same information is available to all your competitors, so how can it be any kind of advantage to your company?
| Type | Consumer Customer | Customer Company |
|---|---|---|
| Individual | Age Sex | Function Title |
| Financial | Household income
Years of education Own/rent | Revenue
Fortune rank Capitalization |
| Size | Number in household | Number of employees |
| Demographics | Average cost of homes in area Median income in area | Industry Region |
| Psychographics | Interests Preferences | A “best place to work” company Company style (top down, consensus) |
17.1.2. Using Relationship Information
On the other hand, only you can know anything about your company's interactions and relationship with customers. This is where your company's real competitive advantage can be found. This information can be either objective (numbers of support calls, size of orders) or subjective (vendor preferences, opinions, feedback). It is usually related to behavioral and attitudinal facts based on the interactions that have taken place between the customer and your company. Because relationship data most often describe events that occur directly with customer contacts (business or consumer interactions), there isn't a significant difference between the types of characteristics captured. The big difference is in the way we aggregate some of these facts. For B2B customers, the individuals we market to are often not the same people who place the order. Instead, we have to assume a link if the company is the same. Because of that, the link between contact and company is a major data quality focus for all B2B marketers. Business contacts inherit many of the characteristics of the companies they work for (as indicated in Table 17-2).
As you can see, business-to-business marketers have different criteria (information, aggregation) for segmentation and targeting. Other differences exist as well. Many companies that sell to consumers don't sell to them directly. The valuable purchase data (point of sale) may be available through third parties. Of course, most B2B companies also have partners that sell at least some of their products. It's important to work with the channel to make sure they understand that you have not abdicated your relationship with your end customers to them. This is a very important source of customer information which too many companies have ignored in the name of avoiding channel conflict. It takes work to turn the channel relationship into a partnership, but for most companies it must be done.
| Relationship Information | Aggregate by | ||
|---|---|---|---|
| Consumer | Company | ||
| Response Behavior | Order, point of sale (from partner or other third party) | Individual | Company |
| E-mail offer responses | Individual | Individual | |
| Interests | Product inquiries | Individual | Individual |
| Product ownership | Individual | Company and individual | |
| Hobbies, sports, etc. | Individual | Individual | |
| Product Uses (Applications) | Reason for purchase or use of product | Individual | Company or individual |
| Number of users | Household | Company | |
| Where used | Individual | Company | |
Another significant difference between B2B and consumer data is in the numbers of individuals and amount of data available. The vast amount of consumer data available always requires the use of automated analysis tools. There are differences between consumer and B2B types, and sources and aggregations of information, but we will see that the methods and tools are the same.
Relationship information gives you knowledge that only you have access to, and because it is behavioral, it's generally much more appropriate for defining and targeting customer segments. Past behavior predicts future behavior better than any other information we can get. The disadvantage is that it takes lots of time and resources to collect enough of this information in your database to do much good.
17.1.3. Using Calculated Information
If relationship data is so much more effective, why do we even bother with marketplace data? The answer is that we don't always have enough relationship data available to determine customer segments. Often, we can get more complete results by using both market and relationship data together. We can combine information in different ways.
Derived Information: One way to improve the accuracy of our segmenting and targeting efforts is to use relationship or marketplace facts in combination with other facts or to summarize them over a period of time.
Inferred Information: The fact is we often use market and relationship information in combination whenever we can. We use relationship facts to determine value, and then use market facts to find other customers who look just like our target customers and (we hope) behave the same. We can infer behavior even for potential customers who have never bought from us before and who have no relationship with us. These new customers can be from internal (e.g., leads) or external customer information sources.
There is a progression of scenarios from simple to sophisticated that most companies adopt as they are building up their CRM capabilities. The three scenarios described here show that progression. It takes a while to build the level of database content required for Scenario 3.
Use marketplace data plus common sense and experience to identify customer segments that will identify actionable groups of customers with similar value.
As relationship data becomes available, use this interaction information to verify segment value and to test and refine your segment decisions.
Look for marketplace characteristic similarities that predict well which members belong in which segments (data mining). Then look for other customers and prospects that match these market facts but don't have relationship facts to include in the same segments.
All three types of information can be used at different times and in combination, depending on the specific project goal. There are definite tradeoffs. Demographic information is readily available and fairly easy to use, but not very predictive. Relationship data is a much better predictor, but it only helps with customers for whom we have captured the required relationship attributes. Calculated information is extremely powerful because it is the best predictor, but it depends on tools and expertise that are often not available. Calculated data also has the advantage that it can be used for “new” customers (prospects or customers for whom you don't have much relationship data).
A number of tools and algorithms can be used to calculate new information from existing information. They are covered in more detail at the end of this chapter. These tools are used for both segmentation and targeting.