If you think having a great sales force is the key to your company’s profitability, think again. The key is what IT can do to make the most of customer data. 


The sales data we reviewed in the last post is just the tip of the iceberg when it comes to making customer sales predictable. Other technologies exist to cleanse and de-duplicate your data as well as glean further intelligence into your customers. This intelligence can then be used to provide more targeted campaigns. The theory behind these more targeted campaigns is that you’ll have a higher level of predictability with which you’ll be able to influence your customer’s behavior.

Here is how it works.

Data Quality: This is usually the first step when cleaning up your customer database. The de-duplication process sends your data through a pretty complex rules engine. It takes all of the data about your customer and matches similar records. Example, Jim Baker at 3000 West Main Street, New Albany, IN. The rules engine looks for records such as James Baker in New Albany, IN, and each variation of the street name (i.e. W. Main St., Main Street, etc.).

Some of the more advanced de-dupe routines will even take into account other known addresses to include summer or vacation homes and name changes as in the result of a marriage. Each match comes back as a score. For example, a score of 10 may be a very high probability of a match and a 1 a very low probability. You, as the owner of that data, can choose the acceptable probability level.

The value of the de-dupe process comes into play in several ways. First, it makes sure that you’re calculating the true lifetime value of that customer. Adding up all the interactions of the various Jim Bakers may move him up the profitability scale significantly, which means that your targeted marketing message may change. This also leads to reduced marketing costs. High-gloss, multi-color mail post cards or other snail mail marketing has significant cost associated with it when you include production costs and postage. Some pieces can be up to $3 each. Think about the cost if you do this over a year or two with weekly mailings at ten addresses that are really the same. You get the idea.

Data quality processes also standardize addresses to ensure that mailings get to where they were meant to go. Bounce-back mail and e-mail also costs money. Another advanced features is linking. This process takes into account address changes and links old addresses with new addresses. Advanced linking can even find relatives and their relations through a process called householding.

Once data quality has been completed, you can use some advanced analytics to perform clustering analysis or other analysis to find similarities in your most profitable customers. These similarities can be used to prospect for new customers that have a higher likelihood of being as profitable as your existing customers. But what prospecting lists do you buy?

Data Append: In comes data append. There are all kinds of append data you can purchase from companies such as Acxiom. This data can tell if your customers drive a Lexus or a Chevy. If they live in an affluent neighborhood or not. Whether they own a home or rent. All of these factors come into play and you get to choose what data elements you think will have the largest impact on whether or not your customer will profit your company or not.

Many of these companies have done a lot of the work for you with regards to customer segmentation. Acxiom has over 70 customer segments that they believe will influence how, what, and even when your customers will buy. Of course, all of this costs money. Typically there is a price per customer record and per data element.

Of course, nothing ever stays the same. Customers evolve. They get raises or they get laid off. They win the lottery and they go bankrupt. They get married and they die. You always have to keep up with how your customer changes. This is a constant effort. The challenge is timing the cleansing and the appending when it’s valuable enough to do it.