David Raab Article Archive

Although many small businesses have been successful with database marketing, the vast majority have never tried it. One barrier–though not necessarily the main one–has been the cost and complexity of database marketing software. The challenge has been building a system that is simple enough for a non-technical user, but still powerful enough to be useful.

TargetSmart! (TargetSmart, Inc., 303-698-2233, www.targetsmart.com) attempts to meet this challenge. It can be evaluated in terms of three core functions: create a list of customers and their transactions, select customers for promotions, and monitor promotion results.

The first function, building a customer and transaction list, is probably the most demanding. Today, most small businesses probably capture their transactions in some type of accounting or cash register system. But they may not also have a customer list or link transactions to individual customers. TargetSmart! provides no real help in this area: it requires delimited ASCII files with customers, transactions and linkages already in place. Users are left to create these files with whatever other tools they have available. The system does make it fairly easy to map the import files to the TargetSmart data structure. This structure includes fields for customer name, address, and telephone; up to 18 user-defined customer attributes to segment the file; up to six additional customer elements for personalization; and basic information on each transaction. This is a tiny amount of data by today’s large-system standards, but should be adequate for most small businesses.

TargetSmart! greatly simplifies maintenance by loading all customers and transactions from scratch with each update of the database. (The system does carry forward promotion history by customer across updates, to enable response analysis.) Reloading the data avoids any need to reconcile the new and old versions of the database, but it also means there is no way to track changes in customer attributes or to include accounts that have been dropped from the current version of the source files. Transactions without matching customer records are simply dropped, although the system does report on how many such records it found.

The second database marketing function is selecting customers. TargetSmart! automatically divides the customer base into four groups, based on purchase frequency and total spending–the Frequency and Monetary Value in direct marketers’ traditional Recency-Frequency-Monetary Value (RFM) scheme. A comprehensive 181-page textbook included with the system explains appropriate marketing strategies for each group. TargetSmart! presents each of the four segments as pie charts, and lets the user slice the pies with one additional variable at a time. These variables may be related to customer attributes, the types of products purchased, or transaction dates. Users can also “filter” the pies to show only customers with one value of a given attribute–say, customers who purchased a certain type of product.

Users select segments for promotions by “dragging” pie slices onto a list. This is easy but precludes segment definitions more complex than the pie slices. The system will eliminate any duplicates if the same individual qualifies for more than one segment. Users can change the priority of the segments by shifting their order in the list.

Promotions are attached to segments by choosing letters, file formats, envelopes or labels. The system comes with over 100 sample letters, categorized by marketing objectives such as “build purchase frequency”. It also provides tools to create additional outputs, including a simple word processor. Promotions are assigned separately to each segment, which is flexible but tedious if the same output is desired for many groups. The system will look at a segment’s size and promotion history and automatically suggest whether to send to a test sample, hold out a control group, or send to the full group. The user can override these suggestions if desired.

Once a segment has been used in a promotion, TargetSmart! ensures that all its members receive the same additional promotions until the file is next updated. This simplifies response analysis, but also may exclude qualified customers from some promotions. For example, if a promotion is sent to all customers in a neighborhood, anyone in that neighborhood would be excluded from a later promotion based on purchase history. This is no problem if promotions are carefully planned and prioritized in advance, but makes it difficult to react to unanticipated opportunities between updates.

The final system function is response measurement. TargetSmart! does not use source codes to tie specific transactions to individual promotions. Instead, it considers a “response” to be any transaction by a segment member during a period starting with the promotion date and ending after three additional file updates. This means if the same individual is sent another promotion in any of those later periods, any response will be credited to both promotions. The vendor plans to let users allocate responses based on date ranges assigned to each promotion, but even then some double counting would be possible.

TargetSmart! calculates key statistics including the number of people contacted, number making purchases, response rate, average revenue per respondent, and average number of purchases per respondent. If a test, control or limited distribution strategy was employed, the system will compare results among those who received the promotion with those who did not. It does not show promotion cost or profit margin, however. Results can be viewed on bar charts or tables. The system can also report on trends in key performance statistics for the four main file segments from update to update.

TargetSmart! is a visually attractive package, with an intuitive interface and built-in guidance for novice users. Its limited file building means the system cannot be used by everyone. But where customer and transaction data are available, its selection and measurement capabilities make it easy for users to get started with database marketing. This is probably enough to justify the system’s $495 cost, even though serious users will soon outgrow its limitations.

TargetSmart! runs on a Windows 95 PC. The system was officially introduced in September 1997 after beta tests at about 40 stores. The $495 price includes technical support with no annual maintenance fee.

David M. Raab is a Principal at Raab Associates Inc., a consultancy specializing in marketing technology and analytics. He can be reached at draab@raabassociates.com.