Often, database marketing campaigns start with a calendar based on previous years' campaigns, with ad hoc campaigns added based on product overstock conditions. Next, the owner/marketing manager determines which products should be part of the promotion, a choice often largely made based on the type of campaign. For example, if it is a "Household" campaign, then house wares, linens and other "Household" products are promoted. The specifics of the offer, such as price discounts, also are resolved at this point. Finally, after all the specifics of the campaign are decided, the database marketer chooses which customers to promote to.
Exceptions to this process may happen when there is a strategic reason to target a specific group of customers. For example, a database marketer might decide that the corporate strategy should be to develop a more loyal customer base among upscale young women. In this case, customers would be chosen for a campaign based on these attributes.
Either way, the campaign still is planned first, without regard as to whether this is the right promotion at the right time for the right customer segment. What is wrong with this approach to promotion planning? First, it ignores customer purchase cycles. The optimal time between a customer's promoted purchases may not coincide with the promotion calendar.
Second, there are not enough controls over the number of times customers are part of campaigns. Controls can be set using suppression rules (i.e., do not select customers who have been mailed to in the last 30 days), but these are often arbitrary and applied uniformly across customer segments.
Third, by finding the customers to fit the campaign, rather than planning campaigns around the customer, many valuable customers never will be adequately promoted to, and you will not get an adequate return on your investment in database marketing. Mailing costs soar; there are high levels of customer attrition; and the company spends a huge amount of money to maintain a customer database that is not providing as much value as it should.
Also important is the way you segment the customers. Most businesses stick to the traditional, tried-and-tested route such as corporate, heavy users, etc. We advocate the use of customer motivation as a basis for segmentation.
An alternative to planning and implementing database-marketing campaigns based on a campaign-centric orientation is to move toward a customer-centric approach. This approach maximises the database marketer's return on a sizeable investment in the customer database. This consists of several stages:
1. Start with customer segments: Segments can be defined any way the marketer usually segments the customer base, using recency, frequency, monetary value (RFM), demographics, types of products purchased or any other method. A valuable approach is to incorporate promotion sensitivity into the segmentation scheme.
2. Review your current calendar with respect to your customer segments: Look at the number of times customer segments are contacted, expected customer spending and the level of marketing expenditures projected to be allocated to customer segments over the year. From this, you can estimate an expected return on investment (ROI) for each customer segment. This provides a base line of information against which you can measure customer-centric promotion effectiveness. At this stage the promotion is still campaign-driven; however, database marketers have gained the ability to understand how many and which campaigns are being assigned to each customer segment.

3. Use a model-driven approach to campaign planning that takes the interactions between campaigns into account: The intent for this approach is to create a promotion calendar for each customer segment that most closely optimises customer segment profitability. This is not based on "pure science." Business requirements may limit the desirability of planning campaigns in a purely customer-centric way, so database marketers must incorporate business and production constraints.
The advantage of taking a long-term view rather than planning each campaign independently is that it allows the system to evaluate the effect of a stream of promotions on the customer. In the short run, there might be a good ROI from running several campaigns one after the other. However, this reduces profitability over the course of the year.
Using this approach, we must look at all past campaigns and develop a model, one for each customer segment that estimates revenue and profit for the customer segment. Each past campaign and its attributes - promoted products, time, markdowns, media mix becomes a variable in the model. We must also include the number of times the customer was promoted to.
The model becomes a basis for a sensitivity analysis. The campaign planner can determine how different promotion schedules and different types of offers will affect profitability for each customer segment. Over the year, the plan can be updated and ad hoc promotions added based on an ongoing response analysis and merchandise opportunities.
4. Allow the system to build the campaigns - including timing, number and specifics of the offers - to optimise profitability across campaigns, based on business and production constraints: Use the customer segment model previously developed as the basis for the optimisation procedure that will lead to a database-marketing schedule for each customer segment. This can be modified by the campaign planner.
The real challenge when making the transition from campaign-centric to customer-centric database marketing is melding the calendars from the two approaches. Experienced database marketers always will need to make choices on how to do this. This process we use moves the database marketer in a customer-centric direction, in which products, timing, markdowns and other details of a specific offer are planned around the customer segment. Database marketing efforts then can focus on making sure valuable customers are marketed to in a way that ensures long-term profitability.

[Innovative Media is a knowledge management consultancy. The Author writes for a number of well-known publications world-wide and has several published works to his credit.]