This is the conclusion of a three-part series on the fundamentals of a successful database marketing program. The first two articles, “Getting Ready” and “Blueprint for Success,” appeared in the January and February issues.
The creation and management of your marketing database were covered in our last issue. Now, armed with a marketing plan, product priorities, an offer inventory and a technology strategy, you’re ready to define how you’ll use the database to create inquiries or sales.
The development of a plan for campaign testing doesn’t happen in a vacuum. It needs to be driven by hard information. Since it’s unlikely you’ll be implementing true one-to-one communications, you need to think about your approach as “one-to-few.”
Grouping customers or prospects into homogeneous segments will let you address them in a way relevant to their needs, wants and desires. There are many ways to handle segmentation. What’s usually most effective is modeling scores based on their behaviors (responses and/or purchases) as they relate to your direct marketing efforts. Of course, if you’re just starting to develop a database marketing program and you don’t have experiential knowledge to work from, you may have to opt for a segmentation strategy based on what others in your business are doing, a clustering method developed from commercially available segmentation schemas, recency/frequency/monetary ratios, or simple logic.
Once the segmentation strategy is in place, you’ll need to devise your creative by targeting the different segments of your audience. If you’ve done the job right, you’ll probably discover the need for variations in your creative by segment, since it’s likely the features and benefits that will appeal to one subgroup may not work for another.
Finally, you’ll need to come up with a method of campaign measurement to provide the metrics necessary to fine-tune campaigns as you roll them out. When creating the measurement methodology, keep these considerations in mind:
-
Be sure the test cells are set up so variables are isolated. If you don’t, you won’t be able to draw any actionable conclusions from the tests. For example, if you had one test cell that contained an offer for a 10% price reduction with the product painted red, and a second that offered a 20% price reduction with the product painted blue, you wouldn’t be able to make anything of the test. Which combination of discount and color would work best in a rollout?
If the first package produced better response, was it because of the 10% discount or the red paint? There would be no way to tell. In order to measure those variables, you would need four test cells, with each possible combination represented.
-
Make absolutely certain each test cell will produce a statistically projectable number of responses. There’s an old rule of thumb in the direct mail business that 30 responses is enough to project onto any size mailing universe. I know people who always insist on at least 100, just to be on the safe side. In fact, there are tables (available in just about any textbook on direct marketing) that will show you how many responses are required, based on the size of the promotion universe, to trust the results at various confidence levels.
One of the wisest and most pragmatic direct marketers I’ve met once pointed out that there are two measures of statistical validity — one that provides a statistically projectable number, and a second, much larger number that the CEO will believe. Suffice it to say that you should err on the side of caution in this regard.
-
If some or all of your campaigns are ongoing, you’ll need to set up a measurement method that takes that fact into account, since the program can’t be measured one mailing or call at a time — it’s the cumulative effect of multiple touches that should be measured.
-
Consider setting up a control group of customers who are permanently excluded from any type of database marketing communications. The idea is to give you a baseline measurement of the behavior of people who are not exposed to database efforts, to compare against the rest of the universe. The control group need not be large, usually somewhere between 2% of the file and a maximum of 5%.
This will help establish what effect your database program is having in the aggregate. Given the small sample size, measurements should be kept relatively gross — metrics such as average yearly revenue per customer, monthly transactions, etc. Over time, you’ll see significant differences that will give you a provable measure of the value of your overall approach to database marketing. This information can be helpful for justifying budgets, or for explaining the value of your program to a new CEO.
With decisions made about how you’re going to segment your customers, you’re ready to begin using your new marketing database for campaigns.
After some initial testing and experimentation to learn what works, many marketers plan a series of continuing communications that span a specific time frame, an approach known as a communication matrix.
There are several advantages to the matrix:
-
To help build the customer relationship and support the brand experience. A series of “soft” communications designed to reinforce the reasons customers bought from you can convince them you really care about their needs and want to earn their business over the long term. This is where you build the foundation for relationship management, and it can pay off in customer satisfaction and retention.
-
To make sales along a customer needs continuum. In most situations, there’s a series of key leverage points in the customer relationship — the times and situations when the customer is most likely to want to buy again, buy up, or renew — that can be identified using fairly standard marketing database cross-selling and activity reports. Selling into those leverage points can create sales you would not always get through traditional product-based marketing campaigns.
-
To automate customer cross-selling and retention efforts. After enough time has passed to learn what works through experience, it often is possible to automate matrix communications. For example, a new customer might get a combination welcome kit and thank-you immediately after buying, followed a couple of weeks later by an offer for an additional product or service linked to his first purchase. Then, a few weeks later, a third mailing could ask him to judge the quality of service he received from your organization. A couple of months later, another offer may be sent for another appropriate product.
At any time, this pre-planned series of sales and relationship-building efforts can be interrupted for more broad-based time-sensitive campaigns, such as seasonal sales.
Response Management
Once the marketing initiatives have been launched, responses will start coming in, and those responses will require handling.
Depending on your type of business, response handling might include product or information fulfillment, assignment of the response as a lead to someone in field sales for follow-up, scheduling of a presentation, and/or notifying the responder that you received the request and are working to fulfill it.
This is an important part of the database marketing process that many marketers simply ignore, reasoning that marketing’s job is to generate the lead and someone else has the responsibility to close the sale. This is shortsighted, since a database marketing effort is always measured by revenue generated, and revenue doesn’t come from inquiries. At the very least, you should sit down with the sales managers in your organization to determine how leads are handled and what marketing can do to facilitate that process.
An inquiry is a perishable commodity. Above all, keep in mind that there always is an inverse relationship between the value of an inquiry and the length of time it takes to handle it. Every day that passes without a reaction from your organization diminishes the chance of closing the sale.
Plot the response-handling process on a flowchart and try to establish measurements at every touch point. This will give you the means to determine where sales are being lost, or what is causing delays in efficient and rapid fulfillment of customer requests.
If you sell through a field sales organization of some kind you’ll need to consider the processes used to notify them about the sales opportunity, and insist on some kind of feedback mechanism so you can tell what they did with that opportunity. What you’re going to find is that some salespeople are very good at turning leads into sales and some aren’t. Unfortunately, any attempt to give leads only to the closers is usually a political hot potato. But if you can classify the leads according to their value, you might be able to see that the best leads are directed to those individuals who can make the most of them.
Lastly, you’ll want to “close the loop” on your database marketing program by using modeling and response analytics to determine what’s working, what needs to be improved, which customer and prospect segments are most responsive, and how you can fine-tune your program to make it more efficient and profitable.
To access the complete series of articles and charts, visit Direct’s Web site (www.directmag.com).
Richard N. Tooker is vice president and solutions architect at KnowledgeBase Marketing Inc. in Richardson, TX.
Network
