Building an Integrated Marketing Plan

The three stages of an integrated plan
It isn't reasonable to expect a multichannel cataloger to go from traditional segmentation to a fully integrated contact strategy overnight. You might want to ease into an integrated marketing plan though a three-stage approach: simple segmentation, advanced segmentation, and marketing mix modeling. (Note: The marketing plan we're discussing below applies to customers, not prospects.)

Stage one: simple segmentation This assumes that you can add a simple level of segmentation — you do or don't have an e-mail address on file for the customer — to your existing methodology. The same segmentation that's always (presumably) worked stays intact, with one added wrinkle. Now instead of launching e-mail campaigns to tens or hundreds of thousands of customers under one single campaign code or a handful of promo codes, you can use the same segmentation practices for your e-mails as you do for your print mailings, and you can measure response down to the segment level.

When the season is complete, you don't measure response and profitability solely for the e-mail campaign or the catalog campaign but rather for the entire communication string — all e-mails, catalogs, postcards, and whatnot that contributed to those orders. Now it becomes less important whether X e-mail or Y catalog “drove” the sale; the focus is put on two more-important questions: How much was spent to get each order, and how profitable was each order when it came in? All associated costs for a segment are considered against all associated sales. No longer are e-mail campaigns measured separately against catalog mailings and decisions made, perhaps incorrectly, that “the catalog is no longer necessary, because e-mail is working so well.” Instead a truer picture of overall performance is evaluated based on the complete set of contacts to a customer group.

Stage two: advanced segmentation Applying a more advanced segmentation model to your multichannel business means moving from traditional RFM, where recency of purchase, frequency of purchases, and monetary spending levels dictate the categorization of customers, to a schema where RFM is enhanced, particularly by channel. While RFM variables alone may determine how deep and how frequently a file is mailed, CRFM (“C” representing purchase channel) looks first at channel to determine the sequence and type of communications (catalogs, e-mails, postcards, etc.).

There are two questions to address for each customer regarding purchase channel: through which channel did he make his first purchase, and through which channel did he make his most recent purchase? Then ask if the first purchase was driven by a mailed catalog or an online marketing effort. If you know that the customer was mailed a catalog as a prospect but responded through the Web, it's apparent that even though the customer is an online buyer, he still responds favorably to catalog mailings. If the same customer is a multibuyer and has always responded through the Web, you might be tempted to say, “Never mail him again.” This could be a mistake avoided by evaluating what drove that first channel purchase.

If a customer is a catalog responder or has responded to a catalog mailing by phone or mail at any time during his life on your file, the mailings are playing a role in his responsiveness, so he should be put on a communication plan that includes both e-mail and catalog mailings. Customers acquired through online marketing efforts who have only ever responded by way of the Web may very well be candidates for far fewer catalog mailings (note “far fewer,” not “none”) without a significant drop in performance.

Stage three: marketing mix modeling Marketing mix modeling builds on the concepts of media mix modeling applied by advertisers to determine the most effective frequency and flighting of space, television, radio, mailings, and other media. By factoring for when communications are put into the market and how response ebbs and flows accordingly, you can take a more mathematical approach to projecting response.

A multivariate approach, marketing mix modeling takes into account the typical RFM variables along with first and most recent purchase channel, purchase frequency by channel, e-mail “type” by response (for instance, if a customer responds to promotional e-mail but doesn't buy online otherwise), monetary by channel (cumulative and average), and more. Working with a qualified modeling company or a database co-op to determine the right channel mix can add a significant level of sophistication to any direct marketer's contact strategy.

Next page: Test and measure