Upuzzling Matchbacks

The marketing manager was suspicious. The pay-per-click Web campaign results looked too good. A matchback revealed that 40 percent of the campaign’s customers, representing 60 percent of its sales, had actually received a catalog before placing their orders. Scary, isn’t it? That’s just one reason why order tracking still matters.

Here’s another: The chart accompanying this article is a real — and typical — example of key code capture rates. This unnamed cataloger captured key codes for 46 percent of its orders that represented 62 percent of its sales. Untracked data represented 54 percent of its orders and 38 percent of its sales. The vast majority of the untracked orders came from the company’s Web site.

A matchback identified the key code for 35 percent more of these orders and 31 percent of the sales. This process increased matched orders to 81 percent and matched sales to 92 percent.

The matchback isn’t perfect, but which set of data would you rather use to plan your next campaign?

The Matchback System
The matchback process was once shrouded in mystery. You needed a mainframe with a high-end programmer. Today, you need a working knowledge of Microsoft Access and a laptop. If you lack either, you can outsource the service to any number of companies. In its simplest form, a matchback mirrors a dedupe process. You simply append the key codes from your mail-file records to their match key dupes on your order file.

If your customers have a multilayer order process, the matchback gets more complicated and requires a few more steps. You may mail your catalog to an engineer in one facility, for example. Then, a purchasing department in a different city places the order, and the order is shipped to the manufacturing plant in yet another town. Tracking that source key code makes Rubik’s Cube seem easy. But like solving the famous puzzle, your matchback just requires a system.

First, don’t assume that every order follows this complicated process. Run your initial matchback to glean the low-hanging fruit. Combined with your tracked orders, you may now have enough data to begin making decisions. If not, gradually loosen the criteria, first to the site level and then to the company level.

To verify your results, request that the mail records and their matched order records be provided on the same row in your returned matchback file. Now you can double-check the matchback results. You won’t get a 100-percent match rate, but your data, and decisions based on it, will be better.

Here’s a final example: For one mailing, a matchback increased all prospect list responses by more than 20 percent. One list’s response increased by 50 percent. For several of the prospect files, that was the difference between success and failure. That’s why order tracking still matters.

George Hague is vice president of J. Schmid & Assoc. You can reach him at GeorgeH@JSchmid.com or at (913) 236-8988.