OK, class, quick test: You mail a catalog, and 21 days later send a postcard. If the customers buys something on day 28, what fraction of that sale was postcard-driven?
Consider this example: A mailer sends out a catalog on the first day of a campaign, and 21 days later sends out a postcard to many of the same customers. If an order is placed on Day 28, what fraction of that order did the postcard drive? And what about that e-mail that went out on day 12?
You can see how murky this is when you have no keycode attached to the sale and rely on the matchback to tell you the order was driven by those mail pieces. And it’s not much better even if you do.
But it proves on point.
As the media mix diversifies, the need to match sales data to promotional files is no longer up for debate. Direct marketing merchants simply must do it. But necessity and ubiquity won’t clear the minefield of business intelligence.
Chief among the potential pitfalls, ironically, is the misinterpretation of results due to questionable business rules driving the matchback process.
To tease the data apart and make it actionable, you must set business rules that all stakeholders will buy into. Given that we’re talking about interpreting customer behavior, this is inherently paradoxical, but I do not believe there is a responsible way around it.
One set of matchback business rules won’t fit all, and a matchback can’t provide all the information required to do the best job in today’s direct-to-customer business environment. So approach the development of solid business intelligence as an iterative and collaborative process.
Jude Hoffner is director of marketing and database services at San Rafael, CA-based consultancy Lenser.