Writing about marketing mix models yesterday got me thinking about why this technique, which started in the consumer packaged goods industry, is now being applied elsewhere. My conclusion—which I found intriguing, and think you may as well—is that even industries in direct contact with their customers are starting to resemble the anonymous world of consumer packaged goods.
Bear with me on this. Most thought relating to customer management has come from industries where companies can link transactions to specific customers: financial services, telecommunications, direct response, travel, retail, gaming). The goal has been to gather this customer information and use it to tailor customer treatments—or customer experiences if you prefer—to each individual. Increasing sophistication has meant gathering more data to understand customers in more detail to make more precise predictions of larger numbers of behaviors. The industry’s proudest boast has been that it can measure the results of marketing activities with a precision that the customer-blind (but vastly better funded) practitioners of traditional consumer marketing cannot.
But a funny thing happened on the way to the future.
Actually, it was two funny things.
The first was simply that more and more targeted messages were sent. This meant the foundational myth—that you could attribute a specific customer response to a single marketing message—became increasingly untenable. Marketers always knew it wasn’t true, but when the number of contacts was limited to a one or two a month, the inaccuracy of accepting it was minor. But now that customers may receive dozens of messages each month—from email, Web site visits, and operational events like invoices, packages and phone calls—it became increasingly absurd to argue that the only message worth counting is the last one before the order.
The second thing that happened was customer-specific marketing moved from being a supplemental tool to the primary mode of contact. Organizationally, this meant the marketers responsible for targeted messages were now also responsible for conventional, mass media. Now they had to find a way to measure the impact of those mass media, both so they could continue to justify buying it (because, deep down, they knew it was providing some value) and so they could include it when they analyze return on all marketing investments.
Both of these factors lead in the same direction. Marketers need to allocate credit for customer behaviors across multiple inputs: the simple method of attributing a single response to a single promotion just won’t do. This is specifically what marketing mix modeling does.
Marketing mix models also incorporate external influences, such as competitive behavior and environmental conditions. Together with internal factors, these should explain pretty much everything about purchase behavior. This type of zero-based forecasting is vastly more demanding than simply measuring the incremental lift from a particular marketing campaign or the relative performance of test vs. control. But it’s also part of taking responsibility for business results: you can still blame the economy or competitors, but you have to prove they were really the cause of a shortfall. (It’s only shortfalls that need explaining. Over-performance is always due to brilliant marketing.)
In other words, becoming more like packaged goods marketers is not a bad thing. Accepting responsibility and dealing with ambiguity are signs of maturity and importance to the larger business. From the perspective of customer experience management, attempting to define all factors that impact sales forces top managers to recognize the importance of operational issues and to measure their impact as well. This, of course, is essential to a full adoption of customer experience management principles.
If all marketing is really becoming like packaged goods marketing, this may have other implications worth noting. I’ll let you know if I think of any. Or, better still, let me know what YOU think.
Thursday, January 04, 2007
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment