External Data Will Help To Explain Customer Behavior

Let’s continue a bit with last week’s post about capturing all the marketing messages received by customers and prospects. Although this is becoming easier and less expensive, it will never be completely simple or free. So there will always be a need to measure the cost of gathering the information against its value.

The value will ultimately be two-fold. Retrospectively, the information will help to measure the impact of past messages, enabling optimal allocation of marketing investments. This works at the aggregate level, and is traditional marketing measurement.

But we also want to use information proactively, to help target individuals. Consumption of marketing messages, particularly in online media, is closely related to customer choices about which Web pages to visit or which offers to accept. This behavior provides vital insights into the customer’s current state of mind and, thus, which future treatments are most likely to be productive. Here we’re moving into a different type of marketing measurement, which involves predictions of individual behavior.

I suppose we could debate whether this type of activity, which itself is not new, really should be labeled as marketing measurement. But the particular point I had in mind today was even complete information about a customer’s interactions with the company will not be enough to accurately predict behavior. There are many external factors as well. So individual-level predictions will be much more useful if they can be based on a combination of internal and external data.

This isn’t a particularly brilliant insight, or a new one. But what is new is the greater availability of external information, such as data about individuals’ backgrounds, news about their companies, and public attitudes revealed through online discussions such as forums, blogs, and reviews. I haven’t compiled a comprehensive list of the information sources, but companies like Jigsaw (online directory of individuals and companies), InsideView (aggregation of news about companies) and Twing (tracking of online discussions) keep popping up. Any predictive system would benefit from incorporating their contents. Benefits include gaining background on customers and prospects (which should yield insights into the approaches most likely to be effective), and identifying specific events (which might prompt new needs for particular products).

The most important value from these sources will come from better sales results. But retrospective marketing measurement will also benefit because it will more precisely identify the factors that impact the results of a given marketing effort. Thus, the focus will increasingly shift from “whether” a particular marketing program worked, to “when” (i.e., under which conditions) it works. This will guide both treatments of specific individuals and the over-all allocation of marketing resources. Once that happens, the distinction between retrospective and proactive marketing measurement will be less important: they will effectively be the same thing.