Better Behavioral Marketing
Yet in this evolving field, there are no clear understandings about what inferences can properly or accurately be drawn from demonstrated behavior. Other than the individual who clicks through to a completed sale, most behavioral targeting at this point is guess work.
Consider several approaches for observing, evaluating or scoring behavior.
Logically someone who does the same thing again and again or visits the same place repeatedly is probably more interested than the average Joe. Assuming that most people only make one or two clicks in error, it is reasonable to guess that someone returning for a 3rd click is probably interested, if not a real buyer.
My wife is a prime example. She likes to visit future purchases frequently before buying. To her, multiple trips to the shoe store, the furniture store or the big box retailer to hover over her intended item are no big deal. In fact she enjoys the process of visiting and revisiting. With each incremental step she learns more, increases her desire and adds layers of nuance to her buying rationale.
Repetition confirms her interest or ratchets up her intent and her commitment to the purchase. The same holds true on the Web. She will click and click again on an item. Perhaps she’ll visit it at multiple sites in search of greater product detail, to compare prices or to discover a deal on shipping or a favorable return policy.
If we tracked her behavior with a cookie or some other technology, the vital questions would be -- how many visits signal her intent and on which visit should we prompt her to buy? Should we dynamically serve her content or intervene by popping up an offer on her third or fourth visit or on her sixth?? How do we know how much repetition is sufficient to encourage her to convert or at what point she might be freaked out by a big brother intervention and abandon interest?
Perhaps if we watched where she went before and after visiting the product, we might get a better idea. If she visits the same product at a competitor site, does that signal intensity of interest or intent? If she looks at a similar product or a product that normally goes together and sells together with the first product can we infer a pending purchase?
If she puts the cherished item in the shopping cart and abandons it can we assume “No” is not really no? And it is a fair expectation that if you abandon a shopping cart, someone may follow-up with a question or an offer?
What about if she makes a beeline for something? Would that be enough evidence to treat her differently from the great mass of web surfers? Say she responded to an email and clicked on a designated landing page or navigated from the home page to a particular product page in the shortest possible sequence (3 clicks?), would that mark her as an A prospect and separate her from the herd?
How about if she fills in a form, signs up for an e-mail newsletter, downloads a whitepaper, prints out a PDF, uses a zoom feature, puts data into a calculator or clicks a “contact me” button? Assuming that only X percent respond frivolously, would using the provided response device qualify her as a hot prospect or merely mark her as either a
well trained consumer or as a tire kicker?
If she fills in only part of the form, what can we infer? Is a newsletter subscriber more interested than a downloader? Can we distinguish between serial subscribers and sequential downloaders, who could be anyone from your next best customer to a high school kid working on a project?
Direct marketers will tell you that even among those who answer the call to action and utilize the provided response mechanisms; most responders are generally interested but not ready-to-buy. So the act of responding, while rarely more than 2 percent of those exposed to an offer, still doesn’t turn you into a qualified, hot lead or give us any indication that for a little extra effort or TLC we can get you to buy.
So what’s a marketer to do?
Charged with generating demand, we want to do so in the fastest most cost effective manner. Accepting the notion that actions speak louder than words and that people are creatures of habit who often act in repeatable patterns, we buy into the notion of behavioral marketing. But how do we draw the right inferences from the behavior we observe?
The short answer is threefold. First we test and learn. It’s classic but it is limited because we cannot project our learning across product sets or categories. It’s even better if we can share data with each other, make generalizations and hypotheses based on products, categories, dayparts, gender and other psychodemographic factors and get collectively smarter at reading the digital tea leaves
Second, we watch behavior over time. Assume that somebody who does the same thing or related things over time is more interested that the person that does it once and moves on. As a corollary, assume that someone who accesses or responds in multiple ways or at multiple, different times is more interested and has a higher purchase intent that a person using only one media channel. If we collect data from multiple channels (web, e-mail, search words, trade shows, purchase history, coupon redemption, etc) we can begin to see patterns that will suggest how we can weight and model observed behavior.
Third, we aggregate the varying dimensions, try to weight them by watching what prospects do over time and then attempt to triangulate purchase intent and intensity. This applies particularly to high value, considered consumer items (cars, stocks, diamonds, real estate) and in B2B marketing where the shopping cycle is longer and where the decision set has a larger number of variables with complex relationships between them.
If behavioral targeting is going to become a standard, it has to have impact. Behavioral targeting must help us sell more things faster to those most likely to buy. If it doesn’t, it is just digital voyeurism.