In a previous article we defined what behavioral targeting is and what it is not. It is not the act of monitoring consumer interactions with a given ad unit. That is called optimization, and it is a rudimentary practice of nearly all ad servers. Instead, behavioral targeting is the aggregation of observed consumer behaviors, the segmentation of the audience using a rigid set of definitions (or taxonomy), and the subsequent delivery of an ad based on that segmentation.
The act of observing consumer behavior is, in itself, a simple matter of straightforward analytics, and is as automatic as counting clicks. Serving ads against a pre-set group of rules is a straightforward practice, too. Making the leap from observation to execution, and using a rigid taxonomy to determine when a consumer is most likely to make a purchase decision, however, is anything but mechanical.
The first step in behavioral targeting is to create an appropriate taxonomy. What special traits define your audience target? What specific behaviors are you going to track? Taxonomies can generally be divided into four key components: content, commerce, identifiers, and advertising. Let's consider the mythical Acme Corporation, an online marketer seeking to sell high-end fishing equipment to adventurous, well-moneyed, well-traveled executives and retirees across the United States. What behavioral points might Acme look to include in its taxonomy?
Content: When potential customers visit your site... or for that matter, any site that can be monitored... what content are they consuming? Travel content? Investment strategies? Issues on taxing retirement funds? Material on fishing expeditions in tropical locations? The opportunities are endless. As they navigate monitored sites, information on what content visitors consume can be recorded and measured against the preset taxonomy.
Commerce: What do potential customers purchase online? What is the average price of the products that they buy? Do they buy fishing equipment or travel guides? Expensive airline tickets? How often do they make higher-priced purchases?
Identifiers: Does a potential consumer live in a high-end ZIP code? Has he or she self-selected to receive information on travel, fishing, and adventure vacations? Have they listed their family's net income?
Advertising: Is there a particular form of advertising that this potential customer typically responds to? Pop-ups, perhaps, or banners? Has a particular frequency of ad delivery triggered clicks and purchases by this potential customer before? Is there a particular day, time, or online property at which this potential customer is more likely to respond to advertising?
A well-conceived taxonomy can help answer all of these questions. At 24/7 Real Media we monitor a robust taxonomy of some 600 consumer behaviors across our Web Alliance, a network of more than 800 Web properties with more than 45 million monthly unique users, and growing. At that scale, up to 21 trillion behavioral data points can potentially be monitored, more than enough points to produce a solid perspective on consumer behavior, gauge purchase likelihood, and serve the appropriate ads.
The ability to serve the right ads at the right times, to people who will respond to them and buy the advertised product or service, is, of course, of critical importance to the advertiser. Advertisers, in the end, are whom this is all about. Show your advertisers a technique that will allow them to fine-tune their ad placement so that their ads go only to those who are interested in their product, and they will love you forever for the value you deliver.