Best Behaviour When It Comes To Behavioural Targeting

The objective of online publishers to maximise the advertising return on every page impression delivered is analogous to an airline’s fixation with maximising passenger loads for every flight - better to sell a seat at $1 than to fly with an empty seat. Yet the airline industry’s sophistication in balancing yields with demand for every slot is but a stunning utopia for most publishers who must be satisfied with impression forecasts based on content or application type.

In other words, a majority of display sales are based on context and the assumptions of audience grades (i.e. AB, C, D socio-economics), and not on specifics like profile data and current ‘clickstream’ behaviour. Yet the question isn’t necessarily about how quickly behavioural targeting (BT) technology is adopted - the appetite for investing in BT and inventory management systems is already very high in Australia - the question is more of how publishers manage the technology and the associated data to optimise the returns.

In other words, what are the ‘control room’ resources being dedicated (or planned for) to complement the technology investment? Who builds and adjusts the business rules which define the target groups? Who monitors impression rates generated by the target groups and culls ill-defined or poor performing audience segments?

The best BT systems will report in near real-time, allowing operators to re-consider target group definitions on-the-fly, bringing a whole new dynamic to inventory management. Changing referring URLs or search terms, the definition of URL channels, the type of profile data used, even the particular links or display ads clicked on, will all have a bearing on the relevancy between the advertising message delivered and the defined audience. And all these combinations will be further impacted by day-of-week and, more specifically, time-of-day.

In short, publishers committed to a BT agenda must be prepared to define and implement quality scores to firstly define their existing premium audiences (as a baseline), and secondly, to ascertain which non-premium areas of their sites achieve similar quality scores. By finding matches between the scores and previously known run-of-site inventory, the publisher is in affect broadening the base of their premim inventory.

Increasingly, the competitive edge provided by BT isn’t necessarily the chosen software, but the ’smarts’ employed to maximise the amount of premium inventory available for sale at any given time of day and to keep those premium inventories rising ever higher a percentage of total page impressions.

Ultimately, during prime-time (say between 12 and 2pm on a weekday), best practice would be achieve premium rates across 90%+ of page impressions delivered in that period. Now that figure would make even Geoff Dixon smile.