• by Paul Knegten , June 14, 2010

In a previous piece, I lauded the efficiency of real-time bidding in some concrete examples of how travel advertisers can use it to micro-segment their audience. The efficiency comes from allowing for bidding differently for different audiences, allowing you to pay the "real" value of each audience (instead of bidding the same for a business traveler in first class and a family researching a trip in six months).

This approach is already far beyond the scope of most advertisers' marketing strategy. But I'm always here to tell you about where the envelope is being pushed next and, thus, I introduce what I like to call variable profitability in real-time bidding: the next real boon in advertising efficiency. And luckily for travel advertisers, having lots of perishable products or offers means they have the most to gain from this concept.

It's not as complex as I make it sound. Micro-segmentation is simple enough -- now what if we took a closer look at each segment as a function of time? A business traveler's propensity to convert isn't a constant: when he or she is researching a trip, there are points in time when the impact of your advertising is higher, and those when it is lower. So the most efficient bid price for said business traveler becomes a function that is able to be modeled by any good demand-side platform that manages bids on your behalf.

The idea of propensity to convert becoming a moving target is just what real-time bidding was built to do. With each impression, you can calculate the expected revenue (probability of a conversion multiplied by the average revenue from a member of an audience segment) and bid accordingly. Now for the next big advance: using real-time bidding to pit expected revenue against expected cost.

Most travel advertisers give consumers many potential options of routes, room types, locations, and frills that carry a certain value to each consumer (affecting propensity to convert), bring in a certain value to the advertiser, and cost them different amounts to deliver. This last bit of datum is key to maintaining the highest possible profitability with your ad campaign. For all the buzz about third-party data providers in advertising today, not nearly enough is being said about the huge benefit of this first-party cost datum.

Optimally, you'll have a feed of some kind (from your bean-counter types) that tells you at any given time the exact cost for each product you offer. As you bid for impressions that show ads that display messaging for one particular airline routing over another, a calculation can be made on the fly that determines the cost of a successful impression (one that causes a conversion) and that cost is weighed against the expected revenue to more finely tune the bid price.

And there you have it: both sides of the equation being taken into account when you bid for impressions. We've come a long way from blind network buys, haven't we?