But I've covered that in the past on MediaPost. Today, I want to talk about a completely different reason to be sub-segmenting: like your audiences, your products have completely different values and propensities to cause a conversion. The difference seems subtle but it isn't; think of your audiences as walking expressions of user intent, and think of your product catalog as the answer to each of those intents. For each traveler looking to fly from New York to L.A., there are any number of flight combinations, hotel options, and package deals that a particular advertiser might want to show. Think of each of these as products you offer to your customers.
To effectively manage your display campaigns on a product level, each ad impression must undergo a decision-making process that chooses the right product by weighing three factors: the degree of match to the consumer's intent, the propensity for that match to cause a conversion, and the revenue or profit potential of a successful conversion.
Here's an example: in a recent campaign with a large hotel advertiser, we analyzed the performance of each city shown in the ad. Each city was chosen initially as a result of a clear intent driver (either the user had previously searched for that city on the advertiser's website or the user was reading something contextually relevant to that city), but the specific property shown was chosen based on its higher propensity to cause a conversion. A high or low profit margin on that property could carry more weight in the decision process depending on the preferences of the advertiser, but we left that factor out of this particular study.
And there's more -- your products also inform the best media buying decisions if you're bidding in real time. In our study, some cities clearly performed better than others (Myrtle Beach was a hot item last month). The bid price becomes another impression-level decision, with higher-performing cites such as Myrtle Beach garnering a higher bid to win more opportunities to show your ad to a high-performing audience, and lower-performing cities (I'll spare those cities the embarrassment of being called out in this column) getting a low bid.
It might sound daunting in the context of display advertising to talk about such granular management of your audience and your specific products (of which you may have millions if you're a travel advertiser). But this approach might ring a bell with some marketers that do exactly that, in a different medium.
Yes, this is essentially rehashing the lessons we've learned from search engine marketing and applying them to display advertising; I've come up with essentially nothing new here. But until now, technology to manage display this way didn't exist; now it does and the marketers best positioned to take advantage of it are the search marketers. So go do it!
Editor's note: Paul will be moderating a panel, "Grill the Vendors: Real Time Bidding," on July 19 at MediaPost's OMMA AdNets conference in Los Angeles. Catch him there!
Also, Gary Leopold's regular monthly contribution to this newsletter wound up in Marketing:Green last week. If you missed it, you can find it here and you'll see why it was posted in that newsletter!