Real-Time Bidding: The Next Round
Aaron Levie @levie
If you a miss a cycle in tech, it's basically over. If you hit a cycle, you've only earned admission to the next round.
If the whole Box.com thing doesn’t work out, Aaron Levie definitely has a career in startup fortune cookie writing. Seriously though, that was one of those Tweets that just smacked me upside my head and, given its retweet and “favorite” stats, I wasn’t alone.
As it pertains to display advertising, owning RTB infrastructure is quickly emerging as the ticket to the “next round.” This means having the bidders, cookie stores, and big-data chops to earn a seat on an exchange and respond to bid requests within 50 ms, without having to use a middleman or license another DSP’s technology. The launch of Facebook Exchange particularly highlighted the distinction between owning and just licensing RTB. There are folks with the tech that have a seat on the exchange (listed here) and there are those who don’t.
To those who are truly able to fulfill the above RTB requirements, congrats and welcome to the “next round.” However, as Aaron Levie’s tweet conveys, there’s no time to pat yourself on the back in the fast-moving tech sector. The first round was really just about building the underlying infrastructure. As an analogy, take buying a new camera--merely owning a camera doesn’t make you a great photographer. The next round is all about options: where to point your camera, how to focus it, and how to adjust your aperture and shutter speed to take fascinating pictures.
The algorithm that powers said camera is one clear opportunity for customization, versus the one-size-fits-all options that are commonly offered today. There are some underlying truths that any RTB algorithm should incorporate: pacing, performance-based learning, cadence modification, viewability, etc. However, as each advertisers’ objectives are slightly different, the key is to create both a flexible algorithm and platform that are capable of accommodating these various requirements.
For example, it’s popular to talk about customer quality scoring today. Unfortunately, there is no universal customer quality scoring methodology for every advertiser. If an advertiser is interested in clicks, people who are “clicky,” showing a history of clicking on ads, should get a higher score. If an advertiser is interested in measuring incremental conversions, “clicky” people might not get a strong weighting if they’re likely to convert regardless of the ad impressions they’re seeing.
Dynamic creative is another area where owning the RTB infrastructure is important. It’s only natural that how much you bid on an impression depends on what kind of ad you can show. For instance, it’s smart to bid higher on an impression if you can match a user intent signal with a strong personalized product recommendation or offer. Unfortunately, creative is often overlooked in the programmatic advertising world, but connecting bidding and ad presentation often accounts for more than half of the performance of campaign. if you don’t do this, you risk running lobotomized campaigns.
The winners of the “next round” will be those that can accommodate all of these different objectives: the ability to shape the algorithm to weight different variables more or less depending on the campaign, the ability to customize audiences and targeting criteria, and the ability to tie bidding strategy to dynamic creative. Without the infrastructure investment in round one, there’s no way to jump to round two.