A Tale of 10 Pixels

In fast-moving innovative industries, it can be quite a chore to be a savvy consumer. Online display advertising has been around for a while of course, yet new innovations like behavioral targeting and supply exchanges are keeping media planners on their toes. More and more these days, the confusing array of technological and methodological options afforded by ad networks is turning media buyers and planners into marketing scientists — figuring out what sorts of experiments and measurements they can make to understand which ad networks will perform the best for their clients.

Testing Isn’t Easy

Recently the agency for a large consumer products company ran a test of over 10 ad networks concurrently, primarily behavioral targeting networks. On the face of it, this sounds like a great way to see how well they perform, so that the agency can pick the best networks for the real campaign. However, the laws of physics of online marketing make this test more like choosing the best sprinter out of 10 runners by putting them all in one lane on a track and firing the starting gun. They’d stumble all over each other and it’s unlikely that the first one to cross the finish line is the fastest runner. (It’s likely to be the runner who was best at throwing a few elbows to thin the pack!)

Similarly when running multiple ad networks concurrently, either as a test or part of a large campaign, it’s important to understand how their targeting strategies will interact. Multiple behavioral targeting networks will likely be using their own techniques for drawing a relevant audience to the Web site, but they’ll also be doing retargeting, and therein lies the rub.

Say ad network A has built the world’s greatest behavioral targeting engine in the world, and uses this technology to drive new users to the company’s Web site. Ad network B has no technology but has its pixels on the company’s Web site and is retargeting the company’s Web site visitors on various exchanges. If the product being marketed has a consideration cycle of a few days, the typical consumer will be drawn to the Web site by network A, then will think about it for a few days. During those days of thought, they’ll see ads from network B. Eventually when they come back and convert, network B is likely to get the credit. Now imagine if there was not just network B but also networks C, D, E, F, G, H, I, and J all with pixels on the homepage, retargeting the same users. It would be a madhouse, almost impossible to tell which networks are good at actually driving conversions, versus which are good at gaming the system to get credit for conversions.

It’s possible to know which networks are driving new users to the company’s web site versus which are only retargeting, but it involves analyzing Web server and/or ad server logs in a very granular fashion that only a few ad servers can do.

Caveat Emptor: When Pixels Go Bad

A media planner for a major automotive brand was recently surprised to discover that a new ad network was delivering results as good as the existing retargeting partners, and took this to mean that the ad network’s technology was living up to its promise. Or were they actually just doing retargeting?

One thing a media buyer can do, to understand what ad networks are actually doing, is to look at what their pixels are doing. Using Firefox and the free add-in Firebug, the network traffic of any Web page can be easily viewed. If an ad network pixel is loaded with a “302 Moved” status, it means that the pixel is redirecting to one of the exchanges, so that the ad network can bid on the user and retarget them.

Don’t Forget About References

Isaac Newton said, “If I have seen farther, it is because I stood upon the shoulders of giants.” What he meant by that is that instead of starting over from first principles when doing his own research on mathematics and physics, he began his work on the frontier of knowledge already established by other scientists. It’s not like media directors and planners all over the world are publishing treatises about how ad network have performed, but a savvy media buyer can definitely ask an ad network to arrange a call with an existing happy customer or two.

Aside from understanding whether ad networks have all the rocket science they claim to have, it’s also key to understand mundane things like how flexible they have been in the past to changing goals, handling incremental spend, adjusting dates, proactively pointing out strong/weak creatives, and other value adds.

Related reading