Why PPC Behavioral Targeting Doesn’t Matter

So the big buzz this week is that Google opened up behavioral targeting to all advertisers, effectively closing the beta test they’ve been running for the last two years. I’m sure many advertisers are celebrating their new ability to target display network ads by user interest categories. But I remain skeptical that this is going to be a valuable asset to your average PPC account holder. Instead, this just seems like another Google PPC doodad that doesn’t quite live up to the hype. Here’s why.

Let’s start with an experiment. Did you know that you can view the interest categories that Google has decided you’re into? You can. Point your browser to this link: http://www.google.com/ads/preferences/view. Now look at the interest categories that Google has laid out for you (assuming that you have any). Do these match your true interest categories? Or are they just representative of the last couple of browser tabs you have open? Chances are it’s the latter. User sentiment is a tricky thing to pin down. Just because you did a one-time search for some shoes doesn’t mean that you’re going to want to be bombarded with shoe ads for the next 24 hours. Google seems to think that I’m very interested in music streaming since I have a Pandora tab open. Sure, I’m interested in that tab that I have open, but I’m not going to buy anything from another music advertiser. My information need (in this case, Pandora) has already been fulfilled by the time Google figured out what it was.

Then you have the problem of shared and work computers. If you share a computer with a spouse, then your interest categories are going to get all mixed up. If a wife is shopping for a new purse, the husband might get ads for Coach on his next browsing session. It’s a total interest mismatch. The work computer problem is even more confounding. Like it or not, most people do the majority of their internet browsing during the work hours. Run an ad preference analysis for your work computer, and compare it to that of your home computer. It’s likely you’ll see quite a difference. At work, I get a lot of ads for search engine marketing, but I can tell you that AdWords is probably wasting their money advertising to me at this point. The same thing could happen to a small business.

So interest categories are a good idea in theory, but the technology just doesn’t exist for it to work in practice yet. My advice is to stick with the more reliable segmentation filters for now – geotargeting, device, and time scheduling all work within a reasonable degree of accuracy. Before you start trying the latest new gimmick, make sure you’ve exhausted all your options to improve your account through tried and true ways.

About Shawn Livengood

Shawn Livengood is a search engine marketing professional based in Austin, Texas. He has extensive experience managing pay-per-click ad campaigns for clients in various industries, from small home-based businesses to large international companies. You can connect with Shawn on Google+.
This entry was posted in Content Network, Google AdWords, Search Engines and tagged , . Bookmark the permalink.

One Response to Why PPC Behavioral Targeting Doesn’t Matter

  1. Good points. The problems you mention could be solved by simply adding some Bayesian weighing factor to each ‘tag,’ rather than Google putting you in or out of a category (maybe they do this already?).

    So your Pandora habit doesn’t say you’re an automatic music-buyer, but it does increase the odds relative to a Pandora non-user. Maybe it’s only a 1% increase, but it is a statistical association.

    Same thing with the purse; the correlation is sound for the proportion of time she’s actually using your computer. If she’s on 10% of the time, again, your computer is still associated with purse buys, simply at a 10% magnitude of association.

    So it’s not a matter of “technology doesn’t exist,” it’s simply a matter of incremental statistics rather than hard categories.