As I logged on to my web-based email account recently (via a portal I won't name, to protect the innocent), I was once again barraged with ads that clearly were not meant for me. One particular ad offered me the romance of a lifetime (although I'm happily married), and another sought to convince me to get my degree online (I already have my degree, thank you very much). Most of the time I just ignore these ads, but this one time I found myself asking, "Where was the targeting behind these ads?" Or better yet, "Where was the technology?" Did the advertiser really know who I was?
There is no doubt that over the last five years, behavioral targeting has made significant strides in improving the delivery and ROI of display ads. However, BT's effectiveness may be starting to plateau for one simple reason -- despite all of the technology behind it, behavioral targeting alone cannot definitively tell who is actually on the other end of the PC. Is it a college student, a businessman, or a soccer mom shopping for that car? All of the above are very different audiences, with vastly different tastes and needs. Without a comprehensive understanding of the visitor, behavioral targeting will continue to deliver mixed results, leading to continued consumer ambivalence or annoyance.
3 steps for high performance behavioral targeting
So what can be done to boost the effectiveness of behavioral targeting? I've found that the most effective behavioral targeting campaigns are based on data-driven decisions supported by cutting edge technology. I think the best analogy for a successful data-driven behavioral targeting program is the good old three-legged stool, where each leg provides critical support to each other and enables, shall we say, a more successful sitting experience for the user.
Knowing where to direct your ad buy is the first leg of the stool. Marketers cannot rely on publisher or network assumptions about who is visiting a site because for the most part publishers and marketing-centric or arbitrage-driven ad networks don't have the advanced technology needed to thoroughly analyze their traffic. This analysis simply falls outside their domain expertise. This knowledge vacuum can hamper behavioral targeting's effectiveness and thereby drive up costs. Advanced data mining and analytics can reveal some surprising insights about an intended audience that could have a significant impact on your media buying efforts.
For instance, one insight that we gained from analyzing our own traffic -- specifically a luxury auto maker on our own network -- is that the majority of visitors to the site were aspiring owners just browsing the website, who tend to have lower incomes than potential buyers. Similarly, we realized that people with much higher disposable incomes did not spend much time on the site, preferring to get information from the dealership before making a purchase.
Armed with this knowledge, media buyers might be less inclined to buy inventory on this site for their other luxury goods clients. Without this strong analytic foundation, most behavioral targeting would likely classify these visitors as high net-worth individuals and serve them the wrong ads, thereby lowering conversion rates and increasing costs.
This erroneous classification leads us to the second leg of the stool -- know thy audience. Now, don't scoff and answer me truthfully: How many of you can honestly say that you truly know the person to whom you're serving ads? Do you have a full 360-degree profile, or are you relying solely on the classifications provided by your behavioral targeting vendor? Many of them simply aggregate online behavioral data and lump people into generic categories.
However, two people who visit the same set of websites do not necessarily fit the same profile -- which is why I get served ads for romantic encounters. Instead, a 360-degree profile would be able to distinguish whether a person is an avid golfer, or a simply spouse shopping for a gift.
Achieving that 360-degree profile involves compiling a deeper range of data such as demographics, lifestyle, life stage, commercial intent, and behavior across multiple networks. To this, we take the extra step of buying publicly available offline data and merging it with the available online data. This approach has helped us to target our audiences more successfully. Is this approach the silver bullet that advertisers need? No, but the more information a marketer can glean in real-time, the better their ability to adjust ad targeting appropriately.
This real-time component is perhaps the most critical leg of the stool. Now that you have a better sense of who your audience is, and where to target them, the final step is real-time analytics and ad delivery. Networks and platforms that can crunch all this data in real-time will be better able to serve the right ads to the right targets at the right time. If they can't immediately ascertain a visitor's identity and potential behavior, chances are the marketer has just missed a golden opportunity to convert a customer.
In a business climate where reduced ad spending translates into an urgent need for greater efficiency and ROI, high performance behavioral targeting presents an incredible opportunity for marketers to stretch their ad dollars even further. By leveraging richer behavioral data, and with the right real-time technologies powering their behavioral targeting efforts, marketers will start to realize greater returns from their campaigns because they'll be finally tapping into the right audience in the right context.
Oh, and for those networks who keep on sending me ads for sunless tanning, what I'm really interested in is a new Tag Heuer. You can't get much deeper insight than that.
Dominic Bennett is vice president of engineering at Turn Inc.
On Twitter? Follow iMedia Connection at @iMediaTweet.