Behind-the-Scenes Innovations in Behavioral Targeting

“It’s the data, stupid” might well be the mantra of the direct response industry. Or a corollary, “Database segmentation and analysis are the key to ROI.” Much of the analytical methodology and discipline are now popping up with a variety of behavioral targeting products and solutions.

Last month during ad:tech, I spent quite a bit of time talking to executives from ad networks, publishers, and other suppliers about what’s new in behavioral during. I may sound like a broken record, but I continue to be astounded at how much of the methodology behind many new targeting solutions is firmly based in direct marketing, data-driven practices now moving into the online space. The level of sophistication is rising: solutions are utilizing the techniques of working with anticipated clickstreams to predict future behavior, as opposed to merely measuring past behavior. This is good news for advertisers!

Epic Advertising

I met with Don Mathis, president of Epic Advertising (previously known as AzoogleAds, plus an acquisition or two), to discuss how his company has been using behavioral targeting and about some upcoming changes. Epic Advertising now encompasses two main units focused on online consumer traffic acquisition: AzoogleAds and Bazaar Advertising. AzoogleAds continues to provide broad-based traffic acquisition services, leveraging the company’s performance-based ad network business. Bazaar Advertising continues to offer search engine management and marketing services, specializing in keyword discovery, purchase, and optimization of online search campaigns.

While the company had previously offered a robust, classic behavioral targeting solution to its advertisers, last week it announced the official launch of Performance CPM (pCPM), a metric that allows advertisers to track performance marketing campaigns by measuring induced visits. Advertisers can now determine campaign effectiveness using a new, more comprehensive approach.

PCPM’s premise is rooted in induced visits — visits to an advertiser’s site tracked beyond direct clicks. A visit to an advertiser’s site is induced if it results in any way from an ad, even if there’s no immediate response or direct click path from ad to site. This metric goes beyond traditional campaign measurement tactics by employing scientific techniques to discern the broadest types of cause and effect, excluding things like accidental clicks.

Along with measuring induced visits, the metric encompasses elements from traditional tracking methods — CPC (define), CPM (define), and CPA (define) — as well as inherent branding impact to offer a more holistic view of an online ad campaign’s success. Because pCPM involves only statistics, it avoids any retention of personally identifiable information (PII).

An advertiser’s ability to use these new metrics to increase performance of behavioral targeting campaigns will increase optimization effectiveness. Mathis stated, “This method lets our advertisers do two new things: they can see the impact of an ad beyond direct clicks, and they can put a solid measure of user-impact on their traditional brand-oriented CPM campaigns. We believe our method, marrying marketing and technology, is the next step in online marketing campaign tracking and will serve as a model for the industry.”

It all started to sound like smoke and mirrors until I spent additional time talking to Bill Softky, the company’s chief algorithm officer. His academic background is in brain science and pattern recognition, and his mathematical work on neural irregularity and correlations is intimately related to the statistics of online advertising traffic.

I was intrigued with Softky’s further explanation of induced visit. An induced visit from an ad is any visit that can be statistically linked to viewing the ad. Because it’s statistical, you can’t just have one induced visit; you have to look at lots of ad views and lots of site visits, then do some careful correlations to discover how many more site visits you got with the ads than you would have had without them. The math behind the metric is just subtle enough that others aren’t doing it, but it’s actually fairly simple and has been used in signal processing and neurobiology for decades.

Collarity

Collarity also offers a very sophisticated behind-the-scenes approach to behavioral targeting (it refers to it as dynamic marketing). A behavioral platform that optimizes content and ad targeting, it’s utilized by major publishers on the back end.

I spent some time talking to Deborah Richman, SVP, marketing and business development. “We’re undergoing a revolution in behavioral targeting because it won’t take so much work to target anymore,” she said. “Publishers and advertisers spend a lot of time categorizing and defining content. With behavioral platforms like Collarity, all people visiting a site can dynamically and effortlessly define what’s interesting instead. This drives far better targeting from the existing ad inventories. In fact, this will be necessary if ad targeting is to work and effectively monetize social and multimedia content today.”

Current behavioral ad targeting is based on preset categories or segments, whether demographics, interests, or site subject matter. Targeting occurs when sites/destinations identify content based on what is mapped to these attributes. Targeting also occurs when people are directly identified or tagged based on where they have visited. One of the down sides is that some visitors are targeted while others aren’t.

Current contextual ad targeting can vary based on the content on each page, as well as on categories set by publishers. Visitors will see the different ads on different content pages. The ads are shown based on words on the pages. Multimedia content and ads depend on keywords tagged to them.

Collarity’s behavioral platform addresses several problems: It applies to every visitor anonymously, without tagging them. It dynamically targets ads based on evolving content and interests. Each visitor sees different ads as she clicks around and is joined to communities of interest. This results in much higher consumption and satisfaction for visitors, with ads that seem more relevant.

While this is once again in danger of falling into the smoke-and-mirrors genre, it really doesn’t, once you look behind the scenes. Publishers using Collarity harness 100 percent of all visitor clickstreams anonymously. These include content and ad consumption of all types — which is a significant statement given the explosion of social networks and CGM-based sites. Community segments form based on intensity (rather than popularity) of interests. Based on these dynamic segments, both content and ads are show to visitors.

Collarity optimizes text, banner, and video ads, relying on ad inventories and feeds used by a publisher. It can deliver the most relevant ads, which typically doubles ad consumption. Ad buyers can see these dynamic profiles from publishers.

The common theme for the two companies is the ability to tap into the nuances of intuited or induced behavior to predict future performance. As our industry becomes more expert and rigorous in technological capabilities, we’ll all benefit from more relevant ads, less ad budget waste, and happier end users. What’s not to like?

Related reading

tencent_emily-ma_featured-image
nurcin-erdogan-loeffler_wikipedia-definition-the-future_featured-image
12919894_10154847711668475_3893080213398294388_n
kenneth_ning_emarsys_featured-image
<