Web Behavioral Targeting
To keep you ahead of the competition!

So what is this Web Behavioral Targeting stuff? Well simply, it's targeting web users based on their previous behavior or activity on the web. And the best way I can think to understand Web Behavioral Targeting, especially when you compare it to say a contextual placement, is that BT (Web Behavioral Targeting) targets people, not pages.

I tend to categorize Web Behavioral Targeting into two distinct approaches: The first uses past behavior to better target an existing audience on in-market sites; many refer to this as Retargeting. The second approach employs context sensitive messaging to target in-market shoppers outside the comfort zone of a company's specific content.

Let's start with the publisher side of things. This is where individual websites actually track the behavior of those people who are on their sites, and they're really trying to identify those consumers that are like minded. You know, birds of a feather that flock together kind of thing. These are people who display the same online behavioral patterns. And by a very sophisticated algorithm and all sorts of other scientific insight, they're grouped together into segments that could be very actionable for advertisers.

The trend in today's marketplace is moving from a focus on quantity of pages to quality of people. Unlike traditional advertising, behavioral marketing treats the internet consumer as an individual, not as part of a mass demographic.

This has proven to be so powerful that one client even admitted to its agency, “I need the interactive medium to make up for the inefficiencies of my other marketing channels.”

There are a lot of people that provide this service. Yahoo! provides it for themselves on an internal basis. But the big players right now are TACODA, Revenue Science, AlmondNet, Accipiter, and Drive PM. These are the folks that are actually working with the publishers to provide Web Behavioral Targeting.

 

Web Behavioral Targeting Section.1

Launching a Campaign

When Web Behavioral Targeting first launched a few years ago, everyone was excited about the incredible possibilities that it offered to advertisers. Pundits endlessly spun out scenarios for advertising tailored to an individual's specific interests. Then the more tactical questions started to arrive: Does it work? How is it bought and sold? What is the scale?

If the tremendous growth in Web Behavioral Targeting over the past year and the predictions for 2007 spending are any indication, these questions have largely been answered. With those operational issues resolved, it is time to turn some marketing attention once again to the more advanced possibilities that Web Behavioral Targeting holds for marketing.

That is to say, Web Behavioral Targeting can help move a consumer through the four key stages of the purchase decision cycle:

1. Brand Awareness
2. Message Association
3. Brand Favorability
4. Purchase Intent

It all comes down to trying to gain a deeper understanding of one thing: which ads work and which ads don't. Not what a focus group thinks works. Not what the client thinks works. Not what the creative director thinks works. It's only about what a million or so potential customers think works.

To achieve this, you need to view ads in the nakedness of their own performance metrics. Like people, most ads don't look good naked. But online has a definite knack for stripping away everything but the grisly truths, like cost and acquisition-based metrics. So that's what you need to look for: grisly truths, in plain up-front detail.

 

Web Behavioral Targeting Section.2

Why test my creative online?

This approach to testing wasn't practical just a few years ago. While online advertising metrics have been around since the first banners, the big change came in the past several years with post-impression tracking.

Post-impression tracking at one time was expensive -- prohibitively expensive, given the ravaged state of online advertising in the early 2000s. But now that's all changed.

What I find remarkable is that there are companies that still don't use it.

Of course, there are a multitude of ways to test almost anything. And I'm not discounting other research methodologies, most of them provide good insight at varying levels of relevance. But strategic and creative concept testing has never really worked well in focus groups, the predominant method for testing them. That's my opinion, and the opinion of many others.

Online testing provides what focus group testing doesn't: a natural test environment. Online advertising happens in real time, in real life. It's real, and so the results are more likely to be real too.

Online advertising is also relatively inexpensive. From producing the creative to generating a small media plan, the costs are marginal compared to most advertising budgets. Ironically, you'd probably spend more on focus groups than on an entire online test. And at the end of the online test, the campaign will already be up and running.

Web Behavioral Targeting Section.3

The first steps

First a project leader would work to define the overall goals that the marketing campaign is intended to achieve. In addition, they'd develop a "creative brief" that would outline more details on the organization's target customers and the actions the project leader wants prospective and current customers to take. The leader would also disclose his initial thoughts for various marketing tactics that would be considered a "stake in the ground" to start generating ideas.

The project leader/team would then be responsible for coalescing the feedback into an initial plan that gets put into action. This is where the real work begins as being a visionary starts to take place.

A communications tool -- such as an email or blog -- will communicate ongoing campaign results, insights and course corrections. Through this process, the entire marketing community can benefit from the collective input and learnings. Undoubtedly, there will be active debate on how to evolve based on the results.

 

Web Behavioral Targeting Section.4

The testing construct

Theoretically, a test should start at whatever point your confidence in your research drops off. For a new product in a new category, that normally happens pretty quickly.
I like to begin with four or five different tests, divided into four phases.

 

Phase 1: Strategic Foundation

Phase 2: Tactical Drivers

Phase 3: Emotional Drivers

Phase 4: Extended Analysis

 

Phase 1: Strategic Foundation

My first goal is to determine which of several foundational strategies will get the best reception. At this stage, however, our goals are to determine core messaging strategies and, with luck, to identify how different market segments respond.

 

Phase 2: Tactical Drivers

Tactical drivers are the blocking and tackling components of the campaign. What size of media units work best? Do the ads need a strong call to action? Will a visual of the product in use be more effective than a straight product shot?

 

Phase 3: Emotional Drivers

This stage builds heavily upon the campaign intelligence that's been gathered to date. It's a little more abstract, as it gets into the emotional aspects of the message. I will usually pit an ad promoting a positive experience against one promoting the avoidance of a negative experience. I'll test polar lifestyle approaches. The idea in this phase is to cover a broad range of styles, placing much greater emphasis on the creative execution.

Phase 4: Extended Analysis

Once Phase 3 is finished, I would take an additional look at using some optimization tools, for example -- 24/7 Real Media -- has to offer. While these optimization tools are very effective, they could skew your early results. So I would introduce them after you have established your strategic and creative baselines.

Once there, though, the optimizers can really boost a campaign's performance

 

Web Behavioral Targeting Section.5

---

 

Methods of evaluation
You would normally monitor a lot of data points throughout a campaign, but here are the key points:

• Visits per 1000 impression
  This factors in both near-term and long-term impact by combining click response rates with post-impression visits to the home page.
• Site visits
  Comparing site traffic levels before and after the campaign is another way to evaluate the campaign's overall impact.
• Visits to purchase page
  Driving traffic is one thing, but this will enable us to qualify the traffic by monitoring what percentage went to the purchase page to look into the price.
• Product purchases
  As we've mentioned we want to approach this realistically, and while under these conditions we don't expect runaway sales.  Monitoring purchase data will help evaluate expected ROI for this and future campaigns

Web Behavioral Targeting Section.6

Final notes on the testing model

There are many ways a test like this can be structured, depending on a number of issues related to the product or brand's situation.

Although I usually use a 12-week test, in actuality, monitoring the performance of campaigns and trying variations of creative should be an ongoing practice.

There are a number of details to be considered with an online test: The creative presentation, frequency caps, the network or media plan employed, and the sequence of the creatives. Those are but a few factors that can skew your results one way or the other. As long as they're accounted for, the results should prove reliable.

The potential for online testing is enormous, and I'm always interested in any input that can further enhance and demonstrate the capabilities of online advertising.

As time allows I will add more to this topic.
Thank you for your interest

 

Web Behavioral Targeting .7