You are here: Home » Uncategorized » Keyword-Level Contextual Targeting Is A Cutting Edge Behavioral Targeting Strategy
Keyword-Level Contextual Targeting Is A Cutting Edge Behavioral Targeting Strategy

Keyword-Level Contextual Targeting Is A Cutting Edge Behavioral Targeting Strategy

If you are a marketer working in the display advertising channel, there is a cutting-edge targeting methodology called keyword-level contextual targeting that you need to know about.  Keyword-level contextual targeting is unique for being a both behavioral and contextual.  It combines the massive advantages of behavioral targeting with the relevancy and reach advantages of contextual targeting.

Let me explain what keyword-level contextual targeting is, how it works and most importantly, how you can use it to bring massive ROI increases to your display campaigns.

Let’s begin…

In online display advertising, there are two primary  methods for targeting audiences:

  1. Contextually
  2. Behaviorally

In the simplest definition, contextual targeting is targeting content (websites, webpages, etc.) that you think your intended audience is likely to engage with.  On the other hand, behavioral targeting is targeting individuals (actual people) that are likely members of your intended audience.

With contextual targeting, you are targeting content that is related to your ad.  With behavioral targeting, you are targeting individual Internet users that are likely members of your target audience.

Big difference.

Since nearly the beginning of the Internet, marketers have been targeting display ads (ie. banner ads) on websites they believe their target audience frequents most often.  For example, the leading fishing tackle retailer Bass Pro Shops would do far better targeting display ads to fishing websites like BassFishin.Com.  It wouldn’t make much sense for Bass Pro Shops to buy display ads on a fashion website like Vogue–it’s just not a highly-correlated site for their target demographic.  If you wanted to sell high-end cookware, would you rather have an ad on or on  FoodNetwork, of course.

So for years, display advertisers would put together media plans consisting of contextually-relevant websites that are closely related to the advertiser’s products or services, and buy ad inventory on those sites hoping they would either reach the most qualified prospects, or reach the highest percentage of their target audience for their allotted budget.  This was a sound strategy, and still is, but behavioral targeting offers marketers a more precise methodology for reaching the most hyper-responsive prospects online.

Behavioral targeting comes in many forms and can be performed in many ways.  Advertisers often target ads to something called audience segments.  These are quite literally pools of cookied web surfers who have been marked as having taken a specific behavior (ie. taken certain actions) that uniquely qualifies them as belonging to a particular audience.

This is best explained with a hypothetical example…

Person A visits and searches for flights to Orlando, FL three-weeks before the date of their travel.  Upon searching for flightsPerson A is cookied by Priceline as being an in-market traveler to Orlando, FL in three weeks.  Person A becomes a member of an audience segment (cookie pool) that advertisers can use to target ads to.  This audience segment would be valuable for an advertiser such as Southwest–allowing them to target this audience segment of in-market travelers to Orlando with display ads that say, “$89 Flights to Orlando — Book Now”.  This audience segment would also be valuable for an advertiser such as SeaWorld, as they could begin raising awareness with a traveler who is likely to be flying into Orlando in the coming weeks.

Behavioral audience segments (aka. cookie pools) come from a variety of sources.  Generally, audience segments are from either 1st-party sources or 3rd-party sources.  You’ll often hear behavioral marketers talking about “1st-party data” or “3rd-party data”…”data” being synonymous with “audience segment”.

What does 1st-party and 3rd-party mean?

1st-party audience segments (aka. data) are generally cookie pools YOU create through audience collection mechanisms YOU own.  A simple example of this is what is called site retargeting, where you as the site owner of an ecommerce store could drop cookies on shoppers who had items in their cart and prematurely abandoned the site before checking out.  Since the cookie pool was created by YOU, this is what’s referred to as a 1st-party audience segment.

On the other hand, a third-party audience segment (aka. 3rd-party data) would be cookie pools provided by OTHER websites, vendors or any other source that you don’t control or own.  Data companies like BlueKai, Neustar (formerly TargusINFO), and BIZO aggregate specific audience segments from multiple sources and sell that “3rd-party data” (aka. audience segments) to advertisers to use in behavioral targeting.  The size and quality of this data varies tremendously.  Some are incredibly useful and will outperform contextual targeting methods substantially, while other cookie pools are virtually trash and could be a potential waste of significant ad spend.

Blending the Best Traits of Contextual Targeting with Behavioral Targeting

Behavioral targeting is effective because you only spend money to target individuals who are qualified as being part of a specific audience that has performed a positive behavior online.  Unlike contextual targeting where you are essentially showing ads to everyone that visits certain websites (ie. or even sections of websites (ie. ‘Congress’ section of, behavioral targeting eliminates wasteful ad spend on impressions that may or may not be hitting your intended online prospects.  Every impression served is targeted to the right person–not just a likely website that the ‘right person’ might visit.

While one would think behavioral targeting is superior in every way, contextual targeting still has its advantages.  Brand marketers optimizing campaigns for reach still tend to favor contextual strategies since behavioral targeting can be a lot more expensive (audience data costs money!), and the potential reach through all available 1st- and 3rd-party audience segments may be miniscule in comparison, depending on the industry or topic.

And it’s important to mention that some industries may have NO available audience segments at all.  A niche industry, such as the elevator space, may have no available 3rd-party audience segments that would be effective for targeting elevator inspectors.  It’s just too small of a niche to make building online audience data lucrative for data providers such as BlueKai or Bizo.

Introducing Keyword-Level Contextual Targeting

So how do you mix the advantages of behaviorally targeting qualified individuals with the additional layer of relevancy and reach that contextual targeting provides?

One answer to this is keyword-level contextual targeting.

Don’t let the name fool you, keyword-level contextual targeting is actually behavioral targeting–targeting a cookie pool that is created through contextual data.

Now this may sound confusing at first, but let me explain a little more clearly using an example so you can understand the true power behind this incredible display advertising tactic.

Let’s say you a marketing consultant for Crest, and your objective is to raise awareness for Crest’s teeth whitening strips among Internet users who are interested in teeth whitening solutions.  Beyond the more direct-response channels such as search engine marketing (SEM) or social media, display would be a valuable channel for raising awareness if you could achieve two things…

  1. Reach users who show interest in teeth whitening
  2. Reach them at an ad frequency level (# of times a user is exposed to your ad) above 3 exposures.

A higher ad frequency is optimal for raising awareness, since brand research studies show that brand recall is significantly increased with ad frequencies greater than three exposures.  Just a single exposure to an individual results in poor probability of recall, and why most marketers fail to make awareness campaigns effective at small budgets–they simply don’t have the knowledge or resources to optimize small campaigns for reach with high frequencies.

For a product like teeth whitening strips, it would be difficult to run contextual display campaigns without wasting a lot of impressions on disinterested individuals.  You could buy ad inventory on websites that are dedicated to the topic of teeth whitening, but there aren’t many of them–your reach on just these kind of sites would be very small, especially for a company like Crest who doesn’t just want to sell a couple hundred products, but tens or hundreds of thousands of teeth whitening strips.  Additionally, if you were to instead buy contextual ad inventory on broader interest sites (ie. Vogue, WebMD, etc), you’d be running into that same problem of wasting ad impressions on people who haven’t shown interest in teeth whitening solutions.

You’re probably out of luck finding existing 3rd-party audience segments of in-market teeth whitening prospects too.  It’s too narrow of a niche right now.

If you’re Crest, you probably have some good 1st-party data available (site retargeting cookie pool), but since this is an awareness campaign, you want to reach people who aren’t as familiar with your products and teeth whitening solutions yet–to bring them into your sales funnel.

So how do you go about using behavioral targeting to reach interested individuals in teeth whitening online through the display channel?  You need to build your own custom audience pool, and two primary ways of doing this right now are:

  1. Keyword-Level Contextual Targeting
  2. Search Retargeting

I’m going to cover search retargeting in another blog post, so let’s just talk about keyword-level contextual right now.

It’s important to note that as far as I am aware, keyword-level contextual targeting can only be performed through the demand-side platform offered by a company named  I am a HUGE fan of because they are at the forefront of making massive amounts of unstructured data available to marketers like ourselves to target and optimize with.

Without getting into the weeds with technical explanations, it is my understanding that crawls pages that serve ads using the major ad exchanges. is able to save all the keword-data (similar to what Google does) of pages on those pages, as well as drop cookies on people who come into contact with ads on those pages.

So to put it simply, has data (cookied users) that can be matched up to page-level keyword data that those users have engaged with.  In other words, if you want to target users who have viewed pages containing the keyphrase, “get whiter teeth” in the last 30 days, you can do that through’s demand-side platform (DSP).

So let me start you off with a real-life example…

Crest might believe that a highly-qualified prospect would be a person who has engaged with a webpage that contains keywords like the following:

  • teeth whitening
  • whiter teeth
  • whitening teeth
  • get whiter teeth
  • teeth naturally yellow
  • yellow teeth
  • healthier-looking teeth
  • and so on…

So with keyword-level contextual targeting (provided by’s DSP), you can create a campaign targeting users who have engaged with pages that MUST contain one of your specified keywords.  You can even set what is called a recency window for how recently someone must have engaged with pages containing those keywords.  You can set your recency window to instant, which essentially ensures pure contextual targeting to pages that contain your keywords.  Below is an example of keyword-level contextual targeting with an instant recency window:

Keyword Contextual Example

An example of how keyword-level contextual targeting can work. Crest could target Internet users who have recently visited webpages containing certain keywords, such as “whiter teeth”, “whitening your teeth”, “get whiter teeth”, etc. It’s like retargeting off of other people’s content.

But you can also set your recency to 1-day, 1-week, 30-days, etc.  So if you wanted to target Internet users who visited pages containing those keywords in the past 30-days, you would set your recency to 30-days and’s DSP will serve your ads to those users across any site on the exchanges that you make a winning bid on–regardless if the webpage is about teeth whitening or not–making it true behavioral targeting.

This is powerful.  It’s like retargeting off of other website’s content.

The buying of ad inventory through this methodology is all done in real-time via RTB (real-time bidding).  For those of you unfamiliar with RTB, you set a maximum bid price (CPM) for a campaign you create, and the demand-side platform (DSP) is the platform that is integrated with the ad exchanges that will bid on ad inventory that fits your defined targeting parameters and criteria.  Think of it as an auction-place of advertising–if you win the bid, your ad is served.  The amazing thing is this all occurs in a fraction of a second!

This past year, I’ve seen massive improvements in both brand and performance metrics with keyword-level contextual targeted campaigns compared to just pure contextual campaigns.  It is typical (with very good targeting) to see 30-100% improvements in brand/performance metrics through this style of behavioral advertising–mostly due to the decrease in impression waste that occurs in pure contextual.

In a future blog post, I’ll go in-depth on how to best strategize and execute a keyword-level contextual targeted campaign.  Leave your comments and questions below and I’ll do my best to answer any questions.

About Kevin Scarselli

Kevin Scarselli is an accomplished Internet marketing expert and Internet start-up entrepreneur. He also provides strategic marketing consulting to some of the largest companies and brands in the world.

Leave a Reply

Your email address will not be published. Required fields are marked *