Segmenting Intent by Referrers Adds Context (and Performance!) to Keyword-Targeted Display

by Tamir Lipton, VP of Business Development
June 8th, 2015

Josh Dreller recently wrote a piece for Search Engine Land appropriately titled 7 Audience Targeting Data Types Versus Search Data (The Keyword is Still the Undefeated Champ). The story was insightful and indirectly highlighted one of the major shortcomings of legacy ad serving technology employed by publishers –the inability to use keywords and other important factors as a means to target ads.

The absence of keyword targeting through the ad server has forced publishers to offer less effective forms of targeting (audience data, demographic, psychographic, third party data). Publishers dismiss the idea of using keywords to target campaigns because they don’t have the ability or scale to do it.

Keywords are certainly a dominant indicator of intent, arguably the most dominant, but they are only one factor and must be considered in conjunction with others. Even Google uses other variables outside of keywords to predict intent. Stand-alone keyword targeting only works at scale (see: search), so in order for publishers to leverage keywords as a targeting mechanism, it has to be algorithmically scored alongside other factors.

The ad server won’t do that – nor will semantic, keyword targeting companies even if they are integrated into the ad server. Keywords are a component of predicting intent, but by no means the only indicator. Keywords are the start – and an important one for sure – but it’s the start of understanding intent, it’s not intent unto itself.

Yieldbot knows this. In addition to keywords, one of the other non-standard identifiers Yieldbot targets against is referring domain.

EXAMPLE: Traffic from to generates a 0.40% (CTR) for campaign X. Yieldbot uses the referring domain data in conjunction with other factors (keywords being one) to optimize the campaigns it powers in real time.

The table below shows the overall performance of campaigns running on desktop and mobile through Yieldbot to a major publisher in the “home” vertical from May 1-20, 2015 against traffic coming from the sites listed.

The chart below highlights top performing keywords of Pinterest Traffic to the same “home” publisher on mobile and desktop from May 1-20, 2015.

Meanwhile keyword sellers, the ones Dreller refers to as having “dominated digital marketing budgets for more than a decade”, continue to reap the benefits of legacy ad servers – but it doesn’t have to be this way. By reframing how publishers think about technology that looks at the page, there are targeting innovations which allow publishers to approach this very set of media buyers – performance marketers.

Publishers are slowly changing the way they run their business, selling to performance buyers, and concentrating on quality of traffic over quantity. And in the long run, performance and quality is what ALL marketers want, making this new tactic a long-term strategic winner for publishers, marketers and site visitors.

Tamir Lipton is VP of Business Development at Yieldbot. Tamir works collaboratively with sales, client services, and account management to nurture relationships with top-tier Publishers, Agencies, and Advertisers who believe that the best ad technology leads to the best ad performance.

Contact Us

Download Case Study