Semantic Targeting: No Cookies? No Problem


anticookie_small.jpgADOTAS – Semantic targeting is coming to be seen as the next generation of online advertising technology. While it is certainly compatible with predecessor technologies, like contextual and behavioral targeting, semantics avoids some of the pitfalls that plague both of these technologies.

In particular, the industry is currently taking a close look at the privacy concerns associated with behavioral targeting, which may severely limit its growth potential as a valuable solution for targeting ads.

Behavioral targeting has become the focus of prominent members of the FTC, as well as various advocacy groups, who wish to protect the privacy of the users who are subjected to this technology. In addition, U.S. online privacy legislation groups have been governing the use of cookies for years, while the EU introduced similar regulations nearly a decade ago and continues to refine the details.

In response, the online advertising industry’s big players are attempting to placate the powers that be with more visible “opt-out” tools. Yahoo’s Ad Interest manager and Google’s Ad Preferences are recent initiatives that easily allow online users to personalize the quantity and visibility of their cookies. These companies have also begun to delete user IP records within a shorter time after acquisition.

Such measures are sure to proliferate, as dependent companies circle the wagons around a technology that directly accounts for at least $1 billion in ad revenue per year, and shows a bright future as a scalable ad targeting solution.

Is the FTC making much ado about nothing? Yahoo, Google and others argue that consumers are willing to sacrifice some privacy in exchange for cheap or free content.

However, the statistics do not verify this. A recent Synovate survey claimed that 35% of Americans would not accept technology that serves ads based on their Internet viewing behavior, even if the data could not be used to uncover their identity.

Perhaps behavioral targeting horror stories have reinforced, or even increased, this suspicious attitude. Even user-friendly Google mistakenly exposed private user information this past year. And some privacy watchdogs still cite the infamous Thelma Arnold case, in which an AOL user was identified through publicly released records despite AOL’s assertion that the information could not be traced to individuals.

No matter what pro-cookie pundits claim about the confidentiality and uselessness of tracking information, perceptions rule.
Semantics Reads Between the Lines

Semantic targeting takes a different approach to identifying the interests of Internet users, and relevant advertising for them. Semantic algorithms can recognize the true meaning of a page of content by analyzing associations between words and phrases in web text.

A basic example is the ability to distinguish between “jaguar” the animal and “Jaguar” the automobile by detecting related words such as “jungle,” “predator,” “car,” “luxury,” and so forth.

Semantics uses this capability dozens of times per page for every page on a website. Once the overall subject of a web page has been defined, a semantic system sends the most relevant and appropriate set of ads to that page.

The result is a combination of advertising and content that is closely related. The success of semantics lies in the belief that someone who is reading content about a certain subject is also interested in learning more about, engaging in or purchasing a related product or service. In the real world, this relationship holds and has been verified through focused testing.

Besides delivering accurate targeting, semantic technology avoids the use of cookies and the wrath of privacy advocates. In addition, semantic systems are easy to implement, feature real-time analysis and allow the application of one piece of content to several advertising categories.

For publishers and technology companies that rely on behavioral targeting, semantic systems are a vital alternative when too many users are opting out. The negative attitude shared by many users regarding cookies and the threat of stronger legislation means that behavioral targeting likely faces a bumpy road ahead. Semantic technology can act as a front end for behavioral systems, combining the advantages of the two approaches to create a very effective tool.

Semantic advertising also exists, of course, as a standalone system that is ideal for premium publishers, networks and exchanges. All of these entities benefit as the entire ecosystem experiences a lift from advertisers that see improved return on investment due to more effectively targeted ad campaigns.


  1. Unfortunately, semantic targeting has no ability to work on the websites that serve the majority of banner ads.

    The issue with semantic targeting is that it becomes contextual advertising on websites where news is served and a variety of story headlines/excerpts are present. When too much data is present, semantic targeting falls flat.

    Additionally, semantic targeting cannot take into account previous behaviors before reaching the website content. Was traffic push or pull marketing? This stuff matters and semantic targeting doesn’t account for it.

    In summary, semantic targeting cannot find audiences. Advertisers want to buy audiences, not a hunch. Semantic targeting is valuable only when combined with behavioral targeting, not independent of it.

  2. John, thanks for your comment. Let me explain semantic targeting a bit further in a way that will address your post. Semantic and behavioral don’t compete with each other. Rather, older technologies like behavioral and contextual are fully compatible with semantic targeting. To be clear on the results of semantics, we see a demonstrated increase in advertising performance and eCPM’s. To illustrate, Peer39 recently hosted a webinar for the IAB (Interactive Advertising Bureau) where we discussed the effectiveness of semantic advertising targeting. I invite you and any other readers to view that webinar in full here

    You are correct in stating that advertisers are interested in targeting to particular online audiences. However, because behavioral technologies rely on cookies to target ads, its effectiveness in delivering audience has its challenges. According to Jupiter Research, up to 40% of internet users delete their cookies at least one a month and people also share computers, thus eliminating the ability of true behavioral targeting. In addition, as has recently been covered by The New York Times and The Wall Street Journal, government regulators in the US and abroad and others are taking a close look at the privacy concerns associated with behavioral. Because Peer39 uses semantic technology instead of cookies to target online advertising, the industry is finding our solution ever more compelling. In my view, we’ll see a more cooperative approach that will leverage the best of all the technologies at our disposal to lift the value of online advertising. In any case, we expect 2010 to be a great year for us and new targeting technologies in general as good, reliable data becomes an ever more essential component of the online advertising ecosystem. – Amiad, Peer39 (


Please enter your comment!
Please enter your name here