Commentary

Why Contextual Advertising Fails

There is something missing from contextual advertising ... context!

Am I actually implying that the entire basis of the revival of the online advertising sector is flawed? I sure am! Got your attention now?

The basis of contextual advertising is a desire to make advertisements in some ways relevant to the themes of articles; it has been this way since Bill Gross and his team came up with the idea, way back when. The idea of linking ads to the context of a page of web content has created a monster. It is the monster,however, that refuses to go back into the box.

Of course the model has been adapted by other leading lights in this sector, none bigger than Google with its monolithic Adsense products. Many a new advertiser has been drawn into this sector and contextual has driven a new revenue model, creating a plethora of "me too" operations.

So what is the problem?

The genre has failed to adapt to the increasing volume and react to the massive reach and diversity that this growth brings about. The increase in volume presents issues with the increased amount of language usage and sheer breadth and depth regarding terminology.

Why is this an issue?

According to language experts, for every world in the English language, there are approximately 2.4 alternative meanings. Added to that are the countless company names, product brands, domain names, people's names; the average soon increases to nearly four alternative meanings. Take, for example, the word orange. It has some 17 alternative meanings - the name of a town, a river, a telecom company, a fruit, a color, etc.

In failing to take this into account, there are some glaring anomalies ... ads for a garden bridge on a page about contract bridges; ads for condoms on a review page for the movie Troy; ads for nursery furniture on a page about sudden infant death syndrome (or SIDS). The list is endless.

There are also some cultural differences. For instance, selling a fanny pack to a British market won't result in many sales. Let's just say that "fanny pack" in the UK is used to refer to a very specific part of the anatomy -- taboo for a general advertisement.

Another issue is with the execution of contextual advertisements. Delivery has been blighted by poor execution. For example, many in-text providers have been somewhat profligate with their links. It is not unusual to see the links in the headlines, links on links and other spurious places. It is also the selection of the keywords to highlight. Take, for example, a web page seen recently where every iteration of the word "its" was emblazoned with an ad for an "International Travel Service."

Many of the issues are faults of the advertisers, and not necessarily the service providers. Poor keyword selection is the downfall of any contextual campaign. It should not be beyond the realms of humans to do quick searches to see what other senses of a word can be found. Some examples are quite hilarious while others are just downright shocking. We recently discovered that "white slavery" was used as a keyword to drive traffic for a well-known international website. What was the marketing department thinking when it made this selection? As you can probably imagine, the resulting ad placements were inappropriate, offensive to many, and potentially damaging to the advertising organization.

With all these pitfalls, wouldn't it be nice to have a solution that took all the guesswork out of the equation and was able to place the ads in real context? Thankfully, just such a solution exists in the form of semantic advertising. This technology examines the page for its themes in its entirety. In doing so, it is able to identify the theme with the most relevance to the core subject of the ad campaign and to link the two together. The semantic advertising system creates a hyper-targeted ad that not only will boost campaign performance, but will also create additional brand recognition for the content. This creates a classic win-win scenario for both advertiser and publisher alike.

It would be churlish to suggest that semantic advertising is going to replace contextual overnight, but unless contextual adapts to the growth in the market and becomes more sophisticated in its targeting, it is not inconceivable that in time semantic advertising will do to contextual, what happened to traditional display advertising back in the heady days of the early internet. Time will tell.

5 comments about "Why Contextual Advertising Fails".
Check to receive email when comments are posted.
  1. Walter Sabo from HitViews, January 12, 2010 at 10:04 a.m.

    This is a brilliant and extremely useful article. Walter

  2. Kacy Maxwell from GAPRC Group, January 12, 2010 at 11:58 a.m.

    I have experienced this issue first hand. Our first foray into using contextual targeting for a previous client, we chose a keyword in which we were assured by Google that our ads would be safe to target. However, we appeared on a very inappropriate desktop wallpaper site because that word was one of their optional designs. The science is still not perfected at all.

  3. Daniel Redman from Evisibility, January 12, 2010 at 1:29 p.m.

    Seems like a pretty self-serving article. Sherpa reports that contextual advertising was the top performing segment of display (as reported by marketers), so it cant possibly be as bad as you've outlined. I'd estimate that you could have taken the stance that you are making a good product better with semantic advertising, rather than making a bad product good. You failed to mention that you can obviously optimize a campaign that is targeting contextually (Google). If you are appearing in non-relevant placements, it's as easy as adding a 'negative' in adwords. Yay Google!

  4. James Oppenheim from Peer39, January 12, 2010 at 3:21 p.m.

    Great article! Very nice job highlighting the areas where Semantics can help make ad targeting more accurate and more efficient. Semantic advertising is looking to be the wave of the future, and companies like adpepper and Peer39 are helping people understand the need for Semantic data solutions.

  5. Luca Scagliarini from Expert System, May 19, 2010 at 10:11 a.m.

    Sacha, this is a very good article that stresses a potentially revolutionary development in the online advertising business. Expert System is your competitor but we agree on the potential of semantic techonologies in this area.
    Hopefully, the opportunity generated by this shift will not cause the usual gold rush that we have seen in other sectors of the software business with companies pretending to address a real business need with half baked products and a big marketing budget. The market is still learning the capability of semantic advertising and whoever will be able to rovide an easy a simple way to present and to design semantic based campaigns (together with a strong and mature technoogy) will win the race.

Next story loading loading..