The relationship between online search and advertising is inextricable. With any new development in search technology comes a correlated development in advertising technology.
Today, the search market is buzzing about semantics (i.e., using the science of meaning in language to produce highly relevant results). It's a departure from keyword methods, which use ranking algorithms to predict relevancy (e.g., everyone's favorite search engine, Google).
Now, the advertising market is buzzing about semantic advertising, because like semantic search, it holds promise to improve the all important issue of relevancy.
Advertisement relevancy is the core issue affecting the online advertising market. Frequently, ads appear on pages that have little significance to the assigned content, thereby rendering the placement ineffective or producing counterproductive effects. For example, an ad for a Caribbean vacation package may appear near an article about a massive hurricane that ripped through that region -- not exactly the association that the brand manger at the vacation company wants!
Let's consider a few examples that further underscore the problems of contextual ad relevancy:
In this example, an ad for a free dinner at Olive Garden is placed (via keyword ad system) next to an article about 250 people getting sick after eating at an Olive Garden restaurant in Indiana.
In this example, we see ad for a Samsung Blast cell phone next to an article about how a man might have possibly died from the explosive blast of a cell phone battery, again driven by the match of the keyword "blast" without using any contextual background to the use of the word.
In this example, a banner ad on CNN for a moving company, Putyourfeetup.com, has the regrettable distinction of being placed above a story about severed feet washing up along British Columbia's shores.
What all these ads say, rather loudly, is that there's a real problem with business as usual in the advertising arena -- keywords do not take into account context, and they should. Not only do these misplaced ads not reach their intended audience, but it's clear that there's also real potential that these ads are reinforcing negative brand impressions.
Real-life examples like the above keep business executives up at night, but with the advent of semantic advertising, executives can now rest assured that their ads are properly targeted and set alongside only the most relevant articles.
Semantic advertising 101
Improving relevancy isn't just about avoiding negative associations with articles on a given page. By improving relevancy, semantic advertising also significantly increases clickthrough rates, which, in turn, ensures that the ad is relevant to the content that the reader originally engaged with.
For advertisers, understanding the interests of the reader/potential customer will mean a more effective, revenue-generating, contextual advertisement. With semantic advertising systems, agencies and publishers can now unlock previously hidden value on the content of a page (as well as blogs and social networks) by publishing more relevant ads.
To optimize the supply and management of these contextual advertisement outlays, semantic advertising systems rely on at least three distinguishing features: precise automation, real-time analysis and monitoring and socio-cultural connections.
Semantic ad systems analyze and automatically understand the text found on a page, identifying with great precision the topics discussed and, by interacting in a transparent way with the advertising server, selecting the most relevant ads to be included on that page.
While traditional keyword systems can also be automated to load advertisements onto a page, they still run into the problem of finding the correct meaning.
Take, for example, the word "jaguar." If a page contained the word "jaguar," the keyword system would not be able to differentiate between the car brand and the animal, as you'll see in the below example on the New York Times' website. Conversely, a semantic system would understand that a reader interested in the story about a jaguar's natural habitat probably isn't interested in buying a Jaguar from the Great Neck, N.Y. dealership advertised next to the story.
Real-time monitoring and analysis
Semantic advertising also offers an additional advantage by increasing a publisher's view of its content through real-time monitoring and detailed analysis of web page text.
This semantic analysis extracts other relevant information on a page -- such as the main concepts, actions, event, sentiment, time, place, people, organizations and other key information. This allows a publisher to automate and deepen the niches in which to categorize content. The traditional means of doing this is to have a small army of staff to read and categorize. Of course, this does not scale well due to costs, so such efforts have proven crude at best over time. All advertising is about finding and exploiting ever more specific niches. Semantics provides the specificity at a cost that is palatable.
Lastly, semantic advertising also allows the ad manager to place ads based on socio-cultural correlations and identifying consumer trends. For example, a user interested in hybrid vehicles may also be interested in organic food. The relations are not direct, and may not include the same keywords, but they are incisive and effective.
Let's sum up where semantic advertising can take companies and providers with the bottom line in mind:
- Lower costs to advertisers. Semantics means better precision, and that also means fewer wasted impressions. Fewer wasted impressions means less cost to the advertiser.
- Better clickthrough rates. Better precision means more clickthroughs. Depending on the ad network's economic model, it can also mean higher revenue.
- Better ad rates. Better precision means ad networks can boost the asking price per ad. So again, more revenue flows to the network.
In the end, semantics significantly changes the revenue and cost flows of online advertising, but in a way that benefits advertisers, ad networks and the consumers who see the ads. It is a win-win-win.
When you're considering making an investment here, it's important to consider whether these vendors truly offer a wholly semantic approach. If the system relies on other methods beyond semantic technology (such as statistics or heuristics), be wary. Otherwise you could be investing in a solution with some of the same inefficiencies as the current keyword systems we've discussed.