Not so long ago, if you wanted to find the right audience online, the people most likely to respond to a brand message or take a desired action (whether that action was get a click or getting a consumer to think differently about a new Cadillac), advertisers really only had one option. They needed to buy a specific placement on a specific site where those types of consumers tended to congregate. Thus, specific sites or sections of sites (Yahoo or Yahoo news) were a proxy for the right audience. In other words, to reach the right audience you had to buy the right content.
Fast forward to today. The world of online advertising has radically changed – mostly for the better.
Now online ad impressions can be found in marketplaces called exchanges. Advertisers can compete with each other for the most valuable impression by bidding via an auction process in real time — known as Real Time Bidding (RTB). But how do advertisers know what to bid on? Of the myriad of options available, the most common solutions are those provided by audience data companies. Audience data can be extraordinarily effective. In its simplest form, audience data tracks a consumer’s web usage and creates profiles based on browsing habits. So in these marketplaces, advertisers can determine which ad impressions have the right audience and then determine what to buy.
Pretty cool stuff, a fantastic evolution from just having to buy content as a proxy. How can targeting get any better than this? Essentially there is no waste, as an advertiser is only buying the audience that matters most to them – it’s almost the Holy Grail.
Unfortunately, this approach does have its shortcomings. Just because we know how and where to reach a target audience doesn’t mean that these folks are in the proper mindset to view, remember or respond to an ad. For example, if they’re engaged in content in a particular way, like emailing friends or watching videos, they won’t be receptive to certain marketing messages. Or if they are looking at pages that contain user-generated content (UGC) laced with profanity, this won’t be the environment that some advertisers will want to serve an ad into, regardless of the knowledge they may have about the user. Clearly, the ad environment is essential. If you don’t think that environment matters to an advertiser, you’re mistaken.
Take control of the environment.
Until now, the marketplaces I mentioned above have been unable to empower advertisers to define the environment in which they want their ads to appear. Online advertising has been disproportionately based on audience targeting, and some advertisers have dismissed environment as an unnecessary concern. But the environment problem is a major reason large brands have barely dipped their toes into RTB.
By incorporating data about the ad environment into the ad buying process, advertisers can not only control the delivery of their ads onto pages they were previously blind to but can proactively dictate their comfort level with quality and safety in these developing marketplaces.
Here’s how it works.
The technology that enables this approach is page-level semantic analysis. Websites are like networks of pages. You can’t designate a whole site to be entirely safe without knowing what is on every page. For example, any news site has content about death, accidents, catastrophes, crime, etc. Advertisers can steer clear of this specific content if they feel it isn’t appropriate for their brand, while still reaching that site’s audience on other, more appropriate pages.
Page-level analysis looks at individual pages rather than sites or sections to reveal data attributes that indicate the safety and quality of each page. This allows advertisers to proactively define their environment upfront. Once the environment is defined, the ad buyer can then focus on targeting ads to users or content within that environment.
For example, suppose an ad impression has become available on a web page that has a few family photos with captions under each one. Through real time semantic analysis of that page, the marketplace can relay to an advertiser data about the quality, safety and category of that page. In this case, the page would have a low text-to-image ratio, be considered un-moderated UGC, is safe from profanity and mature content and be labeled an entertainment page.
This combination of attributes might be fine for some advertisers and not for others, even if their target audience is viewing that page. With this data, the advertiser can decide which combinations of attributes define the ideal environment for their ads before making a buy or bid decision.
Targeting by environment is a new way to think about bidding for and buying ad placements in these real time marketplaces. It’s not just about the audience, but also where and when to reach that audience that matters. Once the guys with the money get comfortable with the audience and the environment, we will see a tectonic shift in how dollars are spent online.
Andy Ellenthal is CEO of semantic advertising company Peer39, responsible for setting and overseeing the company’s strategy and growth. He was previously CEO of quadrantONE, a joint online advertising venture of major media companies including The New York Times, Gannett, Hearst, and The Tribune. Prior to that Andy led sales efforts for PointRoll, making that company the industry’s #1 rich media solutions provider. In addition he held a number of sales management roles at DoubleClick, Inc., where he was Vice President of Advertisers’ Solutions.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.