Barring a massive fiscal collapse, somewhere between 2 and 3 billion transactions will be made on the New York Stock Exchange tomorrow. The people trading and bidding on these stocks, bonds, and other derivative products will fail or succeed, in large part, because of the information they have that drives their trading. Information, as they say, is power. Yet, when applied to the exchange known as online advertising, information is gathered in hindsight. Now that the real-time bidding marketplace is expanding in triple digits, which lack of proactive intelligence must be filled.
RTB has been gaining traction over the past year. While at first it was accused by its critics as too much automation in a business where professionals still value the human touch, there were just as many excited about the prospect of RTB as there were those wary of it. Now it is reality: It's here to stay. During the first quarter of 2012, the supply of RTB impressions expanded 120 percent over the same quarter in 2011, according to results from a new quarterly tracking report being published by independent agency trading desk Accordant Media. The report estimates that the supply of RTB impressions soared 213 percent during 2011 versus 2010. This expansion is being driven by blue-chip advertisers and agencies, which, in turn, is attracting premium publishers to participate in the marketplace.
All of this is great news if the marketplace is approached proactively and predictively. That's where the human touch will stay in this market. We need to accept, however, that the rapid rise of RTB has moved the intent of online advertising analytics and metrics. A year ago, the sharpest companies had the best rearview-mirror results from clicks, to unique users, and even offline brand lift. That's still useful information, but it's not enough. Now leading-edge companies need to define the outcomes they want and predict the best way to achieve them.
Let's break down the "why" for buy-side RTB analytics. Before cookie targeting, a media planner would buy Vibe, SI, ESPN, etc., as a proxy for their audience; let's say the target demo was "sports enthusiast." Since then the industry has completely flipped on its head. The media planner no longer cares about the content. He just wants to find the cookie associated with "sports enthusiasts." He's not thinking about brand safety concerns, and that's his first mistake. His second mistake is not realizing that all sports brands are targeting the same "sports enthusiast" cookies. As it turns out, all auto companies are targeting the same "in-market auto" cookies, all credit card companies are targeting the same cookies, and so on and so forth. This process won't scale, and these impressions are massively overvalued. The targeting capabilities have become so acute, and the pricing for online media is still not efficient by any measure. It is not the future of internet advertising.
By using user cookie targeting exclusively, buyers are accessing only about 20 percent of all available inventory. Keep aiming campaigns at the same 20 percent of cookie-based inventory, and you enter diminishing returns rather than efficient ROI. Again back to the analogy of a stock market exchange. If only 20 percent of the stocks had historical and projected financials, what would happen? Stock pickers would purchase those stocks until some market equilibrium, at which point ROI would be near impossible to achieve. That's what happens if you're only cookie targeting, and it gets even tougher and cookied inventory smaller when you add in regulation.
The problem is that there is a measure of success on Wall Street: cash returned. Clicks aren't the standard in performance nor should they be. There is an industry-wide shift in how advertisers and agencies are measuring success. Those on the front lines of the 3MS initiative (Making Measurement Make Sense) are proposing that clicks and page views no longer make sense -- or define success -- for brands if their ads are never in view. Although standards are not yet set in stone, 3MS defines an impression as viewable if at least 50 percent of the ad is in view for at least one second. As new standards like viewable impressions are adopted, advertisers need new types of data to open up new and valuable targeting opportunities.
Let's start with the planning phase. If a brand is planning a new campaign, it will still want to match brand with audience and audience with media. But it will want to do so at the page level. It will anchor on a cookie and a URL, not just a cookie. Anchoring the ad on the URL gives a brand access to 100 percent of valuable inventory. It gets beyond the 20 percent of over-served and over-valued inventory, and it provides an information set far beyond the 20 percent it accesses by focusing only on cookies. Instead of gaining information on page views, this method produces data on context, engagement, appropriate content, and viewability.
Sure the market and technology need to advance from the binary decisions of the past like, "Does it have the cookie? If Yes, bid $5; if No, don't bid." Fortunately, this shift is already taking place. At AdSafe, we've found that close to 40 percent of all exchange traded ads are never seen to a user throughout the entire user session. That means the ad didn't even scroll into the viewport of the browser. Any algorithm is flawed, whether it's within a DSP or a high frequency trading system on Wall Street, if close to 40 percent of the variable inputs into the model shouldn't be there. It doesn't matter what performance you're after; it's simple stats.
The 40 percent figure is just that -- a stat. It is an indicator of wasted dollars and a call for better viewability solutions. Predictive viewability sets that can inform a pre-bid purchasing decision. This is the knowledge that informs RTB. For example, if a buyer knows that, on a given page, an ad is seen by the user 10 percent of the time, he or she can bid 10 percent of the previous, uninformed bid and only buy if the bid wins. Bidding no longer has to be a "yes/no" model. Now in a more informed, predictive, multi-variant model, buyers must decide both if and how much to bid based on the page's context, level of ad viewability, and other data anchoring on a URL.
Then the outcome (human) has to matched with the media (automation). If a brand is out for awareness, it will want the outcome to be centered around engagement and content relevance. If a brand is out to drive e-commerce sales, frequency, and valuable inventory will be the keys. Finding out after the campaign that the brand got close to these goals is not state of the art. Predicting where the brand will reach that perfect intersection of engagement and content is the right approach. Content rating systems, knowledge of engagement time per ad execution, and predictive viewability is the right approach.
It's not the stock market, but it's a market nonetheless, and it's rapidly becoming more efficient. RTB must be seen as an advantage. The data leveraged by each buyer and seller will be integral. It must be seen as a media strategy that is fact-based but driven by human intelligence. We're not playing with house money anymore.
On Twitter? Follow iMedia Connection at @iMediaTweet.