The inherent value of inventory lies in a combination of the content that publishers develop and the audience that sees it (their demographics, household income, behavior and intent, etc.). Therefore, we believe that inventory should be priced on an impression-by-impression basis. OpenX Lift uses sophisticated predictive algorithms to accomplish this real-time valuation process. Our algorithms assess the user, frequency, volume, geography, time of day, and session depth in hundreds of scenarios to determine a market-tested price for all non-guaranteed impressions. Then, whenever an ad call occurs, the ad server (OpenX Enterprise or another) uses those predictive algorithms to determine what each of the publisher’s demand partners will pay. But CPM alone isn’t enough to determine an impressions true value. The ad server must also factor in each partner’s historical fill and discrepancy rates.
A price offered by a buyer does not necessarily predict how much revenue a publisher will ultimately get for their inventory. The offered price must be balanced against the buyer’s historical discrepancy and fill rates. A buyer may offer a CPM that’s 20 percent higher than a competitive bid, but if, historically, that buyer has a 15 percent discrepancy rate and a 50 percent fill rate, the publisher may be better off placing the impression with a lower bid from a partner with better discrepancy and fill rates. In short, we believe that the ad server should help the publisher determine the true CPM offered by buyers based on historical fill and discrepancy rates.
See how we make this happen in our new OpenX Lift video (above) and discover more about increasing your yield in our
How to crush your online ad revenue goals whitepaper.