Paul Pellman is CEO of Click Forensics, a traffic quality management company which released "an upgraded version of its Yahoo! TQ Forecast feature" last week.
PP: Sure. Most performance-based (cost-per-click) ad networks have a relationship with one or more of the major search ad providers (Yahoo!, Google), either directly with a feed from that provider or through an intermediary. Some networks send all of their search traffic to these feeds (getting back a set of ads from the search provider). Other networks use one or more feeds as a "backstop" for ad categories in which they do not have direct advertiser relationships. The major search providers all employ some form of quality-based pricing which adjusts prices paid by advertisers based on that quality of traffic being delivered from the distribution network. So, for example, ad networks providing traffic to Yahoo! receive a score each week indicating the level of quality that Yahoo! feels they're delivering to advertisers. This score ranges from 0 to 10, with 10 being best quality. This "Yahoo! TQ Score" directly determines how much commission an ad network receives for traffic delivered to Yahoo!.
Unfortunately for ad networks, the formula for computing the Yahoo! TQ score is proprietary to Yahoo!. In order to prevent networks from "gaming the system," it's necessary that it be somewhat opaque, but conversion rates play a large role in the calculation. So this is the problem we've solved for ad networks. Click Forensics provides a Yahoo! TQ Forecast feature that allows ad networks to accurately predict the scores their traffic will receive from Yahoo! so that they can take action before delivering poor quality traffic. For example, an ad network may decide that a specific
publisher has traffic that will damage his Yahoo! TQ score and so he may choose to monetize that traffic differently instead of sending it to Yahoo!
What trends are you seeing in terms of traffic quality from display that you track these days?
While we have historically specialized in performance-based advertising (cost-per-click), we are seeing major opportunities in display emerging right now and it's an area we're particularly excited about. "Traffic quality" in the cost-per-click space is largely about eliminating invalid traffic and scoring valid clicks based on their propensity to convert for advertisers. In the display world "traffic quality" is a broader issue. Advertisers need to know more than just: "did that click come from a human or a bot?" They need both audience verification (how many people saw my ad?
how many times were they in the target demographic/geography?) and content verification (was the ad above or below the fold? was the associated content appropriate for my brand? were competitors displayed more prominently?).
We're seeing that ad networks want to apply the same level of performance management to their display campaigns that they've historically applied to their CPC campaigns. The end result is more transparent, measurable results for advertisers.