What’s in a Name?

Not long after we hit the send button last week, Google announced a major change to the way their Quality Score mechanism works. Quality Score changes, both from the actual mechanical modifications but more often the philosophical rational behind the change, have occupied more than its fair share of articles in DMConfidential. Google decided to publicly announce this change, something they don’t always do, especially when the change targets specific advertisers and advertising segments whose presence they’d like to impact. Those tactical changes we find out about only anecdotally as the flurry of instant messages and emails takes place when affiliates, advertisers, and their networks all try to triangulate volume fluctuations. The larger strategic ones, or simply those they want the world to know about, occur without warning as a post on their AdWords Blog. This was the case last Thursday, in a post titled simply enough, "Quality Score improvements." Calling the changes "improvements" is semantically debatable, as Google introduces changes that improve their business, but they don’t necessarily make life easier for their advertisers. Speaking of semantics, one of the three announced components of the change deals with terminology. The notion of minimum bid disappears, replaced with the term "First page bid," their estimate of what it will take for an ad to appear among the first page of results. The big change, though involves more than semantics, and whether it will send advertisers scrambling, we shall see. Instead of measuring Quality Score when a material change to a text ad occurs, such as creating a new ad or modifying the URL that a particular ad goes to, the system will now determine quality score as an actual query takes place.

Different people have varying takes on what the switch to real time will mean overall. The Google party line states that the changes mean "that Google users are apt to see better ads while you, as an advertiser, should receive leads which are more highly qualified," and they provide the following example to explain how per-Query quality score works:

Perhaps we have verbal lactose intolerance, because that example doesn’t cause any sort of Eureka moment in helping us understand the real implications of the shift to real-time quality score, among other things. Putting on our what behavior does Google want to see happen less hat, we suspect this release attempts to discourage any gaming of the system. While unknown in scope, some advertisers would create a page that has poor performance from a conversion standpoint but high contextual relevance. They would suffer an initial loss, spending money for traffic that didn’t convert, so that they could swap out the content of the page down the road to higher converting text once they’ve calibrated a certain amount of traffic for a particular price. Such behavior goes against the intent of quality score, and because they can’t employ human editors for every ad prior and during launch, Google continues to attempt to refine its technology so that it can automate how their prototypical human reviewer would rate things.

Others, especially search centric blogs /publications, have made some insightful suggestions on how these updates will impact marketers. For us, the most intriguing point comes not with trying to predict the actual impacts, e.g., should advertisers shift from broad match to trying more exact match queries, but in interpreting a concept they reiterate, not just in this, but many previous iterations, the idea of "performing better." In the post describing real-time for instance, they write, "Nancy’s ads perform better on the keyword ‘milk’…", and "Her ads perform better on the query ‘milk delivery’ than on ‘milk’…" In the performance world, we don’t have any ambiguity around the notion of perform better. It’s all about conversion. Take an example of a normal arbitrage scenario, be that in search where you pay for clicks without any guarantee of a result or on a per thousand impression with the same risk. Perform better means the spot converts better and produces a better yield. In a network setting, for example, when running an offer with publishers, perform better implies better conversions, that a certain placement makes more money than another. In other words, perform better has a purely financial implication. Only with Google, one would wonder if this is the case, as unlike normal companies, they don’t speak of perform better and profit.

With the quality score update, we have yet to read anything that disagrees with the idea that perform better has the same meaning for Google as it does for the rest of us, i.e. that change also means Google will make more money. The problem with Google, though, is that they consistently take a better than tough approach to the changes they make and this notion of perform better. In their defense, they come the closest to making sure perform better doesn’t come at the user’s expense, but they never speak in terms of the value to themselves; they speak about the impact to the user, but generally perform better for Google is a term reserved for assuaging their advertiser base that changes really are good for them – like a doctor telling a skeptical youth that this painful shot will actually help them even though all signs suggest otherwise. It doesn’t change the reality of the word though, that perform better will make them more money. Think of our space where to perform better means increasing conversion rates and/or extracting a higher payout, and you can now read between the lines of what will really happen with this iteration. Instead of saying that, Google tells us how they plan on accomplishing it. More than anything, quality score changes are a lesson in managing change and making more money without it seeming like that’s what you are doing. It’s about how you can work the notion of quality into financially motivated changes, it doesn’t seem like you are trying to make more money, especially, if like Google, you can talk about how the end user experience won’t decrease.