While Google can determine keyword relevance to a certain degree, it takes finesse to really understand the elements involved.

Google quality score impacts your paid search return-on-ad-spend. Learn the incredibly delicate feat of balancing your quality score -- that slippery number that is determined based on your ad text, keyword relevance to query and clickthrough rate -- and you'll be miles ahead of the pack because you'll be paying less for keywords and appearing higher in the listing, while (hopefully) not losing anything in terms of conversions.

But what the heck is quality score?

Google's quality score is assigned to every keyword and determines your minimum bid for that keyword and where your ad appears in contextual listings.

According to the Adwords website, Google believes that "high-quality ads attract more clicks, encourage user trust and result in better long-term performance. To encourage relevant and successful ads within AdWords, our system defines a quality score to set your keyword status, minimum CPC bid and ad rank for the ad auction."

A keyword's quality score is determined by "your keyword's clickthrough rate (CTR) on Google, and the relevance of your ad text, keyword and landing page."

It's a pretty basic concept with tremendously complicated implications and, unfortunately, a fairly opaque algorithm. That's because determining the relevance between one thing and another is as much an art as it is a science.

Where's the relevance?
Is there relevance between a dog and a cat? Sure. They're both animals. They're both pets. They both have fur and paws, but what about between a dog and, say, a slipper? You might think at first that there's no relevance between the two, but what's the context? If your pet dog chews up your favorite slipper, then there's an obvious relationship.

My point is that while the science at Google's command is enormous and can determine relevance to a certain degree, it takes finesse and sometimes even a human element to really understand the extent of interactions between things.

In the past, there were human editors at Google who looked at landing pages, and it's possible there still are. As search engine expert Jonathan Mendez, author of the blog optimizeandprohesize.com, says, “We know Google's AdsBot plays a crucial role in determining quality score, but we don't know the extent of human review."

And even if there are human beings looking at relevance between keyword, ad text and landing page, could they possibly look at the majority of landing pages out there? Certainly not.

So, do quality scores really reflect relevance?

To some degree, the answer is yes. Quality scores do reflect relevance, particularly when it comes to relevance between keywords and ads.

But at OTTO Digital, we ran a study around quality scores to see if Google really takes into account the relevance of the keyword and ad text (as indicated by clickthrough rate) as well as the relevance between the ad and landing page (as indicated by conversion).

We found that the ads that get the most clicks are not necessarily the ads that convert the best.

Old news, right? Search marketers know that sometimes ads with a lower clickthrough can result in a higher ROI if ultimate conversions are better. However, in our test, we found that the words that had fewer clickthroughs were served less often, even though they were converting at much higher levels. The problem, of course, is that Google does not know the conversion rate on many sites, so it cannot factor ultimate conversion into its relevance algorithms. And, since Google gets paid on clicks, an ad that receives fewer clicks but more conversion is not in Google's best interest.

What does this mean to marketers?
It means that if you fiddle with your keywords, ad text and/or landing pages in order to increase your quality scores, you may not be acting in your own best interest because your quality scores largely only reflect clickthrough rate and not conversions.

On the other hand, if you get a higher ROI with a keyword that gets a lower clickthrough rate and you don't fiddle around with things to improve quality scores, your ad is likely being shown less often than a word with a higher clickthrough and lower conversion. The issue boils down to volume of conversion versus ROI.  

What can I do about it?

The best thing for a marketer to do is to test. You may already run tests to find the best balance between clickthrough and conversion. If not, the first step is to test to optimize your conversion per ad unit. Next, when you run clickthrough and conversion optimization tests, keep an eye on your quality scores as well as on how often the various ads are being served.

If a highly converting keyword is given a lower quality score and your ad is served less often -- and you're paying more for it -- you might need to play with the ad text to see if you can improve clickthrough rate, which should result in a higher quality score. But watch your ROI. In the end, there is no way around taking the time to understand the balance between your volume goals, your CPA/CPL goal, and how your clickthrough rate and quality score impact this equation so you can optimize accordingly.

Jamie Roche is president at Offermatica. Read full bio.