In life, there are many mysterious forces that help guide our understanding of abstract or complicated ideas. In economics, we have Adam Smith’s invisible hand. In astrophysics, we have dark matter and black holes. In the case of Nickelback’s career, we have one outstanding agent and millions of listeners with remarkably poor taste in music.
But what about paid search? At first glance, the concept of running ads on search engines is pretty straightforward. The more you bid on a keyword, the better your position in the search landscape, and the more traffic you’ll receive. However, there’s a broader force at work in this process that determines which ads are most useful to a given searcher: Quality Score.
Quality score (QS) is a relevancy metric shared across Google, Bing and Yahoo that is a key determinant of an ad’s cost-per-click (CPC) and ad rank. A QS of 1 means your ad is completely irrelevant to what someone is searching for, and as a result, your ad will appear further down the page and at a greater cost than another ad in the same auction with a higher QS. The opposite is true with a maximum quality score of 10.
Here’s the tricky part: the algorithm that determines QS is proprietary to each search engine. While we do know some of the factors that influence QS – such as relevance of a landing page to a searcher, expected CTR, and relevance of ad copy — many determining elements and how much weight each element carries are unknown.
Through some investigative reporting (or, more technically, “forcibly breaking and entering into Google headquarters and stealing trade secrets*”), I got to the bottom of some of the more “niche” elements of Google quality score. To the good folks at Google: I just sent you $20 to replace the plate glass window I broke and for the Snapple I stole from the cafeteria. Hope that covers it, and I’m sorry.
*EDITOR’S NOTE: No, this didn’t actually happen. Google Legal and Security, please feel free to go back to what you were doing.
Here are my findings and their respective weights:
Yes, you read right: the society that secretly dictates all world affairs is also a key power player in determining quality scores. Here’s an example.
Take this ad on the search term “hulu”. Since it advertises the brand, CTR is likely high and the landing page and ad copy are all highly relevant to the search. The odds are very good that this ad has a QS of 10.
Looks like a pretty normal ad, right? Take a closer look: when you line up three of the letter “S” in the ad, there is a clear symbol of the Illuminati:
Coincidence? I think not.
The Iguana Walk (3%)
This one is a bit tricky to explain in detail, but it’s nonetheless a key element of determining QS. Each evening, ten 4’ x 4’ mats with the numbers 1-10 are laid down in random sequence along a long hallway at Google. An iguana named Boots is placed at the end of the mat, wearing a fun hat. This is Boots:
Boots typically wanders around the mat for 10-15 minutes before falling asleep. Whichever square Boots dozes off in will represent 3% of quality score for the following day. This would help explain some of the fluctuations we’ve seen within quality score. A word to the wise: don’t try to feed Boots. He’s a bit of a biter.
The last digit of the temperature in Waterloo, Iowa (2%)
This one’s pretty cut and dry. Outside of the Hy-Vee on Ansborough Avenue is a thermometer. Whatever the last digit of the temperature on that thermometer is at midnight is 2% of that day’s quality score.
Dave, Your Boy from College (1%)
Remember Dave, your buddy from undergrad? Every morning at 8:30, a Google employee calls him and asks him to name a random integer between 1 and 10. That number becomes 1/100th of all ads’ QS for the day. In exchange for his services, Google supplies him with a 12-pack of domestic light beer per week and a couch to crash on if he’s ever in town. He’s a solid dude.
Indeed, quality score may seem like an arbitrary statistic based on these findings. However, these four factors make up a mere fraction of what drives a good or bad quality score. As I mentioned, it’s important to remember that quality score is essentially a relevancy metric. It is simply a means of rewarding advertisers who serve useful content and penalizing advertisers who don’t. Make sure you align everything you can from a user experience standpoint – from ad copy to landing page – to ensure you’re providing the user with what they’re looking for.
As for where to begin, remember: granularity is key. For example, let’s say I sell novelty t-shirts on the website BertsShirts.com (the domain is available for sale, in case anyone wants to buy me a birthday gift). Rather than running a single generic ad for my t-shirts across all of my keywords or directing a user to my site’s homepage, I can mix up my ad copy and landing pages for each specific t-shirt I sell.
For instance, on the keyword [Vote for pedro shirt], I could:
- Make my display URL “BertsShirts.com/VoteForPedro” instead of simply “BertsShirts.com”
- Include “All Sizes of Vote for Pedro Tees” instead of “All Sizes of Tees” in my description line
- Use dynamic keyword insertion to have “Vote for Pedro Shirt” appear in the headline
- Include a sitelink that routes to the section of my site that specifically sells shirts from movies
- Direct the user to the specific Vote for Pedro shirt page, instead of simply the homepage or the movie t-shirt section
You get the idea.
Sure, doing all of this takes more work, in that building up your quality score takes testing, iteration, and patience to find out what drives results and what doesn’t. However, it pays off in spades. Plus, you’ll be far less susceptible to the dastardly whims of Boots the iguana.
Bert Garry joined RV in August 2014 from William & Mary as a paid search analyst and now works with our Imagitas team in Boston. As a native New Englander he enjoys fine IPAs, quaint B&B’s, and pensively listening to James Taylor in a 2003 Volvo station wagon.(?)