Usually, Google ads representatives aren’t very knowledgeable and are prone to give bad advice to unsuspecting business owners.
There are rare exceptions. Extremely rare exceptions.
This is one of them.
The following is from a conversation between a Google Ads representative and Robert of AdWordsAgentur, who was kind enough to let me share it.
I hope you find it insightful.
Quality Score (QS) is based on a great many signals, but in terms of the key players, here is a breakdown:
Average performance across all accounts.
We look at the average CTR of the keyword in question across all accounts. CTR is used to determine the “relevance” that you see in the Ads Diagnostic Tool. Why? The logic is that if people have clicked ads for this keyword, they must find it relevant, otherwise they wouldn’t click. If a KW has a bad overall CTR, it will be hard for the client to get a good QS number unless they have extremely good individual performance. Read on for more about that.
The specific account performance.
We then compare the account performance to the average performance. If the account performs no better than average, then it stands to reason that the QS will remain the same. If it begins to perform ahead of the grade curve, then the Quality Score can rise. Remember, I’m talking about CTR when I talk about performance.
On the subject of account performance, Quality Score is like a web. The individual performance of every element of an account will affect every other element. Ad texts, keywords, ad groups, campaigns and domains all affect each other. The longer there is bad performance of any of those elements, the more negatively affected an account can become. Basically, bad performance “metastasizes” across accounts over time. That’s why we say to delete badly performing KWs.
Sometimes, opening a new account can help if an account was just really badly built for a long time. However, remember that we store information about domain level performance, so if a domain does badly for a long time, all new accounts will be subject to that remembered history. One other point is that keywords and ad texts build up a performance history over time. If you alter an ad text, the system actually creates a new one, so that “QS pairing” is broken, and a little of the Quality Score will be reset.
We only use the exact match data on Google Search to determine the QS that you see in Google Ads. So, any clicks that happen on phrase or broad match do not factor in. I don’t mean that phrase or broad match keywords can never have a good performance history, just that if you have the broad match keywords ‘red shoes’ only when that keywords matches exactly to the search query ‘red shoes’ will the click be used to help determine the Quality Score that you see externally in the keyword tab. You can see the exact match performance data of all KWs in an account by going to the KWs tab, and pressing “Segment” > “See search terms match type”. Try it out!
We only use exact match performance history accrued on Google Search itself. That is, Google.com/de/fr/co.uk etc etc. The Quality Scores for search partners and display network placements are calculated separately and individually for each and every search partner and display network placement. Therefore, it’s possible to have amazing performance history on the search partner network without it positively affecting your QS on Google Search. Remember that the Ads Diagnostic Tool will ONLY show you the QS on Google Search.
Quantity of Performance Data.
It’s common for advertisers to write in and say, “I have a CTR of 33% on my keyword! Why do I have a QS of 4?”. Then, when you look at the keyword, you see that that 33% is over 10 impressions. You should always remember that we need a statistically significant set of data to calculate Quality Score. Keep in mind we can’t define that “statistically significant set of data” in terms of absolute numbers.
I hope this was useful.
I decided to share it because I believe it’s accurate, in a very practical way. There’s what Google says. There’s what makes sense in theory. Then there’s what works in practice.
I’ve listed my takeaways in the comments below, I’m inviting you to share yours too, as well as your experience with optimizing Quality Score, so we can learn a bit more one step at a time.