What Google’s Hal Varian Didn’t Teach You About Quality Score Optimization

I have lied.

Screamed.

Begged, bored and stolen, to get my Quality Score where it is today.

Here’s what I’ve learned.

“Persuade People! Don’t Chase Metrics”
(Tweet it!) 

I like to think about the Quality Score’s 1-10-value as the amount of people who are interested in your ad.

Let me explain what I mean by that.

If you have a quality score of 7 out of 10, then 70% of the people who see your ad, is really interested in it. And your goal is of course to increase the number of people who are interested in your ad copy.

I use this approach to keep reminding myself that I’m actually optimizing for people and not gaming an algorithm.

Some Quality Score Stuff I wish Hal Varian had taught me 4 years ago.

1. You can’t get a Quality Score of 10 for every keyword.

- You might be a great copywriter, an Adwords Ninja or an expert in testing, but some keywords are simply impossible to optimize for a 10 in Quality Score.

And that’s exactly how it should be.

Let me give you an example of a broad keyword and a very targeted keyword, using the first 10 results on Google.

 

“Leopard” – An example of a too broad keyword that is impossible to reach a Quality Score of 10 for.

Some Keywords are impossible to optimize

 

The word “Leopard” refers to an animal, an operating system, a brand-name by only looking at the first 10 results on Google. This makes it extremely difficult to guess what the user searching for “Leopard“, is actually looking for. Is it the animal or the OS or the brand?  That’s why your CTR and Quality Score will bleed if your pick to broad keywords!

Google Adwords and Quality Score optimization is about guessing what the user is looking for in order to meet their need with your ad.

An example of a targeted keyword that is easy to optimize

Now take a look at the SERP for the query “New Era Caps”. All of the results, as shown in the screenshot, contains information about the new era caps. I bet the searcher is looking for a new cap…

Targeted keywords = High quality score

2. Quality Score Optimization starts with keyword selection

Optimizing your Quality Score begins with keyword research and that is what’s going to shape the foundation of your Quality Score performance.

If you start off on the wrong foot by adding too broad keywords, you might never succeed because the keywords are too broad.

Soo..Start off with targeted keywords and focus on relevance. The Basics!

If you pick very targeted keywords and create a reasonable account structure are you on the right track. It’s doesn’t have to be any harder.

3. Negative Keywords doesn’t increase your Quality Score (Auch!)

When I started working with Google Adwords I thought that excluding keywords would make my quality score raise, since the “Visible Quality Score” increases.

This is the most frequent misconception about Quality Score Optimization and it’s tough to wrap your head around because excluding keywords will make your ad more targeted.

And yes, adding negative keywords to your campaign or adgroup, will make your ads more targeted, since your are excluding irrelevant audience. But excluding keywords doesn’t effect your Quality Score, because the Quality Score is only determined when the users searches for a keyword that matches your keywords exactly. The Quality Score isn’t used when your ad matches a search query that you don’t bid for.

4. Does my choice of match type impact my Quality Score?

No. Just no.

5. Is it possible to measure your Quality Score Performance over time?

Yes. Thanks to the guys over at Tenscore. You can use this tool, to track you Adwords Quality Score over time. TenScore will provide you with a historical data on you Adwords-performance and highlight your opportunities and quality score problems. I don’t know why Google hasn’t put more effort in to Quality Score Tracking, when they always preach that it’s the secret sauce of successful Google Adwords Advertising.

Now I’ll hand the mic over to Hal Varian

Hal Varian is the Chief Economist at Google, a professor at University of California and he has a great way to explain the basics of the Adwords Quality Score. If you haven’t seen this Quality Score explanation before, it’s definitely worth watching. Take it away Hal!

What do YOU know about the mysterious Quality Score and what do think is still left untold?

Thanks for your time!Share on FacebookShare on LinkedInShare on Google+Tweet about this on Twitter

7 Comments What Google’s Hal Varian Didn’t Teach You About Quality Score Optimization

  1. Josefine

    Hej frederik!
    Fedt site du har fået! :) Håber at businessen går godt!
    Jeg har lige læst din post, og det jeg reagerer over er at du ikke nævner CTR? som er den vigtigste faktor. Og at matchtypes ikke påvirker – der er jeg ikke enig. Phrase og Exact matches har naturligt en højere CTR (Hvis din Ad copy er i orden) – og derfor højere QS.
    Hvad er din mening til det? :)

    Reply
  2. Frederik Trovatten

    Hey Josefine :) – Thanks!
    (For the sake of my readers I’ll respons to it in english)

    As the headlines goes, is this a post about “What Google’s Hal Varian Didn’t Teach You About Quality Score Optimization”, which is why I don’t talk much about the basic stuff like CTR.. Even though CTR’s of course very important!

    Match Types and Quality Score
    The reason why Match Types in Adwords, doesn’t change the Quality Score of the keyword, is because the QS is only calculated when user search-query matches the keyword precicly. This means that Broad Match keywords, has a higher “Visible CTR”, but the “Real CTR”, is actually the same.

    So you if you three keywords in one ad group, like broad, phrase and exact, each of them should get the same QS. Does it make sense :)?

    Reply
  3. Josefine

    Hi Frederik
    Yes, it does make sense, in a way. But I am still not with you. :) Broad match keywords (greatest way of cashing money out of beginners) (and phrase to some extend) should over time be eliminated as far as I’m concerned. We all know that Google see “relevancy” where no one else sees it. Using broad match is a bit like a lottery, you might rarely “win” and do some great findings, but most of the time you loose. People click on ads that are no where near what they look for. (You know it, it can be great fun to look through search term reports ;))

    You say broad matches has the highest CTR ? They have the lowest, as far as I know, you must have typed wrong here? :)

    What do you mean with “QS is only calculated when user search-query matches the keyword precisely” – So what about QS if you only work with phrase and broad matches, and search queries almost never matches your keywords precisely? (Yes, big no go!)

    Let’s take an example: An ad group “Samsung Galaxy 3″ leading to a Samsung galaxy 3 product LP. We have three keywords: samsung galaxy 3, in all three match types, with same bids. When looking in interface they (seem to) have QS 7/10 (to make it easy) :)

    We all do know that the QS is calculated for each search query. And that competition/bids and your relevancy varies. Now, your three keywords, in different match types have the same bids but the CTR ad/LP relevance is obviously higher for phrase and exact. It is not that Google is so clever that it only display the ad (For Samsung Galaxy) when galaxy is included in search query. (This of course prevent clicks) All three match types trigger your ad. So, when a user search for a samsung tv, and your broad match trigger your ad, this keyword has way lower ad/LP relevance and CTR, therefore it can never be given as high QS as the keyword samsung galaxy 3 in exact match has. So you end up paying way to much. There is absolutely no point in using broad match in this case, in my opinion.

    What do you say, we have to clear this out :) Asap.

    Reply
    1. Frederik Trovatten

      Hey Josefine,

      If you have broad, phares and exact, in the same adgroup, should each keyword get the same Quality Score. Even though exact match might have a higher “Visible CTR”.

      The Quality Score is relative to other competitors bidding on the same keyword and historical data. They know the average CTR of the keyword, in each match-type and takes that into account.

      Makes sense :)?

      Reply
  4. @BulletproofPPC

    Nice one with “leopard”, Frederik!

    I was just debating with someone today why you shouldn’t give up and how you could continue trying new things to raise your quality score, but this is a very good example! That said, I still think you could get a 10 quality score for “leopard” as defiant as that sounds. Here’s why…

    The CTR is the largest portion of the quality score and it is normalized at each average position against the CTR of the other advertisers at that same position. So, if Google finds that for a specific position, your CTR is better than everyone else’s (in that same position), you’d get a higher QS. But how would you know the cause of the increase if you can’t see your competitors’ quality scores? That’s what I want to know… Anyway…

    So, I do like your article, but I would have to disagree and say that with a good negative list (starting off “right”), and by a good negative list I mean FLAT OUT AMAZING, a QS of 10 for “leopard” could happen if your CTR were vastly superior to your competitors’ CTR for that same position. I would shoot for about a 4 avg. position in all honesty, though, depending on the level of competition at the moment.

    Also, there will be cases where even then, you only get your QS to 8 or 9. Using a subdomain with the relevant text “leopard” on a highly navigable page with multiple forms of media (movies, etc.) related to the singular word “leopard”, I say you could get a QS of 10.

    Even if at that point you only had a 9 for your QS, you could use Sitelinks and more specific pages of various leopards to aid in the support of the navigational structure of the subdomain, which would be leopard.domain.com. That could get you there but it’s all about testing and learning. Plus, with the history component of the display URL, you only have so many chances to do it “right”.

    By the way, Sitelinks in the US only show above the organic search results, so the average position of 4 would have to be thrown out. :-) What about in Denmark? Just 1-3 for Sitelinks?

    My question really is not whether you could, but whether or not it would be worth it. That better be a VERY important keyword… and I’m thinking that regardless of the QS, it’ll cost a LOT of money (due to the volume of clicks at the very least).

    It’s probably more likely to get a QS of 10 than be profitable! Haha! :-)

    Great post! Keep ‘em coming. :-) Let me know if I’ve overlooked anything.

    This has been quite fun! Thank you!

    Cory Jaccino
    BulletproofSearch.com
    @BulletproofPPC

    Reply
    1. Frederik Trovatten

      Hey Cory,

      Wow, that was quite the comment :D Thanks!

      Quality Score optimization is a never ending process and you can tweak and optimize it a lot. But what you can’t do is add negative keywords to your adgroup. You can do it, but as I wrote; “You don’t a higher Quality Score, by adding negative keywords”. You’ll get a higher “Visible CTR”, but that won’t change the Quality Score, since it’s only being calculated when the users search for the exact search-query “Leopard”.

      This means that if you add “Operating system” as a negative keyword, and people search for “leopard operating system”, will you ad text not be visible, and you might get a higher CTR (and less irrelevant clicks), but your quality score won’t be effected.

      And as you say, at the end of the day, you have to calculated if a keyword is actually worth optimizing. But if it’s not worth it, I’d think you’re better of removing the keyword :).

      Reply
  5. Pingback: Improve AdWords Quality Score By Building The Proper Framework

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>