• Stephen

    A wordstream good wordstream article wordstream there! I wordstream hope wordstream for wordstream more!

  • Larry Kim

    dude be cool. would you prefer an article that makes grand conclusions and assumptions based on no data at all? I don’t think so. I only have data on the accounts i work on so what the heck are you expecting in a case study, duh.

  • Chande

    @larry_kim:disqus this is nice, but doesn’t tell much about exact niche for this advertiser. I have seen couple of accounts in non-saturated niches where this is easily reachable by optimizing for no1 result (CTR per account usually around 20-30%).

    Things get more tricky in saturated niches or B2B where long tail wins – so if you optimize only for QS or CTR, you would be nowhere. In one particular case, avg position was 1.1, CTR was around 2% (almost everything was long tail) and reaching QS above 3 or 4 was almost impossible – simply not enough volume for given keywords to get enough data to optimize CTRs for specific ads (talking about 1000s of long tail keywords, simply not worth the time & effort).

    So this whole QS stuff is just a bluff from Google to pursue some vague metric, not optimize for really important stuff like LTV :) my 5 cents.

  • Larry Kim

    would love to audit this account

  • Chande

    Sure – send me PM on Twitter and I’ll send some data back.

  • Stu

    Larry mentions wordstream 3x in his article. You mention wordstream 6x in your 2 sentence comment and I mention wordstream 3x in my 4 sentence comment. In the end, Larry is pumping out a RIDICULOUS amount of content because he knows what the bots want, and people like it, and share it. Larry, I am jealous of what you are able to get published and the mentions you are able to receive.



    You continue to be my (and my fellow grad students) go-to for SEM knowledge and insight. I appreciate how much time you invest in your articles and how much inside data you share. I understand that by doing so you are promoting Wordstream. However, I read a lot of your content and never get the feeling that your articles are purely advertisements for http://www.Wordstream.com. I believe you are a sincere educator/author. Never get discouraged by other “experts.” You are appreciated.

    http://www.Wordstream.com (one last plug for you) ;)


  • http://www.toptiertools.com/ Frederick Vallaeys

    Hi Larry,

    The reason Google says they don’t use negative keywords for QS is that they use the position-normalized **exact match** CTR of a keyword on Google.com to calculate QS. Negative keywords only eliminate impressions when there is a broad or phrase match so that’s why it’s not a big factor of Quality Score.

    Thanks for sharing this case study!

  • Larry Kim

    thanks fred. i’ve read and linked to the standard google explanation in the article. Still I’ve found that the higher the CTR, the better the QS (seemingly regardless of match type). it’s almost as if there’s some other undisclosed element – like average account CTR from all search campaigns – that is also being used in the calculation? How else to explain how keywords with 0% CTR are getting QS of 10/10 in my data?

  • Larry Kim

    heh thanks anna. u crack me up! :)

  • Aditya Jagtap

    Hi Larry,

    While the post is informative, I felt you have tried to promote wordstream tool.


  • Guest

    Nice article, but is 8.8 really that great? I manage accounts that have 9.3, 9.9 and a 9.2 according to the tool I use.

  • Mike Shostack

    Interesting data, but the issue comes with the scale. By your own admission, the advertiser is now spending about $500/mo. I could have a phenomenal QS as well if I cut my spend to that level and only put it towards my highest QS/CTR terms and ads.

    Unfortunately, for those of us managing larger spends, optimizing solely towards metrics that drive QS is simply not feasible most of the time as we need to balance things that might improve CTR with conversion metrics and volume.

    That said, it is interesting to see how high they could get their QS at that scale, although I’d be curious about the mix of brand/non-brand terms in there.