How Google Normalises Your CTR For Your Position

Discussion in 'Google AdWords' started by CustardMite, Nov 5, 2008.

  1. #1
    I am in a state of mild shock. :eek:

    I just read this article on how Google accounts for your position when calculating your Quality Score.

    http://searchengineland.com/what-google-adwords-really-changed-last-night-15365.php

    I had always assumed that Google had benchmarks for what a good, bad or average clickthrough rate should be in each position. It does now, but their past methodology was very shaky, particularly if you have a low-traffic keyword.

    Basically, they run your advert in different positions part of the time, in order to work out what would happen if they increased your Quality Score.

    There are many reasons that this concerns me - moving your advert from 8th to 7th will often make very little difference to your clickthrough rate, and there's no way that they could do this often enough to get a statistically significant conclusion.

    I'm glad that they've sorted this out now, but I'm appalled at the apparantly flawed methodology that they used to use. Is the rest of their Quality Score algorithm this dodgy?
     
    CustardMite, Nov 5, 2008 IP
  2. trishan

    trishan Peon

    Messages:
    208
    Likes Received:
    10
    Best Answers:
    0
    Trophy Points:
    0
    #2
    I thought this was already a feature of QS too (perhaps because I saw you mention it in another thread?) It's great though. The other change they made is to do with "top ad position". As PPC Hero puts it:

     
    trishan, Nov 5, 2008 IP
  3. bjewelled

    bjewelled Peon

    Messages:
    230
    Likes Received:
    11
    Best Answers:
    0
    Trophy Points:
    0
    #3
    My assumption also, and, given that such data has been available for years now, quite a surprise.


    Maybe , I am just dumb but why use artificial data from a moved ad being shown to a "very small percentage of searchers"? Why not just use the data from the positions it is showing in anyway? Maybe, it is simply to create a comparison but I would have thought ads move around so much anyway (some of mine rise and fall 20 positions in the course of a day) that it was not necessary.
     
    bjewelled, Nov 6, 2008 IP
  4. CustardMite

    CustardMite Peon

    Messages:
    1,138
    Likes Received:
    33
    Best Answers:
    0
    Trophy Points:
    0
    #4
    I just assumed that they had tables like:

    Financial Services

    Position 1: 7% = Excellent, 5% = Okay, 3% = Poor
    Position 2: 5.5% = Excellent, 3.5% = Okay, 2.5% = Poor
    Position 3: etc

    This makes much more sense that their old methodology to me - how many times do they have to run your advert in an incorrect position to find its clickthrough rate there? How many times has my advert been displaced by other adverts, for Google to find out that my advert was the best one to appear there?
     
    CustardMite, Nov 6, 2008 IP
  5. bjewelled

    bjewelled Peon

    Messages:
    230
    Likes Received:
    11
    Best Answers:
    0
    Trophy Points:
    0
    #5
    I had imagined it was more along the lines of ;

    Page 1, ave. CTR Pos 1: = 10%, Position 2 = 5%, Pos 3 = 3%, Pos 4 = 2%, rest = c.1%
    Page 2 any position; 0.5% (perhaps +0.1 or 0.2 for pos.1)
    page 3 any position; 0.25% (ditto)
    page 4 any position; 0.15% (ditto)
    rest; 0.1%

    with variations for different markets/industries. If you achieve better than the average it boosts your QS, if worse then QS drops.
     
    bjewelled, Nov 6, 2008 IP