#1
  1. No Profile Picture
    Registered User
    SEO Chat Explorer (0 - 99 posts)

    Join Date
    May 2008
    Posts
    14
    Rep Power
    0

    How to speed up AB testing of adword ads?


    Hi, I'm trying to AB test my adword ads and finding it a frustrating and time consuming experience as I've alot of ads. Using adwords to optimise them for me isn't ideal as I want to basically find ads that generate the most sales within a maximum cost of acquistion, and google optimises on CTR. Similarly other tools out there.

    Any advice please??
  2. #2
  3. No Profile Picture
    Contributing User
    SEO Chat Skiller (1500 - 1999 posts)

    Join Date
    Jan 2008
    Posts
    1,834
    Rep Power
    425
    When you have a minimum budget of $5 don't split test 50 ads per ad group otherwise you'll wait a year for decent data. The trick is the reduce the number of ads dependent on your spend. Also set the settings to rotate ads.

    Comments on this post

    • whiterabbit agrees
  4. #3
  5. Contributing User
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    Jun 2012
    Posts
    285
    Rep Power
    235
    regarding your comment on it being time consuming...in the ad creation process and this may be a little basic for you but if you're not already using it, adwords editor helps you make mass changes across adgroups, which takes some of the manual-ness out of it....
    "...using this: sages act out of constructs but they do not rely on them"
  6. #4
  7. Super Moderator
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    Dec 2013
    Posts
    498
    Rep Power
    234
    Test hypotheses against each other. Don't test "variations." I've spent a lot of time testing small differences just for the sake of split testing and learning what version of an ad was better, but in practice these got relatively small CTR or other improvements.

    A better approach is to
    1) See what's working for competitors. Draw inspiration
    2) Come up with 2-3 hypotheses of what might attract a click.
    3) Test ads based on those hypotheses against each other.

    The greater the differences, the better. E.g. What's more appealing, a discount or social proof? "10% off" vs "loved by thousands"

    On the greyhat side, set up multiple sites with multiple accounts and buy more than one spot in the SERPs.

    Comments on this post

    • realityhack agrees : I wouldn't go to that grey hat place. Seems to me if your doing that your site probably sucks and shouldn't be listed at all.
  8. #5
  9. No Profile Picture
    Contributing User
    SEO Chat Good Citizen (1000 - 1499 posts)

    Join Date
    Apr 2009
    Posts
    1,317
    Rep Power
    581
    I would recommend starting with one or two of your top volume decently converting ad-groups.
    Looking at those start to think about wither you have enough traffic to get statistically significant results. Someone pointed me at https://vwo.com/ab-split-test-duration/ I do not know what math they are using but I think it might be a little on the pessimistic side.
    Use only 1 test and the control (2 ads). Set Google to split the traffic evenly.
    It will still take time. You will need to commit to running the test for X period. Do not test every day and assume you have the answer when something shows as significant.

    You might also try something dramatically different the first time through instead of tweaking the headline change the call to action, or even the entire ad instead of one line to see what general direction might be a good one to go in.

    Statistical testing can be amazing but you MUST have the traffic volume for it.

    Comments on this post

    • fathom agrees

Similar Threads

  1. How To Get A Best Crt In Adword
    By sdas2829 in forum Other PPC
    Replies: 4
    Last Post: Oct 23rd, 2012, 02:13 AM
  2. Adword Help
    By pureskill13 in forum Google Adwords
    Replies: 5
    Last Post: May 25th, 2010, 12:05 PM
  3. A/B Testing and multivariate Testing
    By apg1985 in forum Google Optimization
    Replies: 5
    Last Post: May 9th, 2010, 05:45 PM
  4. To Adword or not to Adword?
    By emmaonline in forum Google Adwords
    Replies: 4
    Last Post: Aug 22nd, 2006, 10:47 AM
  5. Email Newsletters: Testing, Testing, 1, 2, 3
    By SEO Chat in forum SEO Chat Articles
    Replies: 0
    Last Post: Oct 4th, 2004, 08:00 AM

IMN logo majestic logo threadwatch logo seochat tools logo