Google AdWords Changes Rotation Options

8th September 2017

Google has been introducing many new options to its AdWords service lately, so it is a little surprising to see a number of options related to ad rotation being removed from the backend of the popular PPC service.

Currently, Google offers a number of options for ad rotation, namely:

  • Optimise for clicks (A setting that shows ads expected to provide more clicks.)
  • Optimise for conversions (Show ads that Google analytics conversion tracking has shown are most likely to lead to sales and enquiries.)
  • Rotate evenly (This is used mostly if you are using a third party tracking system, thus allowing you to make changes at a later time.)
  • Rotate indefinitely (This ignores popularity entirely and does not prompt you to optimise.)

 

New Settings for Google AdWords

The new setting will now be “optimise or do not optimise”. The automatic optimisation setting will favour the best performing ads automatically, whilst the other option will simply rotate the ads endlessly. Given this shift, it is now easier to see why Google has begun to offer to automatically create a second advert for PPC users by simply reversing the two headline fields of the ad. This makes it easier for most advertisers to have a test option and gives the new PPC system a baseline for its optimisations.

 

PPC Users Will Need to Upgrade

Users of older PPC ads, particularly ads that pre-date the switch to multiple headlines, may wish to update their advertising to take advantage of the new services that have been made available over the last year or so. This is particularly true for companies who are looking to make use of the Google provided A/B split testing facility that has been introduced as part of the headline rotation system.

The new options also fit into the broader picture of machine learning’s increasing pervasiveness and influence on AdWords systems as a whole. By introducing this into a paid service, Google is clearly beginning to feel confident in the quality of its machine learning.