Another day, another AdWords update to juggle. Quite frankly, it’s all part of the routine as a digital marketer. While some updates initiate major strategic alterations such as expanded text ads or the seemingly forever ago removal of the right-hand rail, the vast majority are minor structural changes. With this in mind, AdWords announced two upcoming features designed to embrace machine learning at the ad level:

These two updates continue AdWords trend of incorporating automation. Like Smart Display Campaigns and Smart Bidding Options, ad rotation and copy creation will now integrate an element of machine learning. Each update should be relatively easy for account managers to adapt to, although there are a couple curveballs to consider when it comes to strategy and implementation.

Ad Rotation

What Changed:

Google has initiated a major change to ad rotation settings within AdWords, one they promote as a better, simpler option. Starting with the new system, there will soon only be two options for ad rotation. These options will be called “Optimize: Prefer best performing ads” and “Do not optimize: Rotate ads indefinitely”.

AdWords ad rotation settings

Prior to this change, AdWords supported four ad rotation options: optimize for clicks, optimize for conversions, rotate evenly, and rotate indefinitely. While many account managers have held tight to a traditional A/B ad testing system using rotate indefinitely, we’ve recently felt considerable pressure from our reps to evaluate our PPC ad rotation settings. For months they’ve promoted adoption of an automated ad rotation which makes this update that much less surprising. Starting towards the end of September, digital marketers will begin seeing these rotation settings available at the ad set level.

Important Considerations:

If you’d prefer to maintain traditional ad testing practices, the “Do Not Optimize” option will allow you to continue forward. Selecting “Optimize: Prefer best performing ads” will seek to deliver the most effective ad in driving clicks in each individual auction using signals like keyword, search term, device, location, and more. Essentially ads will be geared towards a delivery setting similar to the old “optimize for clicks”. That said, conversions within ad delivery will be prioritized when using smart bidding strategies such as Google’s Maximize Conversions Bidding Strategy.

One major note is that ad rotation can now be controlled at the ad group level. Unlike the campaign setting of the past, this allows marketers a chance to achieve a more nuanced and controlled schedule when rolling out new ad settings account wide.

Beyond strategic changes, AdWords continues to insist that the inclusion of at least 3 ad iterations per ad group will improve performance. Many account managers will likely need to generate additional ad variations in order to put AdWords claim to the test that “the more ads our system can choose from, the better expected ad performance”.

Perceived Impact:

Overall I don’t see this as much of a hindrance or an asset. The additional option to control at the ad group level is nice, although to be quite honest it adds a layer of granularity I’ve never felt a need for. Overall this is just a simplification of rotation options that will likely lead to a higher rate of adoption from account managers. Personally, I’ve seen mixed results when transitioning from a traditional A/B test to an automated ad rotation. In general, these transitions perform better when a smart bidding strategy has already been adopted.

One obstacle I do perceive is for advertisers that optimize copy based on conversion metrics. While conversion rate tends to be less dependent on creative, there are certain situations where ad copy improves lead quality by pre-qualifying an audience. For those not using smart bidding campaigns, ads designed to maximize efficiency and limit irrelevant traffic would likely be incorrectly rated by the new rotation settings.

Ad Suggestions

What Changed:

Remember back when Google Announced “Ads Added by AdWords”? Well, it appears we have a somewhat new and/or repackaged version of AdWords direct role in your ad copy creation process. Presumably, the first iteration of this back in January was to ensure complete coverage of Expanded Text Ads for all account, even those without active managers preparing for the transition. This update, however, seems to be more focused on the expansion of machine learning in every aspect of an account’s strategy.

In short, Google will create ad variations in response to performance data that you can choose to either use or disregard. Suggestions will initially surface on the “Opportunities” page for you to either apply or dismiss. It is important to note that ad suggestions marked as “auto-apply” will start serving after 14 days if no action is taken. Ad suggestions will appear in your account a maximum of once every 7 days and will be in ad groups set to optimize ad rotation.

Important Considerations:

Brand safety and copy management are obviously the major concern here. For highly regulated markets like travel or insurance, many advertisers have to be EXTREMELY careful both about what is said and what is implied. Since AdWords will be running completely off of performance metrics, it will be critical for account managers in these niches to maintain a vigilant eye in the approval process. For those that would prefer to maintain all manual control, you can opt out here.

From a technical standpoint, many questions remain to be answered. AdWords has yet to indicate what the actual copy or suggestions will look like. I’m curious to see how these new variations are constructed — if they will be based on keyword targeting, historical ad copy, competitor insights, etc. I’m also interested to learn what criteria will be utilized for a landing page recommendation, or if it will simply default to the current pages active within a given ad group.

Perceived Impact:

I see this as a helpful reference for times when you’ve completely hit an ad copy wall. I love the possibility of being provided a couple of data backed suggestions from which I can then create my own “spin” on them to better conform towards a preferred brand voice.

I do anticipate this will be much more limited in value for niche verticals or B2B where ad copy is as much about pre-qualifying the right clicks, as simply garnering clicks. Each AdWords update seems to be driving further towards maximizing traffic and CTR, a statistic of relative insignificance in the B2B lead gen world.


Automation and machine learning continues to be the name of the game for each new AdWords feature. The newest development in this trend extends such technology into ad rotation settings and ad copy creation. It appears as though both updates are slated to take action in late September, at which point I’m excited to see many of the remaining question marks answered. If you have any thoughts or additional questions, feel free to reach out to me on Twitter @Will_Larcom.

Cover photo by George Thomas