Our site uses cookies. By continuing to browse this site you are agreeing to our use of cookies.  More information
Media
Google Adwords Launches “Ad Suggestions” Feature
Beginning April 30th, 2018, Google will be adding a feature called “ad suggestions” to certain AdWords accounts.

This feature is designed to automatically create paid search ad variations based off the account’s current ads. Once these ad suggestions have been auto-generated, advertisers will receive an alert within AdWords and via email. Advertisers will then have 14 days from the arrival of this notification to review, approve or remove the ad suggestions, after which they will be automatically enabled within the account.

What Are Ad Suggestions

Ad suggestions will be created based off content within the advertiser’s existing paid search ads and landing pages. They will have the ability to edit, pause or remove the ad suggestions at any point, though Google recommends waiting until the ads have generated enough impressions for appropriate confidence levels to be reached.

New accounts created after the official launch date will be automatically opted-in for ad suggestions, while suggested ads already active within an account will remain active even if the advertiser decides to eventually opt-out.

Why Did Google Add This New Feature

Google’s data suggests more ads per ad group can help improve performance:

“Research has shown that ad groups with 3 or more high-quality ads can get up to 5% to 15% more clicks or conversions than ad groups with only 1 ad, provided ad rotation has been optimized. The more ads you provide, the more options you’ll have to show the ideal message for each user search.”

Advertisers agree, but also understand a sweet spot exists. Ideally there are three to five ads within each ad group to ensure proper testing. Having more than that in the mix would likely produce inconclusive results or require a longer testing runway. As with most facets of paid search — it’s the quality of the ads that matters, not quantity.

Google’s thinking is an influx of ad copy variations will help push the industry to constantly A/B test. This will in turn ensure advertisers are using their best possible creative rotation, getting users to click on their ads as often as possible. The more people click on ads, the more revenue Google generates.

The Good News

With this new feature, Google will be providing the added value of machine learning-based messaging. Advertisers simply have to review the ad copy that’s created — keeping only what they want. Additionally, it isn’t necessary for the entire ad suggestion to be approved. Advertisers can pick and choose parts of the suggested ad, inserting those selected snippets into their current messaging to help their ads better resonate with their target audience.

Advertisers should view this as a value add to their current account features. Empower’s recommendation is to keep the ad suggestion feature live, but monitor it closely — ensuring no suggested ads are activated without being thoroughly reviewed and then approved.

The Watch-Outs

There may be advertisers for which this doesn’t make sense — ones with thousands of ad groups, or ones with inherently complex legal approvals. For advertisers like these, the ad approval process alone may make this new feature unusable. In such instances, our recommendation would be to opt-out. Google has noted though, they will only be providing ad suggestions for a small portion of advertisers, and those select advertisers may only see a few ad suggestions at a time per account. This mention makes it seem as though the number of ad suggestions created will be manageable, regardless of account size.

Additionally, all advertisers need to be cognizant of the messaging included in suggested ads before approving them. Most have legal/copyright guidelines to adhere to, standards and initiatives to follow, etc. that an external, automated process simply can’t account for.

Overall, the new ad suggestions feature seems like a worthwhile opt-in from a “test and learn” perspective, assuming time is set aside to properly manage the process. The big question is — can machine learning craft messaging in an automated — yet tailored — fashion that fits the individual needs and expectations of every advertiser? Automation is the obvious answer for huge, time-consuming tasks that require deep intelligence at an incredibly small scale (i.e. algorithmic bidding), but can it handle the delicate nuances normally associated with the written word? Further, can Google’s machine learning-generated ads convey that human element that still eludes most language processing applications? We’ll have to wait and see.