Google Ads launched ad suggestions last year. By default, you are enabled into this program. What this program does is:
- Create ads for you
- Gives you 14 days to approve or reject the ad
- If there is no rejection, then the ad is automatically added live in your account
You can easily turn off this feature within your account settings.
We wanted to examine these ads in 3 ways to really understand Google’s ad creation automation:
- What is their performance?
- How common are these ads?
- Where is the ad text originating from?
Performance of Ad Suggestions
This is how we analyzed the performance:
- Took data only from ad groups that had both Google created and advertiser created ads
- Examined the CTR, conversion rates, and CPAs for advertiser vs Google created ads for each ad group in the account
- Repeated for multiple accounts
You can easily do this for your own accounts. You can see what ads are auto-applied in your Google Ads account.
If you download the data, you’ll have a column that lists Yes or No for the auto-applied ad suggestion status.
Here’s the compiled data for one account (click the image to see a larger screenshot):
Here’s the overall trend that we saw for every account except one:
- Google’s overall CTR is lower
- They will beat the advertiser’s CTR quite often
- Google’s overall conversion rate is lower
- They rarely beat an advertiser in conversion rate
- Google’s overall CPA is higher
- They rarely beat an advertiser is CPA
In one account, Google hands down beat the advertiser’s CTR. However, it was a B2B account and without the pre-qualification that it was B2B focused ad, the conversion rate was incredibly worse than the advertiser’s and the CPA was phenomenally higher than the advertiser’s typical MQL (marketing qualified lead). A lot of B2B advertisers really think about how to get the correct click in an ocean of B2C clicks. Therefore, the ad prequalification of the searcher being a B2B searcher is very important to many B2B accounts.
How common are these ads?
When this system was announced, we had visions of large numbers of ads being created and Google taking over the ad creation for the advertiser. If you think about it, it takes some thought for a human to come up with a nice creative idea, make sure the character length works, and that the overall ad fits their business while trying to reach the proper prospects.
A computer could just interchange adverbs, adjectives, pull data from the website (like DSAs) and overwhelm the advertiser with ads.
It turns out that’s not the case. In the vast majority of accounts we looked at, there were less than 50 auto-applied ad suggestions. The size of the account did not make any difference on how many ads were created (that was surprising). Some smaller accounts saw the Google suggestions in over 1% of their total ads; but in most cases, the numbers were tiny:
Where is the ad text originating?
With visions of computers looking at thesauruses, dictionaries, website analysis, or just thinking for themselves, we really wanted to know where the ad lines were coming from. Google has access to so much data, we thought they’d examine your sitelinks, callouts, and other ad extensions, try some different CTAs, and use some adjectives pulled from the thesaurus.
It turns out, Google is not that creative. Or, Google is too worried about writing a line that makes you upset and doesn’t fit your business. As these are on by default, creating some benefit that doesn’t exist can cause some legal issues.
All the text is pulled from the landing page of the site for that ad. It might be your headline, company name, a line in the middle of a paragraph, or somewhere else. After painstakingly looking at more than 100 auto-applied ads, all the text was just pulled from the site. Google might add a period at the end of a line, but that’s the only difference we could find from examining the deltas of the ads compared to the on page text.
Your ads are the connection point between someone looking for a service or product your business provides and convincing that user to click your ad and become your customer. This is your main bridge between search results and your website.
As this is how you get the proper traffic from search results, you should control this message.
Now, Google did beat the advertiser on occasion in terms of CTR, and this data is being pulled from text that is already on your site. So, for most advertisers, they should just follow this approach:
- Turn off the auto-applied status
- This will ensure the ads are not just made live on their own
- When you receive a suggestion, examine it. See if it’s something new.
- If you like the suggestion, then tweak any other parts of the copy to ensure it’s still appealing to your customer base (consider your conversion rate)
- Let the human edited, machine suggested, ad go live so you can test the new message
Automation does best when it’s a combination of humans and computers working together. As these are free suggestions, they are worth examining. However, as your ads are your bridge – you should ultimately know what the message is and ensure it’s the right one for your company to gain new customers.
This is a good way of finding new ads to properly test, just ensure you are controlling the final messaging.