Google Ads best practices have evolved throughout the course of their history. Lately, things have been changing more and more quickly. It seems like every platform (including Google Ads) is competing to be the most dynamic and high-tech. Everywhere you turn, you’re being asked to trust a computer to make the right decisions for you. It might seem a little 1984-ish, but machine learning is revolutionizing our world - including our marketing.
Thankfully, machines aren’t yet ready to destroy all humans. But, are they ready to be trusted with our ad optimisation? Read on to find out!
Google Ads the Old Way
When we did ads the old way, we created 2 ads per ad group. The idea was to A/B test them against one another to see what copy “won.” In other words, which copy had the highest click-through rate. Once you’d decided a winner, you’d pause the loser and write a new ad. The process would repeat itself for as long as you were running the account, testing different calls to action, headlines, and other elements. Personally, I thought this was simple and allowed for quick testing and changes.
Google had other plans for us, however. Gone are the days of coming up with just 2 great ideas. Nowadays, Google wants you to let their AI help. That means letting go of your current A/B testing method and trusting your ads (and your performance) with machine learning. It may sound scary, but in this age of smart marketing, it may give you an advantage if you’re just willing to give it a little trust.
Why are Google Changing Best Practices?
You might be wondering why best practices have changed so drastically in such a short amount of time. The main reason? Technology. As the internet evolves, so does paid search. In order to provide a better ad experience for users in the form of more relevant ads and a higher return on investment for the advertiser, Google is using its algorithm to read the signals and optimize ad rotation in a way people cannot.
According to Business Insider, this is just a small part of Google’s attempt to, “help brands run more effective campaigns.” They’re also using machine learning to make improvements to YouTube and Google Map ads. Little by little, we are seeing AI infiltrating marketing, but that’s not a bad thing.
What’s Going on in the Background
A variety of things are happening in the background, which we can’t see. Quickly and silently, each query is analyzed. Then an advanced computer algorithm matches the query and ads based on the relevance they see. Prior performance and statistically proven better performers are also considered when an ad is being chosen to appear on the search page.
Google says this about the optimize setting, “This setting optimizes your ads for each individual auction using signals like keyword, search term, device, location, and more. Powered by Google’s machine learning technology, the "Optimize" setting prioritizes ads that are expected to perform better than other ads within an ad group. All video campaigns are automatically optimized for views.”
The New Best Practices
Okay, now that you know what’s going on and why you should let go of the days of stringent A/B testing, you should know what Google really wants. They want you to produce more ads. Instead of creating 2 ads in the traditional A/B testing method, you’ll now be generating 3-5 ads.
Prior to the use of AI, we created copy that was completely the same except for the element we were testing. Now, you can safely develop variations that have different CTAs, headline, descriptions and more. Since the machine learning is helping to choose what is best, you don’t need to worry about the multiple elements muddying the waters. In fact, you’ll be giving it a larger pool of things to test.
Even though you’re letting Google decide what works best, you should still review your ads regularly. I’d say you should analyze CTR once and month and pause the ads with the lowest CTR. Then, create several new ads to test in its place.
This best practice comes from the utilization of AI settings within the platform. It may not make sense at first, but remember your ads are no being rotated based on their individual performance alone. The optimize setting uses statistical significance to choose the best possible ad. That’s something that takes a lot of fancy Excel work on your part. Who wants to do all that when you can create the ads, select optimize, and Google will figure out what ads do the best?
Paving the Path for Responsive Search Ads
Allowing Google to optimize your ads is just the beginning. The beta Responsive Search Ads brings the many ads strategy to the next level. Responsive Search Ads enable you to generate as many as 15 headlines and 4 descriptions.
Again you’ll be asked to break with old habits and adopt the new best practice of trialing plenty of ideas and messages rather than just 2. These headlines and descriptions should not be too similar. The idea is to test out messaging ideas and give the machine plenty of things to test.
As with the optimize setting, AI is used to choose what will perform best. As it learns, it will match the headlines and descriptions that fit best with the query. In doing so, you’ll see your click through and conversion rates increase.
Google Recommendations and Automatic Implementation
The last change to Google Ads that are making waves when it comes to best practices is Google Recommendations and Automatic Implementation. Traditionally, we’ve relied on our own thoughts and feelings to optimize our accounts. Now, we have an assistant to help save us time and consider things from another point of view. Think of this as the smartest, fastest learning assistant you could ever have.
Google tells us they come up with suggestions, “To help improve performance, a combination of human review and machine learning is used to create high-quality ad suggestions. Relevant content from your account is used to create ad suggestions include your existing ads, extensions, and landing page. Google also uses additional signals such as keywords and targeting in order to optimize the ad copy.”
Now, you’ll know what you should try rather than running down the list of elements. Our new best practice is to take the recommendations and give them a human touch.
Personal Thoughts and Takeaways
As marketers, it can be hard to let go of the wheel and let someone else drive. Giving a computer control is even more stressful for us. I know that I’m still carefully testing out how these settings work and what they mean for my marketing strategies.
I suggest we use the new tools we’ve been given rather than resisting them. Gain a better understanding of the new best practices and how machine learning applies to what we do. If we learn how to use it properly, we can improve our efficiency, return on investment, and strategy. Use caution, but forge ahead with these new developments.