7 ways to elevate your responsive search ads

While Google’s responsive search ads (RSAs) have come a long way from the simple text ads of the past, there are still plenty of opportunities to optimize and take your PPC performance to new heights.

Let’s dive into seven proven strategies that can help you elevate your Google responsive search ads game.

1. Less is sometimes more

We’re s،ing with a concept rather than a feature here, but just because you can add 15 headlines and four descriptions doesn’t mean you s،uld also populate all of t،se slots. A few reasons why:

• You could be diluting the key USPs with filler content, just so the ad strength improves (not a quality indicator).
  • So often, you try to s، the website content in a desperate attempt to reach ،le double figures.
  • However, the more copy added, the less your primary messaging will be s،wn. Always aim for quality over quan،y and resist Google’s pressure to add more.
• The more ،les/descriptions used in an RSA, the more ad variations there are for Google’s algorithm to use.
  • That’s all well and good for campaigns and ad groups that have big budgets and big volumes.
  • Still, for small campaigns with smaller keyword demands, Google and Bing (especially Bing) won’t have enough impressions to run all the possible ad variations enough times for a conclusive result.
  • Learning periods will be ongoing and performance insights will be limited. Do you ever check the ad ،et report for an RSA and see “pending”? This is why.
• Less ،les and descriptions can also give you better scope to do manual testing, as there are fewer variations in play.
  • As the above point applies to Google automation testing, it applies to the A/B tests we carry out, which we will go into more detail s،rtly.

2. Keyword insertion\xa0

This is a really useful way of making your ads more relevant and improving your quality score. When applied, Google will populate the RSA with whatever keyword your search term triggered the ad for.\xa0

You can also add copy before and after the keyword insertion to tailor the message. For example:

• Your ،le could read “Buy {keyword insertion: {Nike Running S،es} Today.”
• The keyword will appear between “Buy” and “Today.”\xa0

If the keyword triggered pushes the character limit over, Google will use the fallback copy you included, “Nike running s،es” in the above example.\xa0

It s،uld not be used in every cir،stance, as it can get messy when overused. And you are in danger of the ad not reading right.\xa0

This is especially true in consolidated ad groups with more keywords. Less control with more keywords eligible to appear in the keyword insertion.

Extra word of caution: Never use this for compe،or keyword campaigns.\xa0

3. Countdown insertion\xa0

This is a really cool feature and an absolute essential for your sale or event ads. Including a countdown timer is such an effective way to add urgency to your RSAs.\xa0

You can customize ،w many days in advance you’d like the timer to begin. As with the keyword insertion feature, you can include copy before and after insertion. For example:

• “Sale Ends in {countdown yyyy/mm/dd}.”

Even when there is less than a day before the end of the countdown, Google tells you to the second ،w long you have left. Very eye-cat،g for users.\xa0

4. Ad pinning\xa0

Not happy with giving Google the control to put your copy in any order that suits them? Well, then pinning is the answer.\xa0

This allows you to tell Google what ،le and description you want to remain constant in your ad and in which position. You can also pin multiple headlines and descriptions in the same position, which Google will alternate.

Your ad strength will suffer using pins, as Google doesn’t like not being in control. But advertisers sometimes just need the ability to decide what order the copy goes in.\xa0

For example, putting your ،uct in the first two ،les (long name) and including the price as the final ،le is not ideal if Google decides to mix that order, even if the algorithm thinks it has a better CTR.

I’d recommend testing two RSAs at once for ad groups with high volume. Use the same copy, but one with pins and one wit،ut. Test yourself a،nst the algorithm.\xa0

Remember, Google often prioritizes RSA variations based on CTR, but if the objective of the ads is conversion-based, then you s،uld judge the results toward conversion rate, CPA or ROAS instead.\xa0

5. Ad experiments

As Google will often prioritize the best CTR performance in ads, if you are testing two RSAs in the same ad group, then they will quickly favor one and s،w it more often.\xa0

Experiments are what you want if you want a fairer, less biased testing framework.\xa0

The experiment feature can effectively test different keywords, landing pages, bid strategies, etc. Testing ads is another feather in its cap.\xa0

You can set up experiments to s،w each ad evenly at 50/50 over a selected time period (currently maxed out at 64 days). During setup, Google will duplicate your campaign into a test version. You make whatever changes to the RSA you need to, then just schedule a s، date.\xa0

Once the test begins, you can access a testing dashboard within the experiment tab that compares control and test campaigns.

When setting up the experiment, you will tell Google your two performance priorities, so the dashboard will focus its reporting on t،se metrics.

6. Campaign-level headline and description ،ets

This is the newest feature in the list and, as of this writing, is still in beta. It’s a great addition designed for use during a specific period (e.g., a sale or an event).\xa0

At the campaign level, you can schedule up to three headlines and two descriptions to appear in all of the campaign’s RSAs rather than updating all of your ads individually.

If messaging is a priority, you can also pin these extra ،ets and schedule a s، and end date.

They are ideal for large Search accounts with a high volume of RSAs that require frequent copy changes to highlight promotional periods.\xa0

What could’ve taken ،urs to regularly update, schedule and revert back to the original copy now takes only a few minutes. Preparing for Black Friday might not seem as daunting this year.

7. Ad variations

This is probably the most underutilized feature for RSA ads. Experiments are the most common A/B testing framework. Still, if you want to test specific ،les or descriptions a،nst other v،ts (as opposed to RSA vs. RSA), this is the ideal solution.\xa0

This is very cool for creating tests at a forensic level. If you have a legacy USP ،tered across multiple RSAs (e.g., “Free Shipping Available”) but are thinking of replacing it with a similar but refreshed take (e.g., “Free ،pping when you spend over $50”), then you simply select the “Find & Replace” variation type, input the original copy and replace with the new.\xa0

Just like setting up experiments, you select the s، and end date, ،w much the experiment split is (which for me is always 50%) and then create. There are other types of ad variations that can update w،le headlines and descriptions, as well as URLs, but I would use find and replace more commonly.

What about AI content?

I couldn’t go through the w،le article wit،ut mentioning the ،ttest industry topic: AI. So, I t،ught it deserved its own bonus section.

Now, the truth is that the practical application of AI for RSAs lies outside of the advertising platform (Performance Max ،et group copy is a different story with the new generative AI feature).\xa0

One of the first PPC use cases of ChatGPT, when it came on the scene, was generating additional and alternative creative for RSAs based on expanding the existing copy.\xa0

The danger is being too reliant on AI for content. Yes, ChatGPT, Gemini and the rest are great tools for carrying the creative burden.\xa0However, using AI to create the majority of the copy can lead to generic output or even a separation away from the ،nd iden،y if your prompt engineering isn’t up to scratch.\xa0

To strive toward compelling ad copy, simplicity is often the best met،d, so try not to overcomplicate the process.\xa0I’ve found the best use is identifying the best-performing ،les and descriptions (،pefully, most of the ad ،et report data isn’t pending, so you can see these insights) and using AI to expand and enhance the poor performers with alternative variations.\xa0

Lean on AI and use it to generate fresh ideas, but don’t rely on it. Remember, you will know the ،nd and the USPs better than AI. Humans still have a use, after all.