Responsive Search Ads (RSAs) are now the default ad format in Google Ads, and whether you actively chose them or not, you’re almost certainly using them.
On the surface, they seem simple enough. You provide a handful of headlines and descriptions, and Google mixes and matches them to create ads. Easy.
But under the hood, there’s quite a bit more going on. Understanding how RSAs actually operate can make a significant difference to how you build, test, and optimise your campaigns.
So let’s take a closer look at what’s really happening.
What Are Responsive Search Ads, Really?
A Responsive Search Ad allows you to input multiple headlines (up to 15) and multiple descriptions (up to 4). Google then dynamically combines these assets to create different ad variations.
Instead of writing a single fixed ad, you’re effectively giving Google a toolkit of messaging options.
Each time your ad is eligible to show, Google evaluates the user’s search query, device, location, and a wide range of contextual signals. Based on this, it selects the combination of headlines and descriptions most likely to perform well for that specific auction.
In other words, RSAs are not static ads. They’re adaptive, data-driven combinations built in real time.
The Role of Machine Learning
At the core of RSAs is machine learning. Google continuously tests different combinations of your assets and tracks how they perform across various contexts.
Over time, the system learns which combinations are more likely to result in clicks or conversions. These combinations are then prioritised in future auctions.
It’s not just looking at individual headlines or descriptions in isolation. It’s learning patterns. For example, a particular headline might perform exceptionally well when paired with a certain description and shown on mobile devices for high-intent queries.
This constant testing and refinement is what gives RSAs their flexibility and potential performance advantage.
How Asset Combinations Are Selected
When a search occurs, Google doesn’t randomly assemble your ad. It evaluates available assets based on predicted performance.
Typically, an RSA will show up to three headlines and two descriptions, though this can vary depending on available space and format constraints.
The system considers factors such as:
- The relevance of each headline to the search query
- Historical performance of asset combinations
- The user’s context (device, location, time of day, etc.)
- Expected click-through rate and conversion likelihood
From this, it builds the “best” possible ad for that moment.
This means that not every headline you write will show equally. Some will naturally become more dominant over time.
Why Some Assets Rarely Appear
One of the more common frustrations with RSAs is that certain headlines or descriptions barely show, if at all.
This usually happens because the system has determined that other assets are more likely to perform better. If a headline consistently underperforms, it gets deprioritised.
However, this doesn’t always mean the asset is “bad”. It may simply be less relevant for the majority of queries your ad is appearing for.
This is why reviewing asset performance data is important. It helps you understand which messages are resonating and which might need refining or replacing.
The Impact of Ad Rank and Relevance
RSAs don’t exist in isolation. They still operate within the broader Google Ads auction.
Your ad’s position is influenced by Ad Rank, which is determined by factors such as your bid, expected click-through rate, ad relevance, and landing page experience.
Because RSAs can adapt to different queries, they often improve relevance compared to static ads. This can lead to higher expected click-through rates and, in turn, better Ad Rank.
In simple terms, the more relevant your RSA is to a user’s search, the more competitive it becomes in the auction.
Understanding “Ad Strength”
Google provides an “Ad Strength” indicator when you build RSAs, rating them from Poor to Excellent.
This score is based on factors such as the number of assets you’ve provided, their diversity, and how well they align with best practices.
While it’s tempting to chase an “Excellent” rating, it’s important to remember that Ad Strength is a guideline, not a guarantee of performance.
A technically “Excellent” ad isn’t necessarily going to outperform a “Good” one if the messaging isn’t aligned with your audience or offer.
Use Ad Strength as a prompt for improvement, not a strict target.
The Role of Pinning (and When to Use It)
Pinning allows you to fix certain headlines or descriptions into specific positions, such as ensuring your brand name always appears in Headline 1.
While this gives you more control, it also limits Google’s ability to test combinations.
The more you pin, the less flexible your RSA becomes, and the less data the system has to optimise performance.
That said, pinning can be useful in certain situations, such as for compliance requirements or when specific messaging must always appear.
The key is to use it sparingly and strategically.
How RSAs Learn Over Time
RSAs improve as they gather data. In the early stages, performance can fluctuate as Google tests different combinations.
During this learning phase, it’s important to avoid making constant changes. Frequent edits reset the learning process and prevent the system from gathering meaningful insights.
Once enough data is collected, performance typically stabilises, and the system becomes more effective at prioritising high-performing combinations.
Patience here is often rewarded.
Common Mistakes That Limit RSA Performance
One of the biggest mistakes is providing too few or overly similar assets. If your headlines all say roughly the same thing, you’re not giving Google much to work with.
Another issue is writing generic copy that tries to appeal to everyone. This reduces relevance and weakens performance across different queries.
Overuse of pinning is also a common problem, as it restricts the system’s ability to optimise.
Finally, constantly editing ads can prevent RSAs from fully learning and improving over time.
In Summary
Responsive Search Ads are powerful because they combine automation with flexibility. They allow you to cover a wide range of search queries and user contexts without manually creating dozens of ad variations.
But they’re not a “set and forget” solution.
To get the most out of RSAs, you need to provide strong, varied assets, understand how the system makes decisions, and give it the space to learn.
When you do that, RSAs stop feeling like a black box and start behaving like what they’re meant to be: a highly adaptable tool for delivering the right message at the right moment.