Google Ads responsive search ads: what to give them to test

A Google Ads responsive search ad is not a form you fill in once and for all. It is a set of headlines and descriptions that Google tests continuously to put together the most relevant combination for each search. What makes the difference is what you give it to test, not the number of boxes ticked.

What really matters in a Google Ads responsive search ad

Google says so in its official best practices for responsive search ads: the quality of an ad shows first in its Ad Strength. This rating scores each ad on four levels, from lowest to highest, and measures the relevance, quality and diversity of the headlines and descriptions provided.

Each ad accepts up to fifteen headlines and four descriptions, as the page dedicated to responsive search ads specifies. The classic trap is to stop at the three required headlines, when the algorithm needs material to put together different combinations depending on the user’s exact search.

According to the averages Google publishes on Ad Strength, moving an ad from the lowest level to the highest brings 15% more clicks and conversions on average. Adding a second ad to an ad group that only has one brings 6.6% more conversions at a comparable cost per conversion, and a third, 3.7% more. These are market averages, not a promise for your account.

Who it is for, and who it is not for

This is for you if one of these sentences describes your account:

• a Search campaign runs with a single ad group and a single active ad;
• only the three required headlines are filled in, with no variants tested;
• nobody has looked at the ads’ Ad Strength since they were created;
• the account targets several search intents with the same ad for all of them.

It is not really for you if each ad group already contains two or three responsive search ads rated Good or Excellent, with headlines that speak to different searches. In that case the job is regular maintenance, not a complete overhaul. If you also run Performance Max, the same habits of variety apply to asset groups: I cover this in this guide before you switch on Performance Max.

If this headline work is just one example among a bigger pile of repetitive tasks in your team, I can help you put a figure on it: get in touch.

What it changes for your budget

Adding headlines and descriptions costs nothing extra on the bid: the daily budget stays the one you set. What changes is the efficiency with which that budget is spent, because a better-rated ad wins the relevant auctions more often for a comparable cost per click.

Conversely, a single ad with few headlines limits the number of searches it can appear for. The budget is spent anyway, on a narrower range of searches. So the lever is not to raise the budget first, it is to give the ad what it needs to be chosen more often.

Two accounts with the same daily budget can show very different results if one runs with a single outdated ad and the other with several ads well stocked with headlines. It is not the bid that changes, it is the likelihood that the algorithm finds, in what it has been given, the combination that matches the exact search of the person at the other end.

The checklist

1. Open each ad group and count the active responsive search ads: aim for at least two, rated Good or Excellent.
2. List the headlines already in place, then add more until you have twelve to fifteen distinct headlines, varying length and angle.
3. Add one headline that repeats the searched keyword as is, one that talks about the benefit, and one that gives a specific action.
4. Check the four descriptions: each one must be able to stand on its own, without depending on a particular headline.
5. Pin only the text that must appear in every combination, in headline position one or two, or in description position one.
6. Check the Ad Strength again after every change, not only when the ad is created.
7. Let it run for two to three weeks before judging, so the algorithm has time to test the combinations.
8. Write down the date of your last change somewhere: in three months, you will know whether it is time to pick this work up again.

My take

A badly filled-in responsive search ad is not a failed ad, it is an ad nobody ever really wrote. Most of the accounts I take over have three headlines and zero variants, set up on the day the ad was created and never touched since.

My view: the time spent comparing third-party tools would be better invested in writing ten more headlines per ad group. It is less technical, it takes an hour, and it is exactly what the algorithm asks for to do its job well.

Nor is it a question of budget size. A modest account with well-written ads can compete on Ad Strength with an account ten times bigger that is poorly maintained, because the algorithm does not look at the size of the wallet, it looks at the material you give it.

The issue is not Google doing too much. It is that a responsive search ad, like a campaign built around a clear structure (in French), needs to be given material before you ask it for results.

Frequently asked questions

How many headlines and descriptions do you really need to fill in?

The technical minimum is three headlines and two descriptions, but aiming for twelve to fifteen headlines and all four descriptions gives the algorithm enough material to test different combinations depending on the search. Below that, the ad quickly hits a ceiling on Ad Strength.

Should you pin your headlines to keep control?

Only the text that must appear in every ad, such as a legal notice or a current offer. Pinning everything amounts to recreating a frozen ad and takes away the algorithm’s room for testing, which is what justifies the responsive format.

Can an ad have an Excellent rating and still not convert?

Yes. Ad Strength measures the diversity and relevance of the text provided, not the commercial performance of the offer. A well-written ad for a poorly positioned product is still a well-written ad for a poorly positioned product.

How long before you see a result after a change?

Allow two to three weeks: that is how long the algorithm takes to test the new combinations on a sufficient volume of searches. Judging a changed ad after three days tells you nothing reliable.

Do you need one responsive search ad per keyword?

No, one per ad group is enough as long as the group stays consistent around a single search intent. Piling up ads on similar keywords dilutes the data the algorithm uses to learn.