In our continuing series on RSAs, we will take an in-depth look at the RSA ad strength. In this post, we will cover:
Ad strength is a score Google gives your RSA based upon the total number of assets (headlines and descriptions) and the diversity of those assets within an RSA.
Google has four categories of ad strength:
As you add or edit headlines, your ad strength changes dynamically. In addition, Google will show suggestions to improve your ad strength.
One of the suggestions often seen is not to pin the assets. This is a controversial action, so we’ll take a look at how pinning affects ad strength.
We used data from several thousands of accounts to find RSAs that met the following conditions:
By following the above criteria in selecting the RSAs we are confident the conclusions reached in this post are as accurate as possible.
When we look at ads with Excellent ad strength, most of them have at least 13 headlines, although some had only 8 headlines. No RSAs with under 8 headlines had an Excellent ad strength.
When the above analysis is done for ‘Good’ ad strength, we found one needs a minimum of 4 headlines to achieve a Good ad strength.
Ad strength isn’t only about the number of headlines you have. Google also checks if your keywords are in the headlines and if you have diverse enough content in your headlines. To confirm that ad strength is not only dependent on the number of headlines but rather on the content as well, we examined RSAs that have Poor ad strength and found the majority of them have 15 headlines.
This confirms we cannot enhance the ad strength by simply adding more headlines. We still need headlines relevant to the ad group that are different from each other.
If you have more than 8 headlines and do not yet have an Excellent ad strength, look at Google’s recommendations for ad strength. If you see a recommendation to ‘add keywords to the headlines’, and your headlines already have some keywords in them, look at the keywords in your ad group and see if some of them need to be in a different ad group to have more relevant ads associated with those keywords. Our previous article looked at RSA creation and how you still need well-organized ad groups even with RSAs and their potential headline diversity.
To examine the relationship between pinning and ad strength, we split our RSA data set into three categories:
We then counted the ads in the above categories segmented by ad strength.
From the above, we can see it’s extremely unlikely to have an Excellent ad strength with any degree of pinning. However, ads that use full pinning can easily achieve a Good ad strength. If pinning is part of your strategy, aim for a Good ad strength (since an Excellent score is much harder to achieve).
If you have multiple RSAs in an ad group, do their ad strengths affect the impressions they get? To conduct an accurate comparison, we considered only ad groups where all RSAs used a very similar pinning pattern (to isolate any impact pinning might have on impressions).
When we examined this data, we see the RSA with the higher ad strength gets more impressions 56.8% of the time. However, the RSAs with the lower ad strength had more impressions 36.6% of the time. We conclude there doesn’t seem to be a clear or strong relationship between ad strength and impressions. Within an ad group, an RSA with higher ad strength will likely, but not always, receive more impressions than the RSA with the lower ad strength.
To further analyze the above, we also looked at how the impressions differed with the degree of difference in ad strength and found there wasn’t much of an impact. For instance, the way the impressions differed in an ad group with ad strength of Excellent vs Poor was very similar to when the ad strength was Excellent vs Good.
To examine any relationship between ad strength and CTR or Conversion rate, we again considered only ad groups where all RSAs used a very similar pinning pattern.
In the above chart, we can see that the higher ad strength ads had a lower CTR than the lower ad strength ads in 51.5% of the cases. This number is immaterial and ad strength and CTR do not seem to have any direct correlation.
For the conversion rate comparison above, we once again see that the higher ad strength ads had worse metrics than the lower ad strength ads in the same ad group.
In looking at these charts, there does not seem to be a direct relationship between CTR or Conversion rate and the ad strength. We looked at this data in a few different ways and reached the same conclusion in each scenario.
Given the above where no strong correlation was found between ad strength and the ad’s performance, it seems ad strength is purely a measure of Google’s ability to test the RSA, and that having a high versus a low CTR did not impact the ad strength.
In a lot of our analysis, we often found higher CTRs or conversion rates achieved by ads that use pinning. However, the act of pinning by itself does not have an impact on performance. If you are pinning well written assets then, just like writing any good ad, it is common to see higher CTRs and conversion rates due to the better ad seen by the searcher. On the other hand, pinning poorly written assets would typically result in worse performing ads.
The more unpinned headlines you have, the more variations of ad combinations Google’s machine learning needs to test and optimize. Pinning good headlines often shortcuts the learning process since you are giving the machine fewer variables to work with. Assuming you are only pinning well written headlines, you will almost always see an improvement in your ads performance.
This might seem controversial, but having an Excellent versus Poor ad strength does not seem to affect your account in any significant way. We conclude from our research that:
As Google’s RSA machine learning improves over time, we may see changes to the above conclusions and other account factors. As always, however, the main goal should be to create RSAs that convert regardless of their ad strengths. As you plan the transition from ETAs to RSAs, feel free to check our best practices on how to Create and Test RSAs.
This RSA Optimization Series brings you all the knowledge you need to manage the ETA to RSA transition and achieve your account goals with RSAs.
We’ve had a few comments around the web asking how ad strength affects CPC.
In this chart, we compared CPC to ad strength and found that your ad strength doesn’t have any material impact on CPC.
Ad strength seems to have very little impact on any of your account’s metrics with a possible exception of which RSA receives the most impressions within an ad group.
Excellent article, Brad. This confirms what we have suspected and seen with our own accounts. I agree that the main goal of RSA’s should be based on conversion performance and not Google’s ad strength “metrics”.
Hi, Brad! Great article! (and one of the hot topics!)
I’m moving from ETAs to RSAs and restructured an account to fully implement Hagakure strategy with RSAs. This means I’m using only 1 Ad Group to get the max volume of data to feed the algo and using Customizers to fully customize my ads. However, I have some questions regarding the Ad Strength. Have you already faced this type of challenges?
– using customizers as the whole headline/description, as to guarantee the whole copy is as customized to each keyword as possible. However, Ad strentgh at the preview is always at the minimum level. (despite of max customization).
it is not clear to me if the platform is able to consider keywords within the customizer itself to evaluate the headline/description’s strength.
Do you have any idea?
– Should I still use bid modifiers when using smart bidding? Does it not make any difference, or do they work as additional signals/directions for the algorithm in any way? How do they work in this case?
Thank you!
Hi Luis,
I don’t think the point of the Hagukure strategy is a single ad group. It’s just how most of us who didn’t like SKAGs have been organizing accounts for 20+ years. There’s not really anything new to it. The overall goal is fewer campaigns so you aggregate your conversion data by campaign as smart bidding works on campaign level data.
While you can use a single ad group, your ad strength is going to be all over the place. It might not matter at all, but don’t expect to see a high ad strength. You’ll also have a lot of poor quality score words. That isn’t auction quality score, but it will be hard to understand ad relevance with a single ad group.
You can do it how you described, but you might be better served to just make an ad group per theme as the ads, and any potential sitelinks/callouts, can be easily optimized on an ad group level.
This article should answer your questions about bid modifiers: https://adalysis.com/blog/how-bid-adjustments-are-used-by-bidding-method/
Good luck.
This for me sums up the futility of paying attention to ‘Good’ or ‘Excellent’ Ad Strength:
More impressions. Lower CTR. Lower conversion rate.
Thank you for putting data behind what most PPCers already had as their gut feel.
My ETAs still kick RSAs’ ass. Also, RSAs in my experience are far more likely to get spuriously disapproved. <– Hey btw do you have any way of looking at data on that?
Hi Mark,
I haven’t thought about looking at disapproval or reasons by ad type. I’m not sure it’s going to matter after June as you’ll have to live with RSAs, but it might be interesting to examine. I’ll think about this one.
Great article.
Would pinning affect ad relevance? For example, your ad group has four keywords. Your ad is completed with 15 headlines and four descriptions; both the headline and description contain your keywords, and you pin a headline and description containing only your primary keyword. All other pinned items are USPs. Would the ad relevance of the remaining keywords in the ad group that are not pinned in the add suffer?
Thanks,
Scott
With ad relevance, you don’t need the actual keyword in the ad for the vast majority of keywords (there are some odd exceptions with attorney, lawyer, etc). You need to ensure the ad is about the keywords. Therefore, ad relevance should not suffer with pinning. With ETAs you can have excellent ad relevance for many keywords in one ad group using static lines, that won’t change with RSAs.
What will become a bit more complex is trying to decipher ad relevance since there are so many combinations possible. At the moment, managing ad relevance suggestions seems to be a problem for Google that they haven’t solved with RSAs.
Brad, are you surprised that ad strength doesn’t affect the number or quality of impressions or impression share? Some Google reps essentially tell you outright that it does, in what feels a bit like a threat.
Thanks for running the hard numbers on this and sharing your data.
Hi Rob,
I’m not surprised at all that ad strength doesn’t affect impressions for a few reasons.
1. If Google slows your impressions, there is less auction pressure, and therefore CPCs drop. In that scenario, Google makes less money.
2. Fully or partially pinned RSAs will be needed by some people for legal disclaimers, marketing coordination across channels, etc. They don’t want to penalize people for using required disclaimers, cross channel marketing, etc.
3. RSAs are additional options. You do not have to use them. If you want the ETA ad experience, you can have it. If you want Google to take over ad serving, you can do that. In many ways, RSAs are very flexible and you don’t lose any options – you just gain some over ETAs.
Due to this, I’m not surprised that ad strength doesn’t affect how ads are served and that ad strength is just a suggestion on how to improve the ads for advertisers who want some guidance from Google.
Great article, very helpful!
One more question on the relationship between ad strength and number of impressions. You stated that when there are 2 RSA’s within one adgroup, the one with higher ad strength mostly get most of the impressions. This kind of makes sense to me from a Google perspective.
Though, what I’m wondering is: What is the impact of ad strength on impressions if there is only 1 RSA in an adgroup? E.g. if I have an adgroup with 1 RSA with poor ad strength which currently get’s 1000 impressions. Can I assume that the number of impressions will not necessarily go up when the ad strength goes up? Since the article says that in the end there doesn’t seem to be a clear or strong relationship between ad strength and impressions.
Thanks!
Hi,
You are correct. If you only have a single RSAs in the ad group, the ad strength of that ad does not affect that ad’s impressions. The ad strength could go up or down, and you will not get more or fewer impressions.
The only exception could occur when you are using broad match. With broad match, Google has said that with RSAs you might be eligible for more auctions (expanded broad match essentially). These auctions are for different search terms than you are currently displaying ads. So you won’t get more impressions for your current words. but you could get more impressions for new search terms.
Check Your Account For Free!
