Posted by & filed under PPCChat.

This week’s PPCChat session was focused on Responsive Search Ads. Host Mark Kennedy sought PPCers views on negative and positive aspects of RSAs, strategies and tips used by them and more in this interactive session.


Q1 – As you know RSA’s are relatively new, but first, let’s get a feel about how many are using them and if you are using them a lot or just testing them a little.


We’ve been testing some responsive ads across our campaigns (at least 2 per campaign), and the performance has been alright so far. We focus more so on what keywords convert rather than what ads convert. @marccxmedia

Just testing them out in a few places. I’m not sold on them yet. @robert_brady

As of now, we are testing them in about 15-20% of our accounts, but so far not too much data to make any huge assumptions. Just small insights at this point. 

I am testing a little bit. The lack of data available is a big reason for my holding back. @NeptuneMoon

With RSA ads, it’s not so much test and learn as just… test. You don’t see which combinations drove performance, which is a big part of optimization. @amyppc

Using them in places where I’m not actively testing messages. If I’m testing, I pause them. @SEMFlem

We are dabbling in RSA but certainly no big rollouts yet. Reps are pushing them HARD. Had one client where G launched them w/o approval, that was fun… @Mel66

We’re testing then out right now; just testing.  @adwordsgirl

Just testing. But when setting up new adgroups I do try to make at least one RSA. And when doing ad testing I now always include at least 1 RSA. So volume is increasing slowly over all the accounts. @StephanieErne

I do test them with some accounts, but so far, they do not outperform ETAs and the data is still too little to commit to them vs. ETAs. @lchasse

Testing them a bit – I have not seen success on every single client/account so looking forward to see what others are doing. @andreacruz92

We have one in each ad group. They tend to have higher CTR, but less in the way of important and valuable conversions. @amaliaefowler

Used them before with previous clients but not using them now. But current client has sent an email at the possibility of testing them. yay. @mindswanppc

Out of several clients, only one saw better performance with RSA. I am not a fan of trusting automation with communication. Its better for bids and targeting than ad copy decisions imho. @DavidMelamed


Q2 – Many of you may be testing them on your own, but I also wanted to get a feel for how much pressure the agency-folk are getting from their reps, and why you think ( in your opinion) that is.


They are pushing the RSAs, but one of the my more experienced reps did say, the best use for them is to find combinations you can then test as ETAs. The more junior reps just say, add them and just use the script they are given. @lchasse

I stopped talking to reps a while ago. I just got annoyed of them pushing us to try things I didn’t think we were ready for. @adwordsgirl

The spin has been that RSA’s mixed with a DKI headline and BMM, gives the ads a 15% increase in impressions, which is why they want these added. @markkennedysem

Huge pressure. Got grilled hard when I turned down our rep’s offer to have the Google team implement RSAs for us. @timothyjjensen

Our agency rep moved on earlier this year so we are back to the quarterly reps – but all the pressure, all the time. I keep wondering what it must be like for SMB’s who don’t have the depth of knowledge we do. @amaliaefowler

They push them hard because they want to increase CTR and thus revenue and depending on the data around total click share of paid inventory. I really think they just want more clicks IMO. Which I do too. Not for you If you qualify people via copy and deterring others. @markpgus

Haven’t had real pressure but am quick to shut down recommendations that don’t serve my clients best interests. @DavidMelamed

We are dabbling in RSA but certainly no big rollouts yet. Reps are pushing them HARD. Had one client where G launched them w/o approval, that was fun… @Mel66

We haven’t been getting too much pressure from reps. I think one we spoke with a while back recommended it for one of our client’s accounts. @marccxmedia


Q3 – We’ll get into some of the negative aspects of RSAs in a minute, but first, let’s focus on the positives. Let me know what you like so far.


I do like the fact that you can see which combos get at least the most impressions (and how they appear), to test as ETAs later. @markkennedysem

You can’t argue the benefit of testing around 43,000 combinations in an easy way (I realize its definitely far from perfect but saving the other half for Q4) @amaliaefowler

Tends to be the exception rather than the rule, but I have seen them outperform ETAs in some cases. @timothyjjensen

Seeing the combination data on each set you create can be helpful in creating new ETAs. Other data is lacking, but that is the positive of the ads so far. @lchasse

The combinations are probably the best thing about RSAs. @adwordsgirl

I like the fact they can test different combos at once because that’s what we do on a normal basis, but only works if we can learn something from it. This prop is works better in H2, this works better than this etc.. It could be really good if they wanted it to be. @360vardi

Conceptually, I really like the idea as an option. IF the machine can truly put the best combo in front of a particular individual, that has promise. @NeptuneMoon

THIS IS WHY G has more signals than me and can determine the better combo in the moment. You just need to write copy that doesn’t need to flow in a specific order/be together/etc. @markpgus

Everything new is fun to discover and play with . I do like to see how Google is getting more and more automated. And I like that it forces me to be really creative (well sometimes.. when i’m on a roll). @StephanieErne

I do love the blatantly obvious of being able to take the best parts of all ads to use in a super ad. @JonKagan

RSAs are good for reach and query discovery. @SEMFlem


Q4 – And as a quick follow-up, are you seeing RSAs outperform ETAs (from a conversion standpoint), even without knowing which specific combination?


It’s hit or miss. In most cases, we’re seeing more traffic with RSA’s but less conversions than ETA’s. @amaliaefowler

In a few cases, yes. Really would like to get more insight into the combos that are performing! @timothyjjensen

No, but that’s to be expected. With X thousand combinations, it takes the machine a while to learn which ones to use where. For small accounts, it can be unbearably long. @SEMFlem

ETAs most of the time still outperform RSAs in the important metric my retail clients care most about (revenue). Some of the combinations of ad copy do better than ETAs in general, so I do use that data to create new ETAs. @lchasse

It’s not even close for us. for some reason a lot of times legacy ads outperform ETAs and those outperform RSAs. Even if you use the same phrases. Even if the legacy ads sometimes look funky. @360vardi

Not really a yes or no. Mostly it is performing about the same. I do wonder if there is a certain amount of clicks & impressions for it to work well. Like with conversions and bidding strategies. I can imagine only a certain amount data will give you real benefits. @StephanieErne

I find RSAs to be lower performing than ETAs in almost all instances. And I also find that they will eat an outsized impression share even though they are strictly worse performing. @CJSlattery

It’s really hit or miss right now. @adwordsgirl

So we’re still in a testing phase, so we check the keywords QS and performance pre vs current with the ads in, then we check to see if the ads are beating the adgroup average metrics. So far, this has been the purest definition of “mixed bag” results. @JonKagan

Again, we haven’t focused as much on RSA performance compared to ETAs and Call Only ads. If ads aren’t performing during ad testing, we pause them. @marccxmedia


Q5 – Okay, diving into the negative side, what’s your biggest issue (or issues) with RSA’s so far.


Lack of transparency. @timothyjjensen

Lack of data, lack of actual even distribution among entities, lack of insight into the “added auctions”, higher CPCs (why??) NO AD CUSTOMIZERS??? like, what the hell is that about? Ad customizers kill it in my accounts and you can’t use them with RSAs. @360vardi

Inspiration to make all these distinctive headlines and descriptions. And Google still advises 3 ads per ad group for good quality. After 1,5 RSA-ad’s I’m out of inspiration to be honest. @StephanieErne

Lack of data/transparency is the biggest negative for me. @NeptuneMoon

The fact that the combination of headlines if you give google the freedom to mix and match on its own are often garbage. Also the generally worse performance and the eating of impression share. @CJSlattery

Lack of transparency and data. Also so much time goes into getting every possible variation approved by client compliance. @Mel66

My biggest issues here are lack of transparency/insight into the data, the fact that the algorithm determines statistical relevancy in an imperfect testing environment with too few impressions and that Google preferentially delivers RSAs regardless of account goal @amaliaefowler

Just started RSAs. Most daunting task is, which Headlines to pin 1,2,3 or just any one out of them. Also which headlines will be good for enabling DKI. @wpoad1

Lack of data and transparency just like everyone else is saying. You can see that in all the answers up to this question. The feature itself is I think great, just need more data so we can modify as needed to create the perfect ad (for a short time anyway) @lchasse

As a starting point – borrow historically high-converting H1’s, H2’s, D1’s & D2’s, throw them into the RSA stew for keyword sets that make sense. Adjust as data accrues. @SEM_PPC_MattV

Negatives 1. Lack of Ad Customizers. 2. Inability to use learnings on messaging to apply elsewhere. 3. Difficult to use when you don’t want to maximize CTR (you only want enterprise leads for example) @markpgus

In a complete and over simplified statement: “there is no transparency to them” @JonKagan


Q6 – A few mentioned already (feel free to repeat) – Anyone have any strategies, theories or tips they are using/testing with RSA’s they want to share


You should launch experiments. never just run RSAs on their own in an ad group. combine them with 3 ETAs, anchor your customized headline. Show thte results to your client and your reps, so they stop with the threats and throwing you under the bus. @360vardi

The biggest thing to me is that you need to have copy lines that actually are substantially different. Obviously, you’re not going to get good information from 5 different versions of “Want thing? We have thing! Buy now.” But 99% of copy briefs are just that. @ferkungamaboobo

I test ETAs and know I have winners before throwing them into RSAs. I’d rather have thousands of combinations running of copy I already know works. @SEMFlem

Balance where and how you use them, watch them closely, mix with ETAs for coverage. @amaliaefowler

Take advantage of their useful features & what we do have data for. If certain combos get great CTRs, test them as ETAs. Run a couple ETAs along w/the RSAs. As a retailer if the RSAs are taking revenue away, you will have to pause them. Revenue does matter! @lchasse

I would suggest everyone keep a close eye on them and run some ETAs in tandem so you can compare performance later. @adwordsgirl

So far we run it in adgroups where it operates as ad #3 or #4. All the assets in the ads operate as additional standalone ads at once, to see if they can outperform each other. @JonKagan

Discuss ad messaging with the client to get all the needed information for enough distinctiveness. @StephanieErne


Q7 – Are you using all title and description lines in RSA’s?


It depends. I try to maximize the combinations, while also avoiding repetition. If I can’t get enough variables without repeating myself, I stop to avoid similar headlines appearing together. @amaliaefowler

Never done that before. I just think that would be overkill. @mindswanppc

Depends on the account and how inspired I am in that moment. @adwordsgirl

More or less we limit it to 5. To be honest, we only want to use the top performers to make the “super ad”, and we just dont have typically more than 5 super elements. @JonKagan


Q8 – If Google had full transparency on the ads (conversion data, CTR, etc) by combination, would you be more likely to use these ads?


Honestly, maybe. We’d treat it the same way we do DSA’s or broad match. To mine the top performers and then add them in manually for better attention. Not sure it would be much more of a game changer than that.@JonKagan

My head is saying yes I should, but I know it will be: no, I don’t see enough benefits of it yet to impact my accounts. Probably just keep testing it lightly untill I really get the feeling they outperform most of the time. @StephanieErne 


PPCChat Participants: 


Related Links:


Stop wasted ad spend with Karooya

Stop the wasted ad spend. Get more conversions from the same ad budget.

Our customers save over $16 Million per year on Google and Amazon Ads.

Leave a Reply

Your email address will not be published.