Expanded Text Ads out. Responsive Search Ads in.
With RSAs going to be the only standard text format starting from July 1, 2022, there’s going to be a pretty big shift in how PPC marketers approach text ads.
In this episode of PPC Town Hall, we spoke to Julie Bacchini and Ed Leake — two well-known PPC experts with great RSA experience under their belt — to talk about how to deal with this transition.
Here’s the full episode with the transcript.
Fred’s introduction to the episode.
Fred: Hello, and welcome to another episode of PPC Town Hall. My name is Fred Vallaeys. I’m your host and I’m also the co-founder and CEO of Optmyzr.
So hot topics in PPC…Well, let’s see. Automation, Performance Max, RSAs.
So let’s talk about one of these. This week we’re going to talk about RSAs. Responsive Search Ads, as you all know, is the ad format that’s going to become the standard ad text format in Google Ads. It’s replacing the long-loved Expanded Text Ad.
And before that, regular text ad. But basically what’s happening is Google is saying, you no longer have to write a fully qualified ad text with a bunch of headlines and a bunch of descriptions.
Instead, just give Google the components. Give them a bunch of headlines, a bunch of descriptions, a bunch of calls to action. And Google’s automation will put it together based on what it thinks is going to be the best ad to serve for that specific auction, that specific search.
So there’s a pretty big shift that’s happening. And we’ve got some great experts with us today to talk about what you would do as the shift happens to RSAs. And we really have to start thinking differently about how we manage messaging and attics on Google.
I’m really excited about this episode. Let’s get rolling with PPC Town Hall.
The panelists introduce themselves.
Fred: All right. Our two guests this week are Ed and Julie. Welcome to the show.
Ed: Thanks for having me.
Julie: Thanks for having me. Yeah.
Fred: Thanks for coming back, Julie. We’ll start with you. So let’s do a quick introduction to who you are, what you do, and what you know about RSAs.
Julie: Okay. I’m Julie Friedman Bacchini, and I have sort of dual roles in the world of PPC. I have my own consultancy, Neptune Moon. I’ve been in paid search since literally the beginning, so I’ve seen it all.
This is yet another change that is happening. And I also manage PPC chat on Twitter. So I am managing all of the weekly chats and trying to be a source of the latest information gathered.
It’s hard to keep up with everything. So I do my best every week to try to gather information and make sure people can easily find what they need to keep up with what’s going on in the ever-changing world of PPC. Yeah.
Fred: And thank you for doing PPC chat. We’re big fans here. It’s just a different format to get your news, basically, and then have a great discussion around it. So I’m a fan of Search Engine Land. I’m a fan of PPC Town Hall if you prefer video.
But PPC chat is awesome because it’s just on Twitter. And even if you don’t have time when it’s live. And Julie, it’s usually what, Wednesdays at 09:00 AM Pacific?
Julie: We have two chats per week now. One is the traditional Twitter-based chat, and that’s at 12:00 Eastern on Tuesdays and then at 12:00 Eastern on Thursdays. We have been doing an audio version in Twitter spaces of typically the same topic that we talked about on Tuesday of that week.
Fred: Nice. So, yeah, check those out. And Ed, first time on the show. It’s great to have you on. So tell us a little bit about who you are, what you’re doing, what you know about RSA.
Ed: Yeah. Hi, I’m Ed. Thanks for having me on. So I’ve been an agency owner for twelve years and my side project is Ad Evolver, which is a Google Ads optimization tool.
Not sure if I’m allowed to mention this, Fred.
Fred: Yeah (laughs).
Ed: You can edit this bit out (laughs).
Fred: There are many great tools out there. Right. So let’s use them for what they’re best at. Not a problem at all.
Ed: And then to keep myself busy, I also started God Tier Ads about a year or so ago, which is a Google Ads training framework and product, that sort of thing.
Fred: So you’re basically saying you don’t do so much.
Ed: I don’t sleep. I don’t have kids, though. So I’ve got an excuse to work, unfortunately. That does help.
Google’s latest update on RSAs.
Fred: So, yeah, let’s talk a little bit about RSAs then. I think maybe we’ll start with what is happening. So I talked about how this is becoming the new ad format.
The latest news we have from Google is that the end of Q2 is the last day that you can still put in traditional text ads. These are called ETAs.
ETAs are expanded text ads because before that they were not expanded. So ETA is just like a regular ad, but basically end of Q2 and then starting in Q3, I think you can still pause and enable ETAs, or you can no longer make edits.
You can no longer make a new one. So that’s kind of the timeline we’re working with now. How aggressive have both of you and or your customers been with kind of this transition?
How are Julie and Ed dealing with Google’s transition to RSAs?
Ed: Julie? Ladies first.
Julie: Oh, thank you. I have been working with RSAs for quite a while because it feels so different when you’re creating. It’s so different from how we’ve done ads up until this point, even the regular text ads and then the expanded text ads where you had more characters to work with and the extra headlines and that kind of thing.
RSAs are just a totally different animal. And so when you’re used to writing and I’ve been doing this for a long time. So when you’re used to writing the more traditional ads, you’re used to coming up with your ideas for testing like, oh, what if we test this offer against this or we move this in this position and whatnot.
It’s really different when you start to create the RSA because you’re putting in so many elements into a single ad. So you’re like, my God, I have to come up with ten headlines. I have to come up with four different descriptions.
It’s a very different way of thinking and a different process that you have to go through. So I’ve been working with them for a while now and knowing that this date was coming when you wouldn’t even have the option to create new ones.
I’ve been working harder on them, and I know we’re going to get into this later, but once you start using our sales in your ad groups, your ETAs don’t tend to serve a whole lot.
So there’s that whole dynamic that you start to deal with as well.
Ed: Yeah, I mean, pretty much the same as me. We essentially adopted them as soon as they were available because Fred shared at the start of this call, actually, sorry to do the thing of sharing information that we haven’t recorded, but what was it? 17% of people haven’t tried an RSA yet.
So the argument is that they’re behind. We always test new features as soon as they come out, to be honest with you. And it’s normally an easy sell to clients anyway, because new things normally mean you get a little bit of an edge.
But yeah, with what Julie said, RSA is a different mindset because it’s less about the A/B. You’ve got one version versus the other. It’s more about making the struggle of writing 15 headlines that make sense, which is why you shouldn’t write 15 headlines, probably.
So our approach to the RSA situation overtime was essentially treated like two ETAs. So you have to do twice the work but compress your two ETAs down to a single RSA. That’s the way I sold it to the team anyway, to make it sound like it’s no more work to do an RSA.
Fred: Exactly. I think when Google launches new stuff, we don’t always jump in feet first because there are plenty of things that they launch and they just go to advertisers and they say, hey, can you please try this?
Because they’re honestly just trying to see if this is any good. In the case of RSA, I guess it was pretty clear relatively early on that, it was going to be a winner for Google and that they were going to push it through.
Right. But always be a little bit cautious. I suppose you don’t want to be the guinea pig in the Alpha necessarily, but once it comes to a beta stage, I tend to agree to jump in, because being early on may give you some benefits.
Even if you don’t get benefits, there’s this whole benefit, right. When Google introduces something new like you’re the only ad on the page doing that new thing. And even if you do it horribly, it looks different and you’re still going to get more clicks because people are like, oh, what’s that like?
There are four lines of something or there’s a little image next to it, so they just gravitate towards it.
Right. But then even the early learnings that you have in those extra couple of months that you’ve sort of figured out what’s the strategy. What’s the best process here that can be really beneficial?
So the RSA study that we did at Optmyzr, so we’ve done two of them.
Check out our RSA study here.
We did it once for SMX Next in 2020, and then we did it again in SMX next 2021. But we ran through a bunch of ads, millions and millions of ads that we run through the Optmyzr system.
And we looked at how many advertisers had not yet tried RSAs at all. Last year, that was at about 25%. This year, that’s down to about 17%. But that’s still kind of a shocking number like you said. Right.
So 17% of people, given that this is going to go away in about four months, have still not experimented with the new ad format. What’s also fascinating is that for those who have tried it, only a very tiny portion of 0.2% tried RSAs and then stopped all RSAs.
So what that kind of tells me, and you’re welcome to take a different angle on this, but it kind of tells me that when you turn on RSAs, they seem to do something good enough that people want to stick with it.
Do you think that’s an okay read, or do you think it’s just that people are like, well, this is the thing I’m going to have to deal with, so I might as well fight the bullet and go with it?
Ed: I think it’s all the above, isn’t it? It’s easier as well. To a certain degree, you can be RSA is going to create a few lazy account managers. Dare I say it? And I know a lot watching this stream, but that’s an opportunity for other people that aren’t so lazy. No offense.
Fred: I’m kind of curious as far as the lazy account manager. So what were both of your strategies in the beginning? So say that a new client comes to you and they haven’t done RSAs? Like, would you go and write entirely new ad units, or would you just do the conversion and take their existing ETAs and turn them into like chunk them out into headline pieces and description pieces and put them in the right place in the RSA?
Ed: Right now? So one of that 17.5% turned up, knocked on the door, and said, we don’t have any RSAs, take the best ETAs. Hopefully, they’ve got a good year’s data and more than two ETAs and take the best components and create an RSA.
I don’t want to jump ahead because now we’re going to talk about optimization and stuff, but there are some sweet spots as well for the amount of headlines and that sort of thing.
But yeah, easy street. And also, I just want to say that when I said lazy, by the way, I don’t want to offend anyone if they’re really good at ad copy and really smart. Then be lazy because you could create an awesome RSA and get away with it for some time, probably.
Julie: Well, I mean, RSAs are definitely testing of ads as we have historically known. It isn’t just not going to exist anymore in the world of RSA. Right.
Historically, up until this point, when you wanted to kind of test different things and figure out maybe what would be more successful in your ads, you would do A/B test between expanded text ads, and then you would know the variable that you were testing, hopefully.
You had a testing plan of like what you were doing. You weren’t just kind of like randomly doing things. But if you had any method to what you were doing as far as what you were testing, it was pretty clear where you were going.
And then you could record those results and then try to build off of that. The way that RSAs function is completely different. Like that traditional mindset and that traditional method of I’m going to test headline A against headline B. Right. Like everything else in the ad is going to be the same.
But I want to see what happens if we have a different headline. You can’t do that, really. Theoretically, I guess you could do that with pinning of things. That’s the whole piece I’m sure we’re going to get into.
But just the whole way that you think about RSAs and how you use them and what kind of information you can glean from them, it’s totally new territory.
Fred: I have this fascinating example that I always give, which is you have an ad group and then the ad group has ten keywords. Okay. And now you have an A/B test, but you actually decide to make the B version the same as the A version, but you just make it as a new ad and you see what happens.
And then traditional A/B testing, where you’re usually looking for what is the best headline, you will often actually find that the A/A test has a winner and a loser. It doesn’t make sense. It’s the same ad. Right. But the thing that people then forget is while you had ten keywords and this ad group is running for all those ten keywords.
And by the way, these keywords might be broad matches. So the thing that the person actually typed in wasn’t that keyword.
It was some fairly out their variation, because broad match, as we know, is not that restrictive. And then it’s a different audience, it’s a different time of day. Like it’s different geography that is coming in from they Typed in a different query and they saw an A/A test and one of the ads starts winning. Right?
So then the question is, was it really the ad that was driving that change, or was it the fact that other circumstances with the keyword and clearly the second. But that’s a mistake people make.
And so that’s where I think you have to be careful. And RSAs are a bit better because if they do their job, as Google promises it, they should figure out what are those circumstances around each search inquiry and then put together the right ad for that person so that even if you’re not doing the AB testing, they kind of do it for you.
Ed: Yeah, but I agree with that. But the example of A/B testing with ten broads is that’s a setup issue as well. Don’t get me wrong, a lot of people do it wrong and don’t have a controlled test.
But yeah, I agree, and people misread them. And as a result of that, you got 10,000 queries driving two ads off the back of a dozen keywords, and then within seven days pick a winning ad. It wasn’t really the winner.
And as Julie said, it’s not that straightforward anymore, RSAs, is it? But also to touch on what Julie said, we’re going to lose that compounding benefit of doing proper AB tests. The trial and error of A/B tests over a period of time does give incremental value to an account. We’re going to lose that.
So I’m interested to see for the people that are quite strict with AB testing, how that will that compounding performance will go away, potentially, unless the machine learning is that good, that this is all mute, then we just stick a light about it and walk away.
Fred: So given that there’s this compounded benefit of continuous AB testing, obviously with RSA you can still do pinning, right? So you could technically make the equivalent of an ETA by pinning everything to a certain position.
Google says, well, you lose the benefit of the machine learning doing its thing. But as you would argue, you get the benefit back of the A/B testing, I guess.
How good and dedicated are you? And then you can do better than the machine? Or is there a happy medium?
Ed: I think analytical people really struggle with ad copy and add writing. I’ve looked at God knows, hundreds, maybe thousands of accounts. I don’t know, I’ve lost count. And I think it’s the single weakness of most PPC managers is the ad copy.
So can they beat the machine? Probably not over the long run. So I’m defending Google now and the machine. So hold on, I’ve offended PPC managers, and now, oh God! Do know evil.
I think, in fact, that is one of the angles that a PPC manager, particularly freelancers and small agencies, can take is to team up with really good copywriters and admit that it’s not their strength.
And if you’re going to fill stuff, if you’re going to fill an RSA, make it worthwhile. But I’m going off on a tangent.
Fred: Well, yeah, let’s hear from you, Julie. How do you feel about that? So you work with a number of accounts. How many people do you work with right now?
Julie: Neptune Moon is just me. It’s been various sizes over the years, but currently, in the pandemic world, it’s just me.
Fred: How important is the ad testing and what do you think of Ed’s point?
Because I do agree we tend to be so analytical, and the reason that I got into PPC was more my technical side and my analytical side more so than being a marketer or a great messenger. How do you feel about that?
Julie: So I come from more of a creative background. I was a marketing person, but I also did design, I did PR and marketing, and design. I have more of that writing background.
So for me, I like writing the ads. That’s not a challenging piece for me. That’s well within my particular wheelhouse.
But I think one of the challenges is this is what happens a lot with stuff with Google. They have a particular type of advertiser in mind when they move forward in different directions, and then everyone else kind of has to figure out how to work within the parameters of what ends up happening.
So I think I’m hopeful that because RSAs, in the grand scheme of things, is relatively new, as it becomes the only option, some things will evolve. So we talked a little bit about this on #PPCchat.
Within the last week, we were talking about how there are quite a few instances in industries where you have to pin stuff like, from a compliance standpoint, not every Advertiser in every industry is on even footing as far as how flexible you can be in what happens in your ad copy.
And I think RSAs are not equipped to deal with that properly right now. I hope that that’s something that is being thought about inside of Google because right now the big push is to go broad match plus RSA plus smart bidding. Right?
That’s the golden triangle that they talked about, in their Welcome to 2022 presentation from last week. And that’s all well and good, but there are quite a few industries, and then there are other businesses who feel strongly like, I want to make sure that my business name, for example, is in that first headline.
So I don’t feel like RSAs right now do a great job with that. They clearly want you to not have anything pinned. Like, the preference is definitely I’m running an experiment right now in a couple of my accounts where I’m running the exact same ad where I have pinned some of the assets, and I have everything else unpinned, and it’s really preferring the unpinned ad.
So I think there are definitely some things that I’m keeping my eye on as far as things that you might want to adopt. Right. Because it’s like, okay, well, Google is telling you this is where it’s going, and we all know that you can’t swim upstream against Google forever.
Right? It’s futile. So you have to figure out how to work within the parameters and with the system that Google is going to provide for you. But I still have some questions, and I hope that these are things that are kind of like on the board over at Google where they’re figuring out, like, hey, we might want to think about this, or this is something that is going to affect a certain percentage of advertisers.
So that kind of remains to be seen. So those are some holes that I’m seeing right now.
What’s the impact of pinning assets in RSAs?
Fred: Yeah. And let’s get into the tactics a little bit. And I know you wanted to get into those two and the strategies, but before we do, like on the pinning topic that Julie brought up.
So part of our study looked at the impact of pinning. So you can see the chart on the screen. And for those of you listening, not watching, basically, if you pin, your conversion rate goes up slightly, but your cost also goes up.
So your cost per click is higher, your cost per conversion is higher, CPR is lower. Right. And that’s kind of the whole premise that Google believes that if you let them figure out how to make the ad, it’s going to be more relevant.
Hence have a higher click-through rate. But then it might be the wrong message. So maybe that’s why the conversion rate gets slightly worse.
And then I think the cost per click is heavily associated with the CTR that’s the ad rank, the quality score component. If you have a lower CTR, you just have to pay more to maintain that same position.
So that’s why the cost goes up. From that perspective, pinning helps you with your metrics to some degree, but it’s very counterintuitive to what Google wants you to do. Of course, if you’re in an industry where you have to do it, then you have no choice.
Ed: Okay, sorry, Fred. To interrupt. Did you do multi-pinning versus single-pinning?
Fred: We did. Where is that one? We did multi versus single pinning as well.
Ed: Because I think that would be interesting because I definitely see that pinning impacts the impression share, particularly top of page rate as well, which, again, that’s Google being.
I’m not allowed to swear, am I?
Julie: Googley. They’re being Googley.
Ed: Yeah, it’s…But it screws over. Found the word that didn’t have an FC or B in it. Screws over the people that can’t… They have very strict brand guidelines that Julie was just talking about because we found a single pin is quite damaging to impression share, not necessary to performance metrics.
I’m just trying to look at the screen. If you took a few outliers there, really conversion rates, I mean, cost per click, look at that bumps. But the conversion metrics look similarish for a pin, but they hurt click-through rate and they bump CPC.
So essentially you’re losing the auction. If you pin so Google’s forcing your hand, it’s like another layer to qualify for.
Fred: So one interesting thing about pinning and this is a tactic people should definitely try. In the beginning of RSAs, you could pin a single thing to a single position. So you could be like, this needs to be my headline one, and maybe that was your brand, whatever.
Right? But then over time, I don’t think they really announced it kind of like snuck in. But now you can pin multiple variations to the same position. So one technique that seems to make sense to me is I want my headline to be something about my brand.
How do I portray my brand? I don’t know. Is it my domain name? Is it my company name? I don’t know. But I’m going to have three variations as headlines that are something to do with my brand.
I’m going to pin all of those to position one, and that’s what this chart on the screen shows is if you do no pinning, then Google prefers your ad because they have the most liberty to do whatever.
If you pin a single thing to a single position, Google doesn’t like it because you’re basically going back to kind of an ETA, very highly defined. If you do the thing where you pin multiple things to one position, then the results do improve.
Now it’s kind of like that happy medium. Google’s happier because they have flexibility. You’re happier because you got some control. So that seems to be kind of the way to go. If you feel like you want more control, does that make sense?
All right, since we’re talking about tactics, Julie, you were talking to about once you start RSA, your ETAs get very little love from Google, their impressions go way down. Do you want to talk about that a little bit more about what you saw?
How have RSAs impacted ETA impression share so far?
Julie: Yeah, I’ve certainly seen it in my own accounts that once you put an RSA into your ad group, it gets the lion’s share of the impressions almost right away.
I feel like even as a predecessor to this, before RSAs got the amount of preference that they get. Now, even Google had been picking the “winning ETA” almost from Go for probably the last two years.
It’s been really fascinating how they’ve been kind of like in the background, moving us towards this idea of we’re going to pick like, you’re going to tell us some stuff, but we’re going to pick what the best ad is.
So even if you had all your settings and you were just using ETA, where theoretically you wanted to have your ETAs, be able to kind of battle it out a little bit. Google has been choosing a winner, like fast, and you will find that one of your ETAs even was getting the lion’s share of impressions.
So it’s been difficult to even have impressions spread out over multiple ad variations. But now if you put the RSA in a group and you have ETAs, forget it. Your ETAs are occasionally if you have a strong performing ETA, it will still get some.
I have a couple of ETAs that are still getting decent impression share. But they said again in that Welcome to 2022 presentation, they were talking about how once ETAs are deprecated that if you continue to run ETAs, they will “lose value” in the same way that if you are still running broad match modified keywords.
Now they have also “lost value”. So they’re very much kind of telling you very clearly you can keep doing it. But I wouldn’t recommend it because we’re not going to show it. It’s not going to have the value for you that the RSA which they want you to do is going to have.
Fred: Yeah, interesting. And so part of the study that we did as far as RSA versus ETA and the difference in impression volume. So to us, the biggest insight was literally how many more impressions you can drive by having an RSA.
So we did comparisons where it was an ad group and an RSA plus an ETA. It was four to five times as many impressions per ad going to the RSA ads.
Ed: And it’s not just that, Fred. I pulled the end of last year and I don’t know if you’ve done this if you got similar data. I was just looking at the data we pulled.
I can’t remember how many accounts it was. I know it was like 8 billion impressions, a 17% increase in top page rate for RSAs, and CPCs were flat. So not only is Google biasing the impression share to an RSA, you’re getting a higher ad rank, essentially.
So it really is rewarding RSA more. That was last year. I got the year before. I don’t know if you saw the same, but to me, that’s the page.
Fred: But that’s the promise of RSA. Right? As Google says, we can show a more relevant ad based on what we know for the auction.
And then yes, naturally you’re going to have a higher ad rank because it’s more relevant ads. Your predicted CTR is higher. And we can see that in the stats where we look at CTRs of RSA versus ETA.
And so even with the same bid, you’re going to score higher, you’re going to take more of those top impressions. But then we did a study on this not two years ago and it was basically, “Look at your keywords and look at when they serve with an ETA versus an RSA and start to understand the incrementality portion of that.”
“And let’s see how many queries search terms exist now that you have RSAs that didn’t exist in your account before.” And that’s really the fascinating thing is that searches that Google before said your expanded text ad is simply not good enough to qualify to show an ad.
All of a sudden with an RSA, they’ve kind of like put together some ad that for some reason now is good enough. And it’s not the only driver of this. But this year we are seeing 4x more impressions.
And then people sometimes complain. They’re like, well, and even if you look at our stats like you get 10% to 20% versus conversion rate from an RSA. But if you get five times the impressions and you’re controlling your bids in a smart way, like five times more impressions, 10% less conversion rate, well, that’s still a hell of a lot more conversions at the end of the day than before.
Ed: Yeah. If your business has got fixed costs, definitely. It makes a lot of sense to scale incrementally and take that additional volume. And there’s a lot of businesses that are volume play insurance and finances and stuff.
I think it hurts potentially sort of small to mid-tier retailers a bit more because they might get more volume. But ROAS, tends to dip, doesn’t it? in some of the conversion metrics of RSA. Excuse me, but then they’ve got shopping, so you can’t have it probably…
Fred: …should be the main thing for them anyway, right?
Yeah. And I know people are very curious too, about Performance Max. So since we’re talking about shopping and automation and all these different apps, did you guys have any thoughts on P Max that you can share with our viewers?
I know it’s not a topic, so it’s totally fine. If you don’t want to talk about it, let’s skip that one.
Ed: Look, Google is heading towards the target is you give me a URL, you give me a budget, and you bugger off and we do everything else for you.
That’s Google’s end game. And is it five years away? Is it ten? Who knows? But it’s understandable because it cuts out all the drift in the middle. Any issues and Performance Max is a bit of a test of that ideology, isn’t it?
But you start to create your ads.
Fred: Exactly. I’ll do a selfless promotion here, but my book came out recently. So this is my second book, ‘Unlevel the Playing Field’, and it’s based on that whole premise that everybody’s basically given the same automation tools from Google and for the most part being told to bugger off like you said.
But what if you don’t want to bugger off? What if we make a living at this, right, Julie and Ed? You find one. Plus your clients expect you to be the difference. So how do you level that playing field back in your favor? You’re right.
I mean, five to ten years down the road, who knows what it’s going to be like? But right now, yes, there are a lot of things you can do to make things better.
Ed: Also, Where’s my free book?
Fred: You’re in the UK. Do you know how there are no ships with containers going across anymore?
Ed: What nonsense? We’re open for business. Come over.
Fred: I actually am coming to London in March and leave it at that point. I’m going to be speaking at HeroConf London. So assuming that that happens also, you’re going to potentially be at SMX Paris that week and doing SMX Munich as well. So little Europe stint.
Ed: Well, let’s not talk about lockdowns and all that nonsense.
What’s the impact of RSAs on ad accounts?
Julie: So I have a question that I don’t know if you’ve looked at the data relative to this Friday when you guys are looking at stuff with RSA, but some of the stuff that’s been going on with RSA, it’s been concurrent with the continued fuzzying of all match types for keywords.
So part of me wonders like, yes, we’re talking about what impact does RSAs have when you’re using that and what data points we mirror. Advertisers don’t have access to what Google does that they’re using to make all of these auction time decisions on our behalf.
What role, again, I think it’s pretty much a black box. So I don’t know how we would tease this particular piece of information out, but I do feel like the Fuzzification, as I like to call it, of the keyword matching definitely is playing a role in all of this, too.
Because even if you’re not following Google’s Golden triangle recommendations and you’re not using full-on broad match with the smart bidding and the RSAs, you’ve got like broad match, like no matter what keyword matching you’ve chosen for your terms.
So that’s a piece that’s at play here at the same time, which is kind of behind the scenes and we’re not really factoring in necessarily, but I think it is playing a role in what happens with your ads as well.
Fred: It’s a great point for those of you not completely familiar with what Julie means with broad match lite, an exact match is no longer an exact match, and it’s becoming less and less exact as time goes on.
And I think that’s what you mean by broad matching. The problem is it’s hard to figure out what part of the incremental impressions and incremental conversions are driven by the match type versus the bid management automation versus the RSA automation.
But listen, one is going off on a bit of a tangent, but one big point in the book. And so, by the way, like the subtitle of the book, the biggest mind shift in PPC history. So what is that biggest mind shift?
To me, it’s like, wow, we’re so used we’ve been doing this for 20 years of managing keywords and search terms and specific bids and specific effects and like all these details and how they come together in these like for this query, we want to do this. For that query, we want to do that.
That’s no longer how ads are going to be managed. So to Julie’s point, it’s a black box and Google does all this stuff that we used to do for us, but the way that we can still influence it and steer the ship if you will, is by managing at the periphery of that system.
So how do we tell Google what we really want, right? We’re not writing them a blank check. We’re writing them a blank check in exchange for a conversion goal that we told them.
But if we don’t directly tell them that conversion goal, then that’s a huge miss because here’s an example. Like, if you say, I want phone calls for my business, well, I don’t want phone calls for my business.
Nobody wants phone calls for your business. You want phone calls that turn into new customers by just having the phone ring doesn’t mean anything. But that’s how a lot of people define to Google what success means.
And then you get all these automations coming together. And how do you find, like, the most conversions for the cheapest price is by giving you the conversions that nobody else wanted to buy?
The phone calls nobody else wanted to buy. The lead form fills that nobody else wanted to buy. The clicks to your shopping site from customers who return 100% of the time, most of 100% of the stuff you sent them. Right. That’s what makes stuff cheap.
And if you’re not, like, measuring that and reporting that back to Google, that’s a huge fail. And so what we have to think about much more is how do we steer the ship at the periphery? So how do we have a better data feed?
And RSAs, I think, is kind of a feed format, too, because instead of writing ads now, it’s like, well, here’s my list of calls to action. Here are my unique value propositions, and here are my brand headlines. That’s my feed. Here you go. Go at it.
But we can still optimize how we position our brand. Like, what is our unique value proposition and how do we communicate.
Julie: It’s harder for smaller organizations. I mean, all the stuff that you’re talking about there, Fred, it’s like, oh, yeah, that’s amazing, right?
Like, yeah, send the value back, figure these pieces out. Make sure that you’re clear inside of your organization. What is the quality of leads that are coming in and trying to put different values on different types of conversions?
That is such a bridge too far for so many organizations. This is where I think this gets really interesting. And this is sort of what happens as these platforms and these ecosystems evolve and mature over time.
They were a very level playing field. And you had a lot of opportunities for organizations of all types and sizes to be able to find success. And as happens as most platforms mature, then it becomes much more stratified.
And I would argue that Google is pretty stratified at this point as far as a lot of the things that the way that the system is built and the way that it’s clearly moving definitely favor larger organizations, retail organizations.
And then you have this whole other swath where figuring out it’s getting them to figure out what that conversion actually is. Or is there more than one type of conversion, and is a phone call more valuable than a form fill or whatever? Right. Like getting them over that line, it’s like, oh, my gosh, that was monumental.
So there’s a large amount of stratification that is happening, and I think it’s going to continue to happen as all the stuff…
Fred: Let me argue the opposite of that a little bit. And I’m the former Google Ads evangelist. Right. So I’ve clearly drank the Kool-Aid. And I need to defend I know how to defend Google’s position on these things.
But from my perspective, I think today it is easier for a novice Advertiser who would have tended to screw things up in the past to not screw it up quite as badly. Right. At the same time, I think those mid-level organizations, they just need to work a little bit harder to achieve that same level of success or a little bit more success than their competition.
And that’s where I think they need us. Right. They can’t go at it alone necessarily. They need expert help. And I think that’s really good for us in this field because as much as it’s become easier, it’s actually also become much harder because the problems we’re solving are no longer problems like keyword selection, but like business value.
And so we become a little bit more business consultants who know how to connect those pieces back into the ad system. Right. If you explain to someone how do you value a phone call, but the math is not that hard. Right?
And you just need to look in your books a little bit to figure that out. But then how do you communicate back to Google? That’s hard.
Luckily, that’s becoming easier. So Google is working on things like GCLID-less conversion tracking. It’s basically the next version of enhanced conversions where you no longer need a GCLID to be put into your CRM, which is really complicated right now. You can just use an email address much easier.
Ed: It’s about time, Fred, that PPC managers, agencies and freelancers became more business-savvy. And people are going to hate me. And I’m divisive probably saying this, but there’s your angle, there’s your edge.
If you know that your clients or potential clients are struggling with this connection between Google and their data, that should be your foot in the door. Not the PPC, not the ads, because every man and his dog does Ads and SEO and so on.
But to actually connect those dots for them, you’ll be their best friend for life. Because then all you got to do is turn on Performance Max.
I’d love that. If that was your book, Fred, it was just completely blank pages, apart from creating a Performance Max campaign. Done.
Fred: I think that’d be hilarious if people paid me money for that book. So, yeah, let’s try it.
Google does see this, or at least the way they position it is an additive campaign in addition to your search campaigns. It will replace smart shopping, but it doesn’t replace anything else.
It’s additive in many ways. But yeah, we’re getting very far afield here from ads, and I think we’re basically all saying the same thing, just different.
Ed: There’s an opportunity, I think.
Here are some RSA tips & tricks.
Fred: Yeah. Hey, so let’s talk about ads. Any other tips and tricks, Ed, I think. Were you talking about multiple RSAs in an ad group or how many headlines you should ideally have?
Ed: Yeah, it’s probably my fault. I know we had questions for this, but I don’t think we followed the script. I pulled some data again just to be Fred like, it’s good to be nerd data.
So essentially what I found from the data is sweet and it’s so difficult to give a one size fits all. So please, if you’re watching this bit of critical thinking, but the sweet spot seems to be around six, seven headlines unless you can genuinely say something unique.
Different. But yeah, six or seven headlines. It’s like combining two ETAs into one.
Fred: So when you say 6 or 7, do you tend to have multiple? Do you kind of classify it into types of headlines?
Ed: Yes. You’ve got probably 25, 30 different types of headlines. It might be a call to action, it might be a feature, it might be a benefit, it might be a location-sensitive thing.
So you have to compartmentalize them into different assets for your ad. What are they? Are they mentioning price? Are they mentioning delivery? Whatever. So obviously then you need to use a bit of common sense to say, Well, I only really need one call to action.
I could test two. I only need one feature benefit. Or I could just test features. I could just test benefits. That’s what I meant in terms of six or seven. Because writing 15 ads headlines, essentially for one bucket of queries is no mean feat to write 15 good headlines, not five good. And then ten shite.
Fred: Shite, you’ve only sworn one so far. You promised many more, Parker.
Ed: I nearly made it for an hour. But yeah, you get the point. So it’s quality over quantity.
If you try and game the system with three or four headlines, it’s very clear that we’ve touched on that. That doesn’t work. But six or seven seems to be at the okay range in terms of where impressions aren’t cut.
But I can see Google moved the goal post on that, unfortunately. But currently, it seems to be where you’re not putting the naughty corner.
Fred: Julie, what advice do you have as far as ad variations and other tips and tricks?
Julie: So in that presentation, I know I’m talking about this a lot, but it was very eye-opening.
As far as what is Google going to be pushing on us here in 2022? They said very clearly that you should only have one RSA going forward just to be like a single RSA in your ad groups.
Part of me can see the logic of that, right? Like, the idea is, okay, you’re going to give us all these assets and we’re going to figure it out and we’re going to serve the best ones, and then you should just not worry about it.
That clearly is Google’s point of view on all of this. I think that again, you have to consider how do you want to test things inside of an RSA? So do you want to create your six, seven, eight headlines?
Let’s say you let those run for a while, see what happens, see which ones are getting served the most, see which combinations are getting served the most, that type of thing. And then when you want to try something different, what will your method be?
Will you add a new headline? Will you create a new version of the RSA that maybe keeps the top three headlines that performed well? And then you’ll create another three or four. You have to think about what you want your methodology to be because again, it’s different.
Like, with ETAs, we would be like, all right, I found one that seems to really be resonating and doing well. Can I write one that’s better? But you don’t have those same capabilities in the same way inside of an RSA.
Fred: I mean, I totally agree with what you’re saying, but for us to do what Google wants us to do, they need to be more forthcoming.
Like, great that you tell me the impression data, but what about conversions? What about costs, right? If that’s not in the picture, I can’t make that decision about what are my best three headlines to keep and maybe vary some of the others.
Julie: I mean, it’s very clear that they don’t want you making those decisions. Right? Like, they want to make suggestions to you and they want to steer you very much in the direction of like, this is what we think you should do.
Now, do I think that that’s what’s going to be best for advertisers? No. Because a lot of times what Google wants you to do, you look at it, you’re like, are you joking?
You read the suggestions and you think like, oh my God, this system really has a long way to go before I would feel confident putting that much trust into this automated system when I see what it’s suggesting. They did also say that there is going to be “more reporting data” available around RSA that is coming at some point in 2022.
Now, what that means, what that data will be, they didn’t get into any specifics. So huge question mark there.
Ed: To be fair, when you’re spending other people’s money, which is what Google is doing, why they haven’t produced that before June, before the cutoff is a bit disgraceful, really.
Again, to touch on what you’re saying, Julie, with testing, I think what I prescribe in God Tier Adsis a dirty A/B. So essentially you run ad variant tests, one RSA, and I agree with you on that.
One RSA, one ETA dies off anyway, typically, and then just run an A/B because you’re forcing Google’s hand to at least try and deliver the impressions fairly across your RSAs. But test one thing in the RSA.
So I call it a dirty A/B because you’re testing a call to action, for example, and doing it on one ad group is probably not ideal. You want to run a theme. So, test the call to action across ten RSAs.
The reason I don’t create a second ad in the ad group is that the experiment works and it gives you a printout of the metrics. So you do actually get to see the A/B result.
And I know it’s not a true A/B, but it’s probably the safest way to do it at the minute, I’d say.
Fred: So I think what we’ve seen advertisers do successfully is a variation on that. So basically, think of so get rid of ETAs. Now run multiple RSAs in one ad group, but think of each RSA as having a theme. Right.
One may be focused on price, one may be focused on the speed of delivery, whatever other value benefits you have. And then don’t stick to just one headline variation that talks about that. But try three, four different variations. Right?
So to give Google some flexibility. But there’s a very clear theme between the two. And then you can start to see, well, does Google prefer one or the other? And then you can start to figure out, okay, maybe this theme is working better.
Now let me pin those thematic points to different positions. And now I run one RSA. Both RSAs are the same theme, but they’re pinning two different positions.
And now you can start to kind of A/B test. So there’s still a process that you can go through even before you get these new metrics where you can just see the end results of the whole entity as opposed to the individual components.
Ed: Yeah. And Ngrams for Ads, you do it. We do it because it makes sense. And I think Ngrams for our stay is useful. Certainly nice if you just did it for us. But yeah, I think the thematic approach is a good idea.
I think the smaller advertisers, smaller budgets struggle with that and might need the dirty A/B. Yeah. There are still options available, so it’s not the end of the road.
How to use ad strength and asset labels when optimizing RSAs?
Fred: So we’re coming close to time here, but there were two more things I wanted to talk about and get your take on.
So ad strength, whether you use that or not. And then the asset label, which is sort of the asset strength. How do you use that if you do currently?
Ed: Julie, do you want me to go?
Julie: Sure, go ahead.
Ed: I didn’t sound very enthusiastic. Did I? So again, I was just pulling data. I don’t know if you agree, but looking at 2019 to 2021 data, I don’t see any obvious correlation between the top of page rate versus the ad strength.
Fred: Yeah, and I agree on that. So ad strength is to explain to people ad strength is something that Google looks at based on best practices of what you submit that ad strength will never change based on how the ad is actually doing.
So there should be no correlation. Right. It’s just what the machine thinks is going to happen. It’s not what it never updates its thing.
Ed: Umm, but the impression share is different. So the top of page rate doesn’t seem to flux. But the ad strength does impact impressions.
Fred: Yeah. So Google has a correlation study, and I think they say for every point you go up and ad strength, you get 3% more impressions to the highest. It’s like 9% or 12% boost in impressions. Yeah.
So what you need to look at is the asset labels, because those are actually based off of historical performance. But the downside on those is that you don’t really get these until you have a fairly high volume.
Ed: Okay. Impressions and then they tell you an arbitrary thing. It’s good. Okay. Thanks for that. I knew it’s good. How good?
I just feel like we’ve got a two-tier quality score system going on, even though it’s not about ad rank, but it is because it impacts it. So I wish they just amalgamate the whole bloody thing and just got it over and just give us an ad score, rename quality score, whatever.
I think it’s tripping a few people up, to be honest with you.
Julie: Well, pinning comes into effect for that too. So when you pin things again, out of my own curiosity, like I said, I’ve been running the experiment with literally the exact same assets, but with things pinned, and then I have one that’s totally unpinned.
The unpinned one has a higher ad strength, like right from go, because you don’t pin anything. So as soon as you pin stuff, you’re dropping your ad, you’re dropping your ad strength right from the get-go.
And if you get it into it doesn’t take long to get to a poor ad strength, even if you had one that was average or good before you started pinning stuff. And then you get that little what does it say?
Like you get that little eligible and then in the parenthesis limited status where it’s like, well, yeah, we’ll show it, but we’re not going to show it as much as we would if you would make this better, except all they want you to do to make it better.
Like you could literally do nothing other than just unpinning stuff to make it better. So there are some areas of the system that I really hope are going to be further developed.
Fred: This is probably a software enforcing loop on the machine learning system, right? So it’s basically looking at what you did and it knows that system-wide bidding is going to hurt your performance.
So it makes a prediction the moment that you pin and says, well, you’re probably going to perform worse. But then it also says, well, I’ve got these million other advertisers for the same keywords and they didn’t pin.
So I’m going to give them preference because the prediction is they’re going to do better. So then the question is how flexible is Google at allowing some variation from the established best practice so that the system can learn?
Maybe, hey, maybe things have changed, right? Maybe pinning is okay in certain situations, but I think that’s where we’re fighting just a prediction mechanism that has made some large decisions that it’s not willing to change anymore.
Ed: And if you’re on a mid or large-size account, you can’t ignore this because you got to lose volume. So you’re going to lose auctions if you don’t comply, essentially.
So if you’re a volume play, if you’re an insurance company where you just want volume overly, you have got to reference this arbitrary score, throw spaghetti at it, and it patronizingly tells you you’ve created a great ad.
And it’s like, well, I didn’t really do much to create that great ad. It needs some tinkering and needs some work. Certainly, ML is all about iteration and it will, fingers crossed, improve, will it?
Julie: I was going to say it’s interesting looking at what it chooses to put together for you, too. You can see the combinations that got the most like it’s in descending order. So the combinations that got the most impressions.
I find that really interesting to look at that’s giving you an insight into obviously it’s showing the ones most frequently that it thinks will be most successful.
So I also think that that’s an interesting little area to keep your eye on too so that you get a little bit more insight into it because it’s one thing again, we’re used to looking at ads as a whole for ETAs, right.
And you know exactly what was in the ad, and you can see what happened, like what was the percentage that it was served and all of the stats for that.
But with the RSA, you don’t see those stats in the same way. So I find looking at that data where it shows you the combinations that can give you some insight too, for each particular client and each particular thing that you’re trying to advertise for.
What does Google think are the most important aspects?
Right. And again, when you look in there, it’s really interesting because again, they favor certain combinations a lot more than others.
So I think that’s really interesting to look at and start to sort of digest as well when you’re thinking about what you want to create and what kind of things you might want to test because you’ll find assets that they never pick ever.
So you can look in there and see that and be like, well, this headline in here is doing nothing for me, or they never serve this particular description. You might as well put something else in its place.
Right. Because those assets that are sitting in there, never getting served, are doing nothing for you. So I think that’s an interesting place to kind of us as humans, right? We’re talking about how to still have where do we have value in the system where everything is automated to a certain degree?
I think looking at that particular set of data, lets your human brain think about some things that, like the machine can’t think about as well.
Wrapping it up
Fred: I love that. I think let’s wrap up on that. Great advice!
Ed: I’d like to add one thing. Do you know the frustrating thing about looking at the combinations, seeing that one headline isn’t getting any impressions, then removing it from the RSA and your ad score dropping, saying, oh, you’re out of rubbish now it’s like, oh, you right?
Fred: And I think this is really important for people to understand. The ad relevant score is a prediction. It has nothing to do with your performance. And the asset labels are based on your performance. Two completely separate systems, so they don’t talk to each other and they’re going to tell you stupid stuff like that.
Ed: Good point.
Fred: Yeah. Well, this has been a great conversation. I feel like we need to do another one as this topic evolves and as more metrics hopefully become available, a final wrap of anything that we didn’t cover or just tell us where we can find you. Remind us. Let’s start with ladies first, Julie. Yes.
Julie: I mean, this is definitely a topic that we’re going to be talking about, I think, all through 2022, so I’d be happy to talk about it again anytime.
As for where you can find me, I’m really active on Twitter, so that’s the best place to find me.
My Twitter handle is @NeptuneMoon. I am always active on the #PPCchat, so you can find me in either of those places. And I’d encourage you to check out. We have a ton of resources and links to past chats and that type of thing on the PPC chat website as well, which you can find at officialppcchat.com.
Fred: All right. And I know your dinner is getting cold, so wrap it, Ed.
Ed: So, me, I’ve got a really awful load of websites. I won’t bother. I’m not very active on social media, but if you really want to find me, go to Edleak.com.
It’s a rubbish website, but you can find your way to my other stuff from there. That’s it. Really. I mean, I don’t have a lot more to add. I won’t get another invite.
Fred: Julie should we invite him again?
Julie: Yeah, I think so. Yeah.
Ed: Thanks. All right.
Fred: But only if you meet me in London.
Ed: Hey, I want a copy of that book. Did you say two books I got two books?
Fred: You still didn’t get the first one. It was like back in 2019. Anyway, bringing up the book so folks ‘Unlevel the Playing Field’, it’s on Amazon.
Ed: Where’d you find the time?
Fred: I’m efficient, I guess, but I hope the book is good. We did a couple of iterations of it. I wasn’t very happy with the first one I wrote. Kind of went back and made it I think, really good.
It talks it’s really about this evolving nature of automation in PPC and how we deal with it. RSAs are one great example.
So anyway, thank you both for being on. Julie, I think we’re going to hang out in Austin at Hero Conf.
Julie: No, sorry. I’m not going to be there with you. It’s a bummer.
Fred: Well, anyway, it’s been a pleasure having both of you on folks.
If you enjoyed this episode you want to see more of them. We’re doing them about twice a month subscribe to the YouTube channel. That way you’ll find out.
And if you have any questions for me or Optmyzr, I’m @siliconvallaeys on Twitter and you can always find me Optmyzr.com.
So thanks so much for watching. See you for the next one.
Connect with the panelists.
- Do not pin assets in your RSAs. But if you absolutely have to then pin multiple texts to a single position. This gives you control over what the ad says and gives Google some flexibility.
- Provide as many headlines and descriptions as possible to let Google generate the maximum number of ad variations.
- Learn to become more business-savvy because with automation taking care of the basic stuff, the success of your campaigns depends on your expertise.