The Right (and Wrong) Way to Grow with Google App Campaigns — Ashley Black, Candid Consulting

David Barnard:

Welcome to the Sub Club Podcast, a show dedicated to the best practices for building and growing app businesses. We sit down with the entrepreneurs, investors, and builders behind the most successful apps in the world to learn from their successes and failures. Sub Club is brought to you by RevenueCat. Thousands of the world's best apps trust RevenueCat to power in-app purchases, manage customers, and grow revenue across iOS, Android, and the web. You can learn more at revenuecat.com. Let's get into the show.

Hello. I'm your host, David Bernard, and my guest today is Ashley Black. After 6 years leading the app ads team at Google, Ashley left Google to found Candid Consulting, an agency that helps apps run effective marketing campaigns for Android and iOS. On the podcast, I talk with Ashley about what makes Google app campaigns so effective for growth, proven optimization strategies, and why lower CPMs can sometimes be a warning sign rather than a win. Hey, Ashley, thanks so much for joining me on the podcast today.

Ashley Black:

Happy to see you again.

David Barnard:

We're going to be talking about Google app campaigns today, and I wanted to kick things off with your experience because you do have a kind of unique perspective compared to a lot of folks that we talked to on the podcast. So, why don't you share what you've been doing the past 10 years and then what you're doing now?

Ashley Black:

Yeah. I think there's a lot of UA experts out there and a lot of people obviously who work in apps. I guess, my specialty lies within Google app campaigns, their ads product. For the last year, I've been out on my own. But prior to that, I spent nine and a half years working at Google. Six of those years, I led the app ads sales team, so my team worked directly mostly with game developers, so the marketing teams within those games. I guess I have a lot more inside knowledge as to the product and how things actually work, so my perspective tends to be a little bit more unique.

David Barnard:

It's really cool that you were in Google for so long. And then, a little bit of the backstory that you've told me and shared publicly is that it was frustrating at times to work with customers where you knew certain things that could maybe be helpful but weren't really appropriate for you to share as somebody working at Google, and so you did start a consulting company where you can say those things and maybe even be more directly helpful to the customers.

Ashley Black:

Yeah. And this happens at any large company where you feel that there's a lot of processes that you have to go through in order to actually be able to say anything, or there has to be so much data that's collected in order for Google to say, "Okay, this is how we can position this product." I always felt really stifled by that because people just want to know and, I think you guys understand better than anybody, people want to test things too. So for us to have to wait years, in some cases, to be able to communicate to people what we feel like maybe we already know but product or legal hasn't given us the stamp of approval to do so, I found to always be really infuriating. Now, I offer a lot of those hot takes that I wasn't allowed to do before, but always with a grain of salt. It's testing at the end of the day but, yes, I do have. I've seen thousands and thousands of these accounts, and so I probably have a little bit more credibility when I make these statements.

David Barnard:

Yeah. I was just talking to Sharina, and it was actually the very last episode. For those of you who didn't listen, go back and listen to that. One of the things we talked about quite a bit in that episode was how important it is to take risks, and how important it is to foster a culture of taking risks when you're doing user acquisition and trying ads because you have to take some risks, you have to try some new things. You have to kind of go out on a limb in order to find those things that are going to double your performance. And that's one of the things she talked about. They were able to double their performance over the course of a few years through hundreds of tests, and so it is kind of cool to hear that.

When you're at Google, it's hard to tell people to take a risk. It may lose them tens of thousands of dollars. That's not what Google wants to do, they want to only share the tried-and-true strategies. It's interesting to hear that one of the bigger frustrations was that you couldn't help people take those kind of risks, but it's those kind of risks that really do optimize their performance over time.

Ashley Black:

I think one of the things that I felt most passionate about, and it's really how Candid Consulting came to be, was on the iOS front. Google has a very complicated relationship with Apple, and so we were so limited on what we were able to say when it was in regards to SKAdNetwork and how Google's model was working and what we were rolling out. I would always advocate internally when I was there to say, "Hey, people just need some bit of information." I know that we can't say definitively this is what we're doing or we may have to pivot later based on what Apple changes, and people I think understand that. What people don't like is silence. And that's what I was really frustrated by a lot of times was just the silence because there was all of this back-end complications that honestly I didn't always have visibility into.

They were happening I think a lot of times with the engineers and the product teams and people very high up, but I was like, "People just want to know something. Can't we just tell them something?" Because I can tell them little things but it was like, "No, no, no, not yet," and so leaving has obviously enabled me to be able to share those insights.

David Barnard:

Let's just kick things off from the highest level overview then of what is Google app campaigns and why should this audience of subscription app growth practitioners, generally. We have indie developers and CEOs and a broad audience of folks who are going to be listening to this, but why should people be thinking about, care about, or what should they know about Google app campaigns?

Ashley Black:

If you have an app and you're starting to think about how you actually pay to get users to download that app, you're probably going to start looking at Google. Although, I find a lot of people start with Meta first. Or if they're heavy on iOS, they'll start with Apple Search Ads. But typically, Google is first or second when they're looking to explore. Google app campaigns, if we take the history of it, I want to say it was probably 8 years ago now, maybe 9, where Google actually this was their first machine learning ads product. It was very novel because... Well, people did not like the idea that Google was taking all of this control. They were basically saying, "Tell us how much you want to spend on a daily basis. Tell us how much you're willing to pay for whatever it is, and install..." I think at the time, it was mostly just installs. But now, it can be how much you're willing to pay for a subscriber or somebody to start trial. Throw in some creative, and we got it from here. You don't need to tell us anything else.

What's interesting is that's the direction Google has moved with a lot of its products. I'm not as familiar on the website of things, but people talk about PMax and I think you're starting to see an expansion of this. And the reason why is it works, it really does. At the time, Google's app ads product contributed nothing. And by and large, in comparison to search, it's still a very small percentage of their ads revenue. But it has grown massively, and so the product does really work. It started out as UAC, universal app campaigns. That's what most people call it because it's easy, everybody loves an acronym. They dropped the U a couple of years ago, so it is just app campaigns. Basically, it allows you to serve across all different inventory types, but you don't have control over where your ads serve-

David Barnard:

We can get into some of the fun little hacks to get a little more control. That was one of the frustrations when UAC came out was that it limited so much control, but that trade-off was the performance and we're see Meta going that direction as well. I'm giving you less specific control, but then the algorithms have just gotten so good both on Google, Meta, and other ad products where the algorithm is going to end up making better decisions and that's why they did it. So, what would you say are some of the key benefits then of Google app campaigns? And then we can get to some of the drawbacks, and we've already discussed some of them.

Ashley Black:

Whenever I talk about Google, you really have to talk about it for Android and for iOS. I feel like with Meta people, they don't really care too much. I don't notice people separating the two and talking about them separately. But I feel with app campaigns, or I'll just say UAC, it is really critical. Because on Android, you have the benefit of actually serving your ads in the Play Store. That's super high intent. You cannot get that inventory anywhere else, so that just doesn't exist. You can't buy it, it's not like part of another network, so I think that's really the main advantage of UAC when it comes to Android. Also, there's a lot of other properties, you have Google Search and YouTube, so there's a lot more that you can leverage there too. But by and far, the Play Store is huge.

On iOS yet, Google's actually not as competitive, but I really don't feel it can be overlooked. You still have Google Search, which again, you can't buy search from anywhere else, and you have YouTube, which still is the number one app. It's like the most downloaded app in the world I think, not just in the US. Those two pockets of inventory are really... those are huge for you to be able to access and be able to advertise your app.

David Barnard:

This is one of the things about Meta and why a lot of apps are finding success right now is that between the Facebook app and the Instagram app, you're just getting such broad exposure. But then Google takes it even multiple steps further, like YouTube, inside the Play Store, across Search, across their Display ad networks. It is just a massive scale that is hard to access in any other way. What would you say then are some of the drawbacks that come to mind with UAC?

Ashley Black:

Going back to the controls, it can be really challenging if you're testing in particular. If you're like, "Oh, I really want to see how these videos perform," on Android, it is really challenging to try and get your videos to get any meaningful traffic. People can do it, and it depends on the vertical of your app. But again, you have the Play Store, it's going to be really high converting traffic. I always liken the Play Store traffic to branded search. When we talk about when you used to have your branded keywords, it's like you're reaching people at the highest intent moment possible when they're actually searching for something or they want something. And that's kind of what the Play Store is, too. They're grabbing them right at that last minute.

Being able to really diversify where your ads are showing, that can be a big challenge, or the opposite of trying to force and get traffic somewhere else, getting in a way. I was talking with a client yesterday, all of their traffic's flowing to Display and we're like, "How can we force it over here?" Aside from really paying a lot more, like manipulating your bids, which isn't always feasible, that can be a big drawback, and so I think a lot of people despise Google's products for that reason.

David Barnard:

But as we were kind of pointing out, it's is all trade-offs, so you trade off some of that control for reach, for the high intent, things like that. We've kind of been hitting on this already, but any other contrasts you want to paint between UAC and the other major kind of user acquisition channels, TikTok, Meta, Apple Search Ads?

Ashley Black:

Oh, yeah. I think one thing that people look at is... A lot of people are like, "Oh, it's an app network. We're using that product to distribute our app," but you have to think about it. Let's just use Meta as the primary example. Their two main sources of traffic are Facebook and Instagram, and most of that is app specific. If you're doing an app campaign, it's just going to serve there. On Google, let's talk about the five primary sources of traffic that you get with your Android campaign. You have the Play Store. You have Google Search, which is mostly done in a browser. It's not in an app, so it's in a browser. You have YouTube, which is probably pretty heavy on app, but there actually is still quite a bit of it that happens in a browser.

You have AdMob traffic, which is ads in other apps. That's part of Google's network. And then you have what they call mGDN, which is mobile Google Display Network, which is ads in browsers, on websites, mobile websites. A lot more of Google's mobile traffic is served in browser in comparison to these other app native platforms like a Meta or a TikTok, that just doesn't exist on web. You have to really consider that, and that's the piece that you really have to understand, especially when it comes to iOS. That's why Google, I think, had so many challenges for a long time. SKAdNetwork wasn't even counting web to app conversions. We had cases of some clients where 99% of their traffic was coming from Search, and SKAN wasn't capturing any of those conversions, like zero of them. Now with the SKAN 4.0, that has changed. But there's still a lot of limitations, and people realize how unique Google's platform is compared to some of the other big players.

David Barnard:

One of the things you had shared previously... We did a webinar and, by the way, the webinar might be a good adjunct to this if you want to get back and listen to Ashley's webinar shift slides and all that. So, it's kind of a nice complement to this conversation. But you also mentioned Play browse. Within the Play Store, you can actually buy ads where if people are browsing around looking for apps, they're going to see those ads. Is that serve much traffic, and does that perform well?

Ashley Black:

Nobody knows, David. That's the thing. That's the other reason why people don't like this is because you really limited visibility into where your ads are showing. In the Google Ads, like in the front end, if you're looking and you're managing your campaigns, you'll be able to see a breakout of the four different inventory sources. They lump it as Traffic Search Network, Search Partners, which usually is a really small percentage. If it's high, that should be a warning sign to you. And then there's this other one called Google Display Network, and that's what we'll talk about in just a sec, and then there's YouTube. But again, what people don't realize is what's included in those. So if we're talking about Search, yes, it searches on google.com but it's also searches in Play. Well, you don't have a way of seeing the breakout if you have a really good... I always tell people, "If you have a really good Google rep, you can." They do have the ability to see it, they have to have the right access, but they can show you the breakout.

There's google.com searches, there's Play search, and that's all included in Search. And then when we talk about the Display Network, that is the AdMob traffic a lot of times, it's the mGDN, which I was talking about is website images or ads, and then it also can include Play browse. That's the tricky part is you're trying to figure out, "Okay. Well, is this not working because of this, or is this happening..." You're making a lot of assumptions about things and so, a lot of times, it can be hard because you're a little bit blind to what's actually at play.

David Barnard:

But overall, the combination of all these things and Google's algorithm that tries to figure out where things are going to be performant, it does seem like overall it is a high quality source of, and especially on Android, potentially very high intent traffic as well.

Ashley Black:

Absolutely.

David Barnard:

Even though you do have to depend on the algorithm to sort through, there is really high quality sources of traffic that are flowing through UAC.

Ashley Black:

Oh, yeah. I don't think that that's undeniable on Android, like we were talking about, and also on iOS. You have Google Search, which again, I just don't think people understand. You can't buy that anywhere else. You can't get those types of users who are actively querying for product that you have anywhere else, and so that can't be ignored. And like you mentioned, the sheer volume of people. I know a lot of people, you can get good scale on a lot of these other platforms. But I do think a lot of those other platforms have demographics and audiences that they may be skew a little bit more toward. Where I feel like if you think about Google and all the products that we just listed, you're likely to find a little bit of everyone across all of these.

Everybody's looking at YouTube, at some point. Everybody uses Google Search. If you have an Android device, you're sort of required to use the Play Store. We don't have to get into those requirements, but you know what I'm saying. It just feel like it's something that can't be ignored, and there is a really high-quality user that exists out there. You just have to put a little bit of faith into the system that it will find it for you.

David Barnard:

Yeah. I think this is a good message for the audience that even if you're finding incredible success on Meta and doing really well, at some point, you really should be looking at UAC because it's going to give you access to these sources that are high intent and can potentially be really well-performing ads that you just can't get access to those people in that traffic any other way. So, you definitely should be considering UAC even as other networks are performing really well.

Ashley Black:

One thing I want to talk about, and I think we sort of discussed this offline before, is I find that a lot of people like using the other networks because of the testing capabilities. I feel like in my LinkedIn influencer world of people who are into user acquisition, it's a niche group but they seem to be very active, there's so much conversation around creative testing. And people are developing 900 videos a month to test out on Meta or on TikTok, and nobody really talks about Google in that conversation. I think sometimes people like that control, they like the ability to test, which we've talked about as well. That testing environment is exciting, you get to see what's going to be produced. And that doesn't exist to the same extent on Google.

I think if you're really savvy and you kind of know the right levers to pull or what sort of data to look into, you can force some of that testing. But it's not as obvious to be able to run a side-by-side test between two videos and see which resonated better with an audience. I think people are looking for that a lot of times when they're going to market their app, and that can be a real challenge to do with Google Ads.

David Barnard:

Speaking of that, the next topic I wanted to talk about was measurement. So, how is Google app campaigns measuring the performance, and how should people be thinking about that measurement when they start considering or start running ads? And then we'll get into all the tips of how to run in-app strategy, and stuff like that.

Ashley Black:

Sure. Again, this is another example where I very distinctly call out the difference between Android and iOS, because they're night and day. On Android, Google still has the GCLID or GAID. I always forget which is which. But they still have the ability to track at the user level, at least for now. I know there's things that they're rolling out. But most of those conversions that you're seeing in Google ads are deterministic, meaning they have confidence that those actually happened because they can see that they did. So if you're using a platform like an MMP to measure your performance, generally, you're going to see that the measurement there, what Google shows you and what your MMP is reporting, it's pretty close. We can talk about discrepancies, but the discrepancies are usually a lot smaller when we're on Android. iOS is a whole other ballgame.

Google introduced modeled conversions. Oh man, people hate modeled conversions. This is Google, they don't tell anybody about it. I was there when they rolled it all out for iOS, and I feel like I got nothing from how they actually were modeling these things. They talked a lot about stuff, but it wasn't all that practical. I couldn't really understand how it was going to roll out. But the modeling is Google basically saying, "Hey, we have confidence based on all of these signals that we have that this conversion happened." And by the way, there's very few deterministic conversions that Google actually has. I'll touch on that maybe in a bit more detail in just a minute, but Google models their conversions. Then if you were to go into your MMP again and say, "Okay. Well, how many conversions did Google produce?", you're going to see a huge discrepancy. And then when you talk about SKAdNetwork, you're also going to see a really enormous discrepancy there. So, Android's pretty reliable in the measurement. iOS, we could do a whole other podcast on that.

David Barnard:

Is Google using SKAdNetwork for now, and are they capturing those web to app or... I mean, the kind of dirty secret of the whole industry at this point feels like is that both Meta, Google, and everybody else doing any kind of measurement is using the IP. And so the quote, unquote, and for those of you listening in audio, we're both doing finger quotes, "modeled conversions" is probably pretty heavily reliant on IP address, fingerprinting, which then can be fairly confident, but then you're going to lose some of those signals. But then, SKAdNetwork should be more deterministic, but just not on a user level. So yeah, do you have any understanding of how they are and aren't using SKAdNetwork to help model those conversions?

Ashley Black:

Just to be clear, on SKAdNetwork, Google, at least I haven't heard anything changed in the last 6 months, does not ingest any data from SKAdNetwork back into their models. The only thing that they ingest back from SKAdNetwork is the installs that happened, nothing deeper. And I think it is relatively complicated, right? You have to figure out what all of those conversion values mean per each individual advertiser. There's not a standard set of them to be able to say, "Okay, did this happen?", so Google's not really using SKAN in, its models. Now, the fingerprinting thing is separate. They just announced as of recent that they will start doing this, the IP. I shouldn't say fingerprinting. I don't want to get in trouble, but-

David Barnard:

Well, all of these big companies are doing fingerprinting. It's a bad word in the industry, but Google and Apple aren't cracking. I mean, Apple specifically, especially post-ATT, is not cracking. They say in all the rules like fingerprinting is not allowed, but then everybody's doing it and nobody's getting rejected for it.

Ashley Black:

Yes, they did announce that they would be using IPs as part of it. Now, whether they had been all along and now they're just being forthright with it, I don't know. We'll have to see. But I think the one thing that I wanted to call out... because I get calls all the time. I do audits for a lot of people reaching out and they're like, "Hey, just tell me what's wrong with my setup," or, "We have questions," and so they bring me on to do stuff like that. One of the big things that people are confused by is the deterministic conversions. I always liken it to pre-iOS 14, what the environment used to look like. Again, this goes back, David, to what I was talking about where you have to really think about Google and all the different inventory, all the different traffic sources it serves on in comparison to some of the other players.

One reason why you might see Google with zero conversions in your MMP deterministic conversions, or just a fraction of what Google is reporting or what SKAN reports, is deterministic conversions can still happen if you're able to collect the user's IDFA. But again, remember, so much of Google's traffic happens on browsers on websites and the opportunity to collect user consent, the ATT prompt, does not exist on browsers. I think people really overlooked that. And then probably the biggest one, and again, this is one of those things that I was talking about where I wish Google had been more vocal about it, is that they did not prompt for it on YouTube. That was their biggest source of app traffic on iOS, and they were not collecting those users' IDFA for a long time. They launched it finally back in July of 2024, started to roll it out, and actually prompt for consent on the YouTube app.

But I noticed even still, Google does phased rollouts of everything. They'll launch it to 1% of users in one country, and then they slowly bring it out. They launched that in July, and I personally didn't get the prompt until December to opt in. They hadn't even shown it to everybody. Anyway, that's all to say that, Google, because of their traffic differences and the choices that they made back in 2021, people don't understand why those decisions had a big impact on the measurement side of things.

David Barnard:

Yeah, that makes a ton of sense. Well, I did want to shift gears now. I think we've given a pretty high level overview of all this stuff and gone into weeds on different topics. But I think what a lot of people want to know at this point now is like, "Okay, I'm convinced I should try UAC. How do I do that?" I think that the first question most people ask is budget, it's like, "If I want to start spending on UAC, what does that look like? How much do I need to spend?" One of the things we didn't mention earlier is that it does sometimes take longer for UAC, for the algorithm, to get trained, so sometimes it does take higher budget. But yeah, what are your thoughts on getting started?

Ashley Black:

I'll be consistent here. I've said it somewhere else. I know people come at me and say, "Well, that didn't work for me," and they're like, "Do I really need to spend that much?" I'm going to use this as if you're starting in the US, okay? Because if we're talking about other countries, then that becomes much more complicated. But in the US on Android, if you're saying, "What is the minimum budget? I need to start testing things," I would say anywhere from two to 300 bucks a day. Yes, you could do less and sure you can do more. But if I were to say, "Okay. You're just starting out, that's what I would probably want to spend on a daily basis to get enough information, to get it fast enough to actually be able to say, 'Okay. Is this working, or what adjustments do I need to make?'"

On iOS, it's got to be more significant than that. But it's also a more expensive platform. Installs are a lot more expensive on iOS than they are on Android. The inventory is more expensive, the CPMs are higher, so all of those things need to be taken into consideration. So on iOS, again, I'm just saying this as a figure, I would say at a minimum about 500 bucks a day to see if you can get enough traffic, and then actually have something to do with that traffic to be able to make adjustments as needed.

David Barnard:

Once you start spending that and seeing it work, how do you then scale up from there? What are your tips on scaling budget?

Ashley Black:

Slowly, slow. You don't want to wake the dragon, or whatever it is. And I've experienced this personally where I've always... The engineers at Google were always like 20%, right? No more than 20% in any given day on bids and budgets, up or down. So, if you are looking to scale... I kind of got burned not that long ago where I was like, "I bet I could push it a little bit more." And then I did, and what happens is that the system goes crazy. It's like, "Woohoo." It's like a 10-year-old or something that just got more money is like, "Let's go spend it. Let's go figure it out," and it's like, "Whoa, we have to be cautious. Let's try and figure out new pockets of inventory, and do this in a responsible manner." So yeah, that 20% rule does really apply. I even am maybe more conservative where I only adjust it or I try to only adjust it by 20% every 2 to 3 days. I try not to do it every single day because then it just keeps kicking it back into learning mode at too great of a frequency.

David Barnard:

We'll dive more into campaign management in a bit. But with that budget, would that be on one campaign or can you have multiple campaigns running at the same time with that $500 just getting started budget?

Ashley Black:

That's one campaign. But you really don't need more than one campaign if you're only targeting one geography. So if you just have one US-based campaign, actually the guidance is to not have a duplicate one doing the exact same thing. There's a lot of theories out there that you're cannibalizing your traffic if you do that. And that was certainly what I think when we would submit internal tickets sometimes and say, "Why isn't this working?" That's what we called a G-Tech, and that was the internal team. The Google Tech team would come back and say, "Well, it's because they have multiple campaigns." But that was always, I think, really hard to prove that that was the case. But if you're going to set up your campaigns, you only need one campaign per geo. And then if you do want to get more granular, you could have a separate one if you were doing different bid types. If you're optimizing for installs versus let's say start trials, you would have two separate campaigns because you do have to set your event optimization at the campaign level. That's pretty rare, too. I don't see that happening as often.

David Barnard:

Any other tips as you're ramping up? If you're spending $500 a day, how long is it going to take before the algorithm really figures out where to show your ads, how to show your ads? Any other tips on ramping up?

Ashley Black:

Usually within 2 to 3 weeks, you can start to see how things are going to likely trend. One other thing that I notice as a mistake that a lot of people do is they build out their campaigns as the bare minimum. They will create one campaign, and then they'll have one ad group. And the ad group, if you're familiar with Meta, is like an ad set and it's sort of a grouping of creative that you're testing. Now, you can't assign budgets at the ad group level in Google. Google's just going to make the call and say, "Okay. We're going to send 80% of our traffic to this one ad group and 10% to this other one and 5 to the last one, et cetera," so you have no control over that. But what I always encourage people to do is you have to build out multiple ad groups and theme them.

Think about use cases and or features of your app. And then that's what I would create your ad groups around, and see what kind of engagement you're getting on those. I'm just going to use a golf app, maybe it's swing performance. That's one ad group. And then the next one is about caddies, that's your next one. All of your creative should sort of correspond to that central theme, and then you can kind of see what people are interacting with. I usually like to build out maybe about 5 ad groups per campaign, get a little bit of traction, see what's resonating with folks as opposed to just doing one. If it doesn't work and then you're like, well, "I tried, right?", it was like, "No, you didn't."

David Barnard:

Yeah, that makes a lot of sense. You already again touched on the next thing I wanted to talk about, and that's created strategy. I know there's all different kind of assets that you would put into those ad groups. How do you think about creative strategy with UAC?

Ashley Black:

Like I was saying, I theme everything for the most part. I think another common mistake that I see people make is... Google, not unlike the other platforms, has character limits when you're creating your text assets. Text assets, by the way, are the most crucial part of your Google Ads campaigns. Contrary to maybe other platforms where people aren't even reading, the text asset is just fundamental. Again, if you're thinking about Search, Search only shows text, search assets like, "That's just a text strategy, so you got to make sure that that's really fine tuned." And then even in the Play Store, they will pull in your app icon and then a bit of text that you write in Google Ads, and that's what your asset, the total asset is. I think a lot of people oftentimes overlook them or they use Google's recommendations to put their tax asset. And I don't know why Google doesn't abide by regular punctuation and grammar in those but, anyway, the text assets is really, really critical.

The other piece that I talk a lot about... Again, going back to understanding what Google's traffic is, I think it's really easy to see in Facebook or Instagram that, "Oh, you always get a 15 or 20-second reel." And that's what you see quite often. But I think on YouTube, and you don't have reporting into this, you don't realize all the different pockets of inventory within YouTube. When you're creating videos, I always tell people probably want to 6-second video because that's one ad format in YouTube. You want a 10 to 15-second video because we just want to test out that duration. Around 20 seconds is great for YouTube shorts, 30 seconds or longer is probably going to serve in stream. Those are more of those skippable ads that you would see if you're watching just a regular video. You have to have a lot of variety in terms of links on Google, and then also the orientations. I mostly default to portrait, especially if we're talking about iOS. But on Android, you can't overlook the landscape orientation because a lot of people still do watch their YouTube videos in landscape mode.

David Barnard:

And then, what about image assets? Those are the three primary assets that you provide to Google, right? Within each ad group you're posting text assets, image assets, and video assets, and the algorithm is just figuring it out.

Ashley Black:

Yes, usually combining some aspects of that. Usually, text is always with some component with image or with video. But yeah, images are used a lot. Also, images are used sometimes in the Play Store too. Again, on Android, you'll notice the theme here, we talk a lot about the differences between the two. But on Android, I over-index for landscape orientation images. As of recent, I've sort of taken the stance that I'm aiming for an 80/20 rule for images. 80% of them on Android, I want to be in landscape. And then I'll use the remaining 20% for square or portrait, just because landscape always gets more traffic. Google also has the option where you can... They have the AI generator assets now where you can actually type in text and create your own. One thing I tell people is you don't have to have a fully designed image asset because Google overlays so much within your asset, so they're going to pull in your app icon.

They will, in most cases, pull in a install button. It says close or install a lot of times, and then they even pull in a rating. A lot of times, people don't realize that your app store rating is pulled in to the actual asset too. So if you have a fully-designed image asset with a bunch of text and an image and a call to action, and in your rating, if you're calling that out for whatever he said, it's really redundant because Google's already doing it. I usually tell people, "Use the stock image." They work just as well a lot of times, and they're free. You can just search for them as you're creating your ad groups, and grab a few of those and see what works. You can test it, and by test, I mean putting something else in there and seeing if it gets traffic. You can't run an A-B test on an image asset, so you got to just throw it up there and see if it sticks.

David Barnard:

So, the recommendation here then is you don't hire a graphic designer and have this polish what you would typically perceive as an image ad. Like you said, with a clear call to action, a subtext highlighting your favorite feature, whatever, the image assets should just be very plain, no text at all, or minimal text or-

Ashley Black:

Again, it's not a hard-and-fast rule. I haven't seen the ones that are fully designed, at least for the accounts that I work on. And a lot of my clients, I haven't seen them get as much volume as the ones that are just images. Again, the whole concept is because it's already overlaid anyway, in a lot of cases. If you're finding that those aren't working for you or the click-through rate's really low, then sure, go ahead and design a fully... But I don't think you necessarily need that in order to get started.

David Barnard:

How do you look at the performance, or can you even look at the performance of individual assets in UAC?

Ashley Black:

Yes, you can. But I think there's the difference. When I was working at Google and working with a lot of games, they would get really frustrated that they couldn't see what the combination of assets produced. They wanted to be able to see, "Okay. Well, did this specific text plus this image perform better than this text with the same image?" Google will not let you do that. They're not showing you that level of granularity. But you absolutely can see like, "Hey, is this text asset performing better than another text asset?", and so you can break it down quite a bit. There's actually a performance rating in Google. If you pull up your account, it will tell you Google has a performance rating, how your asset is, was is it like best, good, low, and then learning. Just don't pay attention to it.

I was joking with somebody, I'm like, "It's 5 columns that you can't remove. They serve almost no purpose, at least when I'm optimizing an account." And the reason why is that that performance rating is based on the scale of the campaign and the click-through rate, so the engagement. It's kind of like how much can Google serve this asset, and then that's what the rating is based on-

David Barnard:

Not necessarily how it's performing-

Ashley Black:

Performing, right, like your actual metric that you want to optimize towards. So if you're new and that's how you're approaching it, just go ahead and ignore that. Focus on whatever you're optimizing for your CPI or your CPA, how do you actually look at the performance of it? So yeah, you do the asset reporting. I pay a lot of attention to where traffic is serving. We've talked a lot about all the different inventory types. So a lot of times if I'm like, "Well, why is this ad group performing worse than this one, or vice versa?", a lot of times I'll pull in the field so there's a way to segment it, so I can actually see the percentage of traffic that's flowing to Search versus Display or YouTube. And that, to me, is a really helpful tool. It guides me to understand what might be happening here.

Well, Google doesn't... you might have to create it as a custom column. I actually spend a lot of time looking at CPMs. I don't know if other people do that quite as much, but I noticed some big differences. Again, you have to do some guessing work on what's happening here because not all of it is exposed. A lot of times, if I notice a big decline in performance, then I check out, "Okay. Well, was there also a decline in CPM?" A lot of times, cheaper CPMs mean worse traffic. Not always but a lot of times, so that's another thing that I pay attention to when I'm going through and actually optimizing for performance.

David Barnard:

Are there any got yous to all of this as far as in any ways to point the algorithm in specific directions? Of course, you would never say this, but I've heard folks say that Google is going to optimize toward the things that make them the most money. So to your point about looking at CPMs, if you're getting lower CPMs, it may be that it's being served to lower quality traffic. And they're hitting your budget and you're still getting the performance you want, but maybe you could be getting better performance. How do you think about that?

Ashley Black:

Yeah, you have to pay attention to it. One of the things that people, again, don't like when Google rolled out UAC is that it blended a lot of the traffic. So all of a sudden, they can serve at the highest performing traffic. They know that that's going to convert. And then they serve the crappy stuff and it all averages out, and people are like, "Well, I don't want the crap. I just want the premium stuff," and so you do have to be on top of it. Those I was mentioning, the CPMs, where your traffic is serving, you can to some degree see what sites or apps your ads are serving in or what YouTube videos they're serving in. I pay attention to that sometimes. If it's a significant volume and I'm like, "Well, this is just a crap site," it's super spammy, there's 40 ads on there, I don't want my clients to be included in that type of inventory but-

David Barnard:

So, you can actually go in there and say, "Don't show on this site specifically," or you would even just say, "Don't show in this type of inventory"?

Ashley Black:

No, not on this type of inventory, although I think you can do it for certain apps. But it does block it at an account level, not just at a campaign level, but that doesn't usually cause too much issue. But yeah, so that's a reality. You do have to make sure that you're staying on top of it. I know a lot of people are like, "Oh, you just set it and forget it and it keeps doing it." It's like, "Yes, if you're okay with that, if you're okay knowing that you're getting some of the crap." And you can't avoid the crap altogether, but you can try and minimize it or try and increase the high-quality stuff.

David Barnard:

Would you also then be looking at certain images or texts that's hitting those lower CPMs that are giving you a hint that maybe it's lower quality traffic and actually just remove those assets from the ad group? Is that how you would approach that?

Ashley Black:

I did it yesterday for one of my clients. This one's a little bit tricky that I find is in this particular case, they were getting a lot of traffic from certain videos and that was fueling the majority of their scale. That's where all of the volume was coming from. Initially, I was really resistant to just being like, "Eh, can them, get rid of it." Because then, what can happen is that that will tank the whole campaign. All of a sudden it's like we've got no conversions, and so we'll just die out. You have to sometimes be delicate with that. It's like a balance between how much you're removing at once and so, in that case, I went in cautiously. I was like, "Okay, I see these videos. They have a super high CPA." Sometimes it's also important to look at your even further down funnel metrics.

Maybe the CPA is okay, it's within target. But in this particular case, the ROAS was crap. Then I had to make the call. I just slowly started kind of removing those assets so that it didn't cause the campaign to just have no conversions, and then give it a week. And then after that, I got to pull them all because Google's trying to get away with something.

David Barnard:

Got you. We talked about structuring campaigns early on, and that first just launch with one app campaign and then do your testing in 4 or 5 ad groups. How do you scale that up or what are your kind of campaigns structuring and testing? When do you layer on more campaigns? How do you do that smartly?

Ashley Black:

Well, layering on more campaigns in any specific geo, maybe you don't want to do it. Although, it can be a hack sometimes. I've noticed some people where they're like, "I'm not getting enough scale." I've seen people with budgets of $100,000 a day and just because they want to signal to the system, "Spend my money," and it won't sometimes. One hack, this is definitely not a Google best practice, but it's something I think I've learned more so as a practitioner doing this all the time, is sometimes just duplicating the campaigns. For whatever reason, you duplicate it and that signal to Google is like, "Oh, we have something new to go out, and we have to go test it." So then you might start to see more scale, whereas that existing campaign had just kind of remained very steady.

That's one hacky way to do it. But again, the scaling piece is just sort of slow and steady. I think it's hard if you're reaching your max capacity in one market to really force more scale without getting really aggressive with your bids, so then I start to look at other markets that might perform well.

David Barnard:

And then, what about ad groups? Are you constantly adding or removing ad groups? As you said, if you're targeting specific kind of messaging or talking about specific features, is that an area that you actively test? Like add a new ad group more focused on this feature versus that feature and remove that ad group that wasn't performing?

Ashley Black:

Yeah. A lot of times what you'll find though is that they won't, the new ones that you add in. For whatever reason, Google favors their first child. The first ad group that you create, I've just seen it happen so many times, that is the one that they're like, "We're running with this one." Even if another one that seems to be performing better, it doesn't always get as much scale. So yes, I do add in new ad groups. Around the holidays, for example, if we're testing a new holiday theme, I put it in a specific ad group that also makes it easier to manage. Right? I don't forget about it later on that, "Oh, we're running Halloween creatives." That's labeled separately. It wasn't mixed in with a bunch of others, so I can just pause that ad group. If there's something seasonal that's going on, like I work with some apps where natural disasters, for example, that happen, that's when they see spikes in their traffic and so we run individual natural disaster ad groups.

Like if a hurricane's happening or a tornado or if it's avalanche season or whatever, we have those separate and test out those things. So yes, I am constantly going in and trying new ones. But a lot of times, they won't get meaningful volume to really be able to say, "Was it worth it?", so you might have to leave them on for a little bit longer.

David Barnard:

And then with that kind of first child, as you called it, the kind of one ad group that the app campaign kind of preferences, would you then start to expand the assets in that ad group as kind of an experimentation of adding some new themes and other stuff in that high performing ad group?

Ashley Black:

Yeah. I think a lot of times I try and just what I've noticed, I don't hear people talk about it as often. Prior to Google, I worked at Pandora and that was more like a brand's plays. One thing we talked a lot more about there, keep mind this a really long time ago, was ad fatigue. I don't think that we talk about that quite as much. Maybe we do in the UA space, but I think that's something that people need to realize is that even in those high performing ad groups, at some point, people are just tired of seeing the same thing over and over. It's just hit the same audience too many times and they've tuned it out, or Google has just gotten in this habit of showing it on the same inventory or to the same people.

So even making a small tiny adjustment, actually, if you had something that said, "buy it today," and you changed it to, "buy today," you took out the it, that sends Google into a whole new learning. It's like, "Oh, we have to figure it out." So it's not a new concept, but to Google it is. It'll go out and test something new, and the ad isn't fatigued quite as much.

David Barnard:

We've weaved it into this entire conversation. But any other kind of troubleshooting or analysis tips on like, "If things aren't performing well, these are the first places to look."? And how do you troubleshoot campaigns and budget, and things to get optimal performance?

Ashley Black:

That's why people bring me on is that's what I spend a lot of my time doing. That's what my team does as well. The people that actually know the platform pretty well, I don't think this comes as any surprise. Everybody uses that, the segment called networks. That's again where you can see, at an asset level, how your ads are serving across all those different inventory sources. That is so important. You really need to understand where those ads are serving. And again, I did an audit for somebody just this week and they were like, "Well, we don't understand why we're getting this crap traffic," and I'm like, "It's because all of these assets are serving on the Display Network, and you're not getting any of that high quality search traffic." So if you don't know to look at something like that, then you're like, "Oh, Google's just not working."

So now, you have to go back and figure out ways, how do you force it in some ways to spend in these other areas or test those other areas. I guess, my big tip for people is don't just assume that it is doing what you think it should be doing. Sometimes it'll go rogue on you, and so you do have to know how to manage it and also what metrics matter. Like I said, I pay a lot of attention to CPMs. I know a lot of other people pay attention to IPMs. I don't tend to pay attention to that as much because of the... we talked about it. Extensively, the network changes, so your IPM on search is going to be really different than your IPM on display, so I think that's kind of unfair to hold them to the same bar. Anyway, I tend to look at those other early metrics to determine if something ultimately is going to perform well or why maybe it is or isn't performing well.

David Barnard:

It sounds like it's a trust but verify a situation where to use UAC, you have to trust the algorithm because you just can't get around having to trust that it's going to get you performance. But then, you verify that it's actually giving you the performance you want and then kind of subtly nudging it different directions with all these techniques that we've been talking about.

Ashley Black:

Absolutely, that might be my new tag.

David Barnard:

Trust but verify?

Ashley Black:

Trust but verify, yeah.

David Barnard:

Any subscription app-specific considerations? Most of our audience is subscription app practitioners as opposed to games. Is there anything with free trial optimization, targeting the install versus a trial versus a conversion, any kind of subscription app, specific tips?

Ashley Black:

I don't think that this is unique for Google, but maybe just if somebody is starting out to understand what goes into this. I typically don't like people to optimize their campaigns for purchases or trial complete or something like that because those tend to happen way too far out. If you're a 7-day trial, I think most subscriptions are, for Google to get that data back 7 days later is just too much of a lag, and so really optimizing for those earlier signals. Most of the time, it's a start trial in most cases. But you can play around with different events to optimize for. I think, yeah, for the subscription apps, that's probably the piece that I would pay most attention to. The other thing I'm getting a lot of questions on, and I'll be totally honest, I don't have a super sound opinion just yet, I noticed that Google... I know Google is pushing their value-based bidding. This is TROAS, that's what most people refer to it as, and I think it's still TBD if it really works for subscription apps.

It wasn't designed for subscription apps, it was really designed for IAPs. And then they have another product that's designed for ad-monetized apps. Same name, still TROAS. But I think with subscription apps, they've had a 50/50 hit rate on whether or not that bid type can work for their business model. So for that one, if you're feeling like you're being pushed at all on it, I think it's okay to be skeptical. But at the same time, it has worked for some folks, so it might be worth a short test.

David Barnard:

Yeah. Speaking of being pushed, I know one of the questions a lot of people would be asking, shouting into the car as they're driving, listening to the podcast or otherwise, would be, "You worked at Google, how do I get in touch with somebody at Google? How do I get a campaign manager? How do I work with Google?" And then you already mentioned one, and I know you've mentioned others in the past, that a good account manager can actually get you some data that you wouldn't get otherwise so-

Ashley Black:

It's the easy answer, but it's also the honest... you got to spend more. You have to think about it, right? I don't hope people don't get mad about that. Google has to pay these people, and so they have to make sure that they're getting an ROI on these individuals. So if you're spending 200 bucks a month and you're like, "Well, I want to scale," they need to see proof in that. They're not going with the trust and verify mentality like, "No, we need hard evidence that you can pay this, you got the budget for it, and then we're going to give you the support. We're paying these people X amount of money, we got to make sure it's worth it for us." That that's really how you get it.

The one thing I would call out though is that depends on your market, and ultimately you get assigned based on your billing country that you have set up in your Google Ads account. I have a client, he's based in the US but his account is billed out of the UK, so his rep sits in Dublin. There's different spend criteria. Obviously in the US, they're going to have the highest thresholds. You got to be spending a decent amount in order to get support. But if you're in a smaller market, let's call it Vietnam, that threshold might not be as high.

David Barnard:

Is it a, "We'll call, you don't call us," kind of situation? Once you start spending enough, they reach out to you? Or how does that kind of first contact work?

Ashley Black:

Yes, that is the case. They will reach out to you. And typically if you're wondering like, "Well, is today the day that they're going to reach out to me?," no, it happens on a quarterly basis. All of these things happen, call them like book refreshes. Right? This is when the reps actually, they have their assigned book of accounts that they work with and their goal against. Yes, they have revenue goals, so they will be pushing you to spend more money. And those get changed out. Rarely, can you make a mix up in the middle of a quarter. So if you're wondering, "Hey, am I going to be able to talk with somebody?, and it's May, April, May, June, that's Q2, no, you might not hear from somebody until July if you think that you are spending significant amount.

But I would say once you get a rep, I feel like people under-utilize them sometimes. Maybe my Google folks are going to get mad at me and they're like, "No, we're bombarded with requests." But I think that if you're smart, ask for more, ask for more information, be specific. They have a lot of it. They can't always share everything that you might want to see. But I think that a lot of times, people don't realize how much they do have and that they can share, so make those asks.

David Barnard:

For people before Google reaches out, are there any support channels? Is there a form that can fill out that is actually helpful? Or do those just get routed to people who can't really help you anyway?

Ashley Black:

When you're reaching out, those are all going to go to vendor teams. And that's not to say that they're not helpful sometimes. I haven't found them to be as helpful most of the time, and so... Oh, I don't have a great answer. Send me a message if you really have a question that you're like, "I can't figure this out for myself," but that's kind of the best bet at that point.

David Barnard:

Yeah, it makes sense. And like you said, and this has been a long time frustration of developers working with Apple as well as Google that like, "Oh, I can't talk to anybody." Well, yeah, because anybody that you would talk to who could make a decision is being paid total compensation, probably $300,000 that it costs Apple to have that employee to be able to talk to you, and so it makes sense. Like you said, they have to strategically allocate those resources in a way that's going to give them a return on investment. These are for-profit companies who are beholden to their shareholders and optimizing towards shareholders, not you as an indie developer or even as a large developer.

That's something I've kind of come to internalize a little bit more in these last few years is like, "Okay, yeah. Apple's not my friend. Their fiduciary is their shareholders." When you kind of really internalize that and you look at all these decisions and the way they do things, and them pushing so hard on collecting their fees and all those kinds of things, it's like that's just how the world works. It's just how it works.

Ashley Black:

Yes. Actually, the ads team at Google is pretty substantial. From what I've heard, Meta has made some major cuts on their support to the point that nobody has a rep anymore at Meta. I'm exaggerating, but you know what I mean. There's not a lot of support there. Google's Ads team, I think, is still pretty healthy as it stands today. I feel like a lot of people do have a rep, even if it's somebody... By the way, there's different tiers of reps based on your spend, so there is a difference in support level. But I think maybe even more frustrating is on the Play side of things where that team is actually really, really lean, and so I feel like nobody has support from Play.

And then also you have to think about it on Ads, they're like, "As long as you're spending money with us, great." Play only cares if you're making money for them. So if you're heavy on ad funded, they don't care. They're not making any money on you. You could be a massive app that has 100 million downloads or whatever. You're not making any money for Play, so they're not as concerned. But that team is actually incredibly small, and so there's not a lot of developers. I'm not trying to say they favor them but, again, if you're thinking about how these companies make money, they work with a lot of the gaming studios. That's where a lot of their support flows.

David Barnard:

Yeah, it just makes sense. But I will say that if you're struggling with these things, this is why you created Candid Consulting. So, how can people reach out? Are you accepting new clients?

Ashley Black:

Yes. Please, reach out if you're interested in talking. I have sort of an agency side of the business where I actually do all of this for my clients. But I find that a lot of people do reach out just hoping to get some information, so I do partner with people in that capacity too. You can reach out to me on LinkedIn. My name's Ashley Black. There's not too many of us, you should be able to find me. Or you can reach out directly through the website, which is candidconsultinggroup.com. Yeah, I'm happy to chat with people. I'm always interested to learn what other people are working on, what problems people are seeing. And a lot of times, people just want to know what actually is going on at Google. Shed some light or lift the curtain a little bit, and so that's always fun to talk about too.

David Barnard:

Well, it's been so fascinating and fun to get the kind of inside scoop on a lot of this stuff. But then also now that you've been out of Google for over a year and worked with so many different clients, so many of these insights are things you probably didn't even know when you were inside Google. You've done a lot of testing as we've talked about, and so thank you so much for being so open to share your experiences and how you optimize all this.

Ashley Black:

Absolutely. I appreciate you having me.

David Barnard:

Thanks so much for listening. If you have a minute, please leave a review in your favorite podcast player. You can also stop by chat.subclub.com to join our private community.