Marketing Your App More Efficiently with Apple Search Ads — Dilip Reddy, Search Ads Optimization

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 Barnard, and my guest today is Dilip Reddy, a data engineer turned marketer working on an Apple Search Ads platform called Search Ads Optimization. On the podcast, I talk with Dilip about the impact of Apple Search Ads on organic search, how to save money on brand defense, and why ROAS shouldn't be the only thing you optimize for. Hey, Dilip, thanks so much for joining me on the podcast today. 

Dilip Reddy:

Yeah, I'm excited to be here talking about Apple Search Ads. 

David Barnard:

Yeah, I've been meaning to do it episode on Apple Search Ads for a long time, and there's so many different frames of reference to talk about it. And so, there's a few people I've talked to that I probably will still have on, but a little more growth hacky, more marketer-focused. But one of the things I'm excited to talk to you about search ads is that you're a data guy, and you stumbled into marketing as a data guy. So, let's just kick it off with that story. How did you get into Apple Search Ads as a data scientist?

Dilip Reddy:

The reason is I did computer science and always data is my go-to like this what fascinates me. And I started Bank of America, there's no app thing and all. But over time I joined unicorn app companies in 2016 and that's when the Apple Search Ads came in. Because I'm a data architect as well, so I manage all data sources and design architectures. I noticed Apple Search Ads is different in a way that it gives so many data points, so much of data that extreme granular including high-intent keywords, how the users are reacting, and it just was great. And also, I created my own hobby apps that don't solve any real problem, but I started tinkering with Apple Search Ads. 

And in a weird way, this my on and off, I've been developing all algorithms and all this stuff, and that's how I ended up, and yeah. 

David Barnard:

Yeah. And then you had told me at one point that one of the things that really fascinated you about Apple Search Ads was that you were digging through the data one time and saw some really high bids and some high cost per acquisition. Yeah, so tell me about that. 

Dilip Reddy:

Yeah. There are some categories like finance, health, and fitness. The bids can go crazy for some market, some keywords, but they also bring in high revenue. That's what is interesting because if you look at in a spray and pray networks like Meta, everything is usually constant. You expect this and the same revenue performance because this not intent-based, it's just shooting creative across millions and hoping some will fall for it. But I would suggest that even the CPS is high I find that there's something we can tinker with and experiment with because we cannot stop just because it's high CPA because they're the high intent. 

And in a way it's a limited inventory, demand and supply, so it depends on the app, it depends on what keywords we're targeting and what kind of users we are looking for. And especially for very high competitive, finance is one thing, health and fitness, we may see bids go really high up and CPA can go high in USA. USA is one of the most expensive storefront.

David Barnard:

And that is the interesting thing about search ads, it's why Google is a what? Trillion-plus dollar company. It's that when people go search for a specific term, they're very clearly demonstrating intent and I don't remember all the different categories, but there's things like searching for a divorce lawyer on Google, whereas the bids just seemed insane. But if you're searching divorce lawyer, something's going on in your life and the lawyers make a lot of money on divorce settlement and so it's like they can bid really high. 

And I think there's similar market dynamics at play in the App Store where for things like finance and health and fitness where there's high intent, there's high willingness to pay, it makes sense in a competitive bid environment that some of the keywords will get bid up. And it's cool that part of what got you into Apple Search Ads was seeing those high bids and be like, "Wait a minute, this data must be wrong." But no, it's like, "Actually the data was correct, it's just that you can have a return on ad spend with a $50 CPA if it's the right keyword and the right app and the right monetization." So, that's really interesting. 

Dilip Reddy:

And yeah, the good thing about it is we can control each and every aspect at granular level. You can also stop some keywords like bid really low. So, either forcing Apple to stop giving those impressions or give more installs for the same, not more installs but similar installs with the lesser CPA. That's the best part, we are in control of everything. It's not like Meta, we don't know why the CPA is going high and I can't do anything about it.

David Barnard:

Yeah. A lot of granular control and a lot of data to understand what's going on. Yeah.

Dilip Reddy:

It is a double-edged sword, it can get overwhelming, especially if a company is running all 60 countries and 70 countries, like thousands of keywords for each country. We have some solutions for that and there's some campaign structure we can set up to make it the mental overload much better. 

David Barnard:

I wanted to kick things off talking about trends for 2024. Things are always changing, Apple introduced the brand, or I guess not brand but the inventory on search and they've been expanding inventory over the last few years. Anything especially interesting in 2024 that it is changing in App Store Search Ads that folks should be paying attention to?

Dilip Reddy:

A lot of apps now, usually, it used to be like we just run in Tier 1 in US, UK like Germany, France, and forget the rest of the world. But now, I'm noticing the trend of adding as many countries as possible. Because if you think about it, there's only five, six localizations that we have like for example, if there's Spanish, you can maybe have an entire Latin America, not entire, but most of the Latin America with the same kind of keywords. Same with Portuguese, German, France, the way is we identify the keyword builder building for these localizations, but you can expand to pretty much every country in the world. And for extreme local language that we cannot support or Apple defaults to English anyway, so expanding to more countries is what I'm noticing. 

David Barnard:

Apple introduced new geographies this year, right? 

Dilip Reddy:

Yeah. Brazil and I think six other Latin countries. So yeah, this is the best way and people are seeing some surprising results. So, my advice is test these countries because we can't just assume, "Oh, because this country is this, we should not run it there." Especially the effort is pretty much the same. It takes maybe a few minutes to just copy and paste the campaigns.

David Barnard:

But you don't want to just copy and paste the campaigns, you want to copy and paste and then spend a little time honing it in for that market and then check your CPAs and your return on ad spend and stuff like that. 

Dilip Reddy:

Yeah. Whenever there's a new storefront, my go-to is start with bids like too stupidly low that will never get impressions, but that's fine we can always increase it. If you start bids high for a new country, we don't know what is the base. I can start with $10, Apple happily give me $10 bids, but maybe I can get the same number of installs or 90% of installs if I bid half. And also, the trend I'm noticing is the CPA is going a bit high, and I'm assuming it's also across all ad channels because I don't deal with other ad channels, but the data suggests CPAs are going a bit high year by year, so we are dealing with that.

David Barnard:

Yeah. It makes sense because as apps get more sophisticated about monetization and really get their monetization dialed in, then they can spend more. And so, the market dynamics is as consumers get more comfortable spending, as inflation causes higher prices to make more sense, as apps work on their funnels and paywalls and all the things we talk about on the podcast, it's like we're creating the problem. As more apps get more sophisticated, then they can bid higher because they're going to still get a return on ad spend even with that higher CPA. So, it makes sense that there's this rising CPAs, especially with the most high-intent search ads. 

Dilip Reddy:

Yeah. I think there is also a trend of, I call these apps hyperscalers. It's like Temu, they bid 100,000 keywords for some reason. But the reason is maybe, I don't know why, but they must have billions of marketing. So, unfortunately, it'll raise up, it's a way of lifting up all the boats kind of thing. So, because Temu is bidding for each and every possible English character. This letter and word in the dictionary, it may mean that the CPA may be going high. I'm not saying this is the reason exactly, but this is a trend I'm noticing as well. 

David Barnard:

Yeah, probably contributes to it for sure. 

Dilip Reddy:

Yeah. And there's a new search inventory came in that didn't exist before ChatGPT is the AI, Everything AI, YogaAI, pretty much there's entire new generation of search terms come in, so the inventory went up as well. So it's not all high CPA for the same inventory but also the apps skyrocketed, number of apps released. Ironically, AI is making easier to create many apps at scale as well. So yeah, that's the trend so far. 

David Barnard:

Yeah, that's really interesting. And the explosion of apps thing is interesting. It does seem like AI has accelerated that. Like you said, for every major category you might want to bid on there's probably somebody who is either built a more really sophisticated app or rebranded an existing app to say it's the AI version of it. And then people are looking for more apps because they're thinking, "Well, hey, I've been using this productivity to-do list. Is there a better AI version of a to-do list that's going to make me more productive?" And then they go back and search, whereas maybe they weren't searching some of those terms before because they had the app that they used. 

Dilip Reddy:

It would be interesting to see the iOS new version with the built-in AI, how it's going to impact a lot of apps. Yeah, let's see how it goes with the AI. 

David Barnard:

So, one of the hotter topics I wanted to get your take on is brand defense. So, you have some people say never do brand defense, you're just like it's extortion from Apple and Google to force you to do brand defense on SEO and ASO. But what are your thoughts? Can it work? Can it be profitable? How should you structure it if you do do it? How do you think about brand defense on Apple Search Ads?

Dilip Reddy:

Yeah. Obviously, it's a touchy subject. I have some opinions, but what I would do is I ask and inquire with the company like, "What is your plan? What do you think of this?" Because it depends on each and every company. On the surface level, this is what I would suggest. If your app is easily replaced with a competitor app, you have to run it back. The unfortunate situation here is that if we lose impression share that we cannot measure that because it's a lost opportunity. We can only measure something incrementality, but it's a loss of revenue. If you think about it, let's say a company is running Facebook, TikTok brand awareness, all the touchpoints made that user come to the app store right before downloading. If a competitor solved the exact same problem, it's a really huge risk.

We made all the hard work and now the competitors are benefiting from it. There are some apps with apps with network effects and we know these apps cannot be replaced with another apps. They do not need to run brand, for example, maybe Bank of America app. No one creates a online application, but there's an opportunity for FinTech ads, right? Because that's their opportunity. Bank of America doesn't need that, but FinTech can capitalize on that opportunity. My suggestion is test with bidding lower for the brand keywords, you'll be surprised you may be paying, let's take one example. An app is getting 100% impression share, to be honest, it's not 100% because there's never 100% impression share. Let's say 90% impression share. 

David Barnard:

Is there a way to measure that incrementality testing or does Apple actually show you this impression share? 

Dilip Reddy:

This is the best part. Apple does give us impression share per search term per country. So, that is one thing we can keep track of. That's the reason why it allows us to conduct this test is bid lower for brand and test it out and if you retain same impression share, you essentially have the cost still getting same impression share. It is a test because we don't know how impression share can take a hit and maybe some companies are okay with 70% impression share if it reduces cost by 70%. So, it's a nice trade off you still retaining, we still letting go of our impressions for other competitors, but we can use that budget for generic high intent and these are actual new users. 

And regarding Apple extorting, I mean it's not unique to Apple. It's every ad channel there's no trademark on search terms. So yeah, it's unfortunate reality of everyone has to do it, but it's also an opportunity. 

David Barnard:

The impression share is an interesting way to approach it though because you're right, it's like if you can spend half as much money and get 90% impression share versus twice as much money to get to 100%, that extra 10% is probably not incrementally better than just spending less and getting the 90%. And then something you and I were talking about was how the search ads algorithm does tend to prefer the actual branded app where you could potentially even bid a lot lower than your competitors and still win the bid. Because Apple's not just going with the highest bid, there is some amount of them trying to fit the best app, right? 

Dilip Reddy:

Yeah. It's because of one mechanism is that cost per tap model. Apple only charges an app company if someone taps on it. If you get a million impressions with zero taps, you basically got all this free 1 million impressions. So, if that's the case, if Apple just takes highest bid, like Temu will be showing up for a research term, but it's not. The reason is imagine if I'm a yoga app, a famous yoga app and someone is searching for the yoga app of my brand, a competitor may not get same level of engagement. Lot of users may see the ad and "Oh, you know what, I don't care about it. I heard about this app so I'm going to download that." Apple is losing too and it's also not great user experience. 

If user is typing for yoga brand of XYZ, they are expecting to see XYZ not so eventually Apple may shift the terms to brand. Even if you're bidding like let's say half of the competitor because there's some calculations, it is also a test because it's not guaranteed outcome. Imagine I'm a just generic yoga app and the competitor has way better screenshots, way better messaging, way better ratings. I may lose that and Apple may think, "You know what? This competitor app is actually working better more than this brand." So, Apple may give the impression share. So, it is a bit of test, but that's where it's ideal to monitor the impression share and make sure we are not like tanking.

David Barnard:

Yeah. No, that makes a lot of sense because yeah, you're right. In your example earlier I think is maybe even better, it's like if somebody's searching Bank of America, they're probably not going to tap on an ad for Chase Bank, and so Apple doesn't make money when somebody searches Bank of America and they show Chase because people are looking for the Bank of America app because they bank at Bank of America. They're not searching Bank of America, they're looking for bank, they're looking for the app. And so, it's probably the exact opposite on like SEO. It's like if somebody's searching Bank of America online or other kind of keywords, it might make more sense. 

But on the app store specifically, when somebody's looking for a very specific app, Apple's not making money if people are having to scroll past the ad. So yeah, it makes a lot of sense. So, on brand defense you really experiment and try and bid as low as possible to maintain as much impression share as you're comfortable with to make sure you're not losing out on the opportunities. That advice alone, I don't know it does seem like a lot of companies, they just throw money at app store search ads and don't operate in a very sophisticated way. And of course, tons of them hire consultants or have internal people who are really good or have agencies working on it who know these kind of things. 

There's probably folks listening to this podcast who don't know that and aren't paying close enough attention and you may have just saved them like tens of thousands of dollars just go in there and lower your bid and watch impression share. 

Dilip Reddy:

There's one more thing about the ROS and case studies and one thing I noticed is a lot of case studies don't mention if it's generic or brand. This brand usually has highest ROS of any kind. If think about it, that's what we made the user come to app store and their highest intent possible. So, when we seeing the revenue ROIs I would consider brand as its own entity, be treated as we do not compare brand with any other high intent generic keywords or competitor keywords or broad match keywords. I treat it as two different things, I do not even mix up the data as an entire opposite strategy. 

David Barnard:

Yeah, that makes a lot of sense. Yeah, if your overall ROI looks amazing, yeah, because you're cannibalizing, but if you can pay half your ROI is going to look even better, so that's great. We already talked a little bit about the trend in CPAs going up, up, up, but I wanted to dig into that a little bit. So, we talked about some of the high CPA like finance and health and fitness and some of these categories that are going high. What about low CPAs? There are some categories that are opportunities, some arbitrage available in the app store in certain categories? 

Dilip Reddy:

Yeah. Even with the high competition categories, we can have some CPA, we can improve CPA better by just expanding high intent and low intent keywords and bid accordingly. For example, I know yoga, if someone's typing yoga, yoga is high intent for my app. I'm talking about yoga as because it's such a generic app, but I also know if someone is doing Pilates or some kind of, maybe there's a low intent, they may not be interested in my app, but it still, if someone is looking for health and fitness and Pilates, maybe they're low intent, but I won't bid the same as I would bid for yoga. 

So, I classify that as low intent, that way you can expand keywords a lot and also keywords, we can think of it as this where the night and day difference of performance can occur in opposite chat. If I'm choosing bad keywords like have 10 keywords that are not related to app and bidding $10 each, we are going to have some troubles. This is a human thing we know much better. You can use AI ChatGPT to have some ideas, but the final say should be yours of, "Okay, we are using this keyword because it's slow intent and we'll bid lower and in hopes that we get more solutions." And the best part about Apple Search Ads is there are two types of keywords, exact is not 100% exact, but it's more like yoga for woman for example, is that if I'm bidding as exact keyword, that's fine. 

But there's another keyword type called broad match. What it means is any variations can also, we can all right automatically shows up. For me if yoga for women is same in intent as yoga for women in 30s, I know it's a long tail but yoga for women, I don't care what the following search are because I already know yoga for women. And these keywords usually are dynamic because if you look at the app store app, if you type yoga, you'll see Apple such just some search terms and these change very often so the competition may not exist for some of the search terms and that's the low hanging fruit run the broad but bid obviously bid lower. 

I'm a big fan of classifying keywords and bidding different for each based on the group brand and regarding low it's also very dynamic. Even with the high competition, some keywords may not have any competition at all and there's think of stock market, there's so many people trying to buy stocks and we cannot predict what's going to happen. So, this is where we try to have as many as relevant keywords and we expect some may go down in impression share, some may go up. Overall, we just see improvement in performance. 

David Barnard:

So, is that how broad match works? I'm a little naive here. I have not unfortunately, you and I were talking about I actually want to start running after search ads campaigns for my weather app. I've been holding off because I'm like, "I don't know if they're going to be profitable and it is tough in a category like weather," so I'm a little naive to exactly all of this works, but maybe that's good because I think a lot of the folks who listen to a podcast like this don't necessarily day in, day out run campaigns. So, it's helpful to define some of these terms and describe in a little more detail how some of this works. 

So, broad match, if you do a broad match for yoga for women, will it broad match anything related to yoga for women or does it still have to have yoga and women in the search terms? How dangerous is it to set a broad match? Will it just pull any random stuff in there? 

Dilip Reddy:

Yes, just because we said yoga for women, it doesn't mean that Apple may show for truly unrelated. If Apple algorithm thinks, for example Peloton, if Apple thinks Peloton is related to yoga for women our ad, we may be paying for Peloton keyword. But the good thing is we can even track this, there's a report called search terms report. It's better to occasionally we check this report and make sure, okay, all our spending is going to the right kind of keywords. Usually, the search terms is not perfect, Apple only shows search terms if we spend enough and if you get enough impressions. But that's fine that what we want anyway, we don't need thousands of search terms report. 

David Barnard:

So then in this case, if you have a yoga app and you see in the report that it's bidding for Peloton and it's a high bid and you're not getting good ROI, do you go in there and then specifically say don't bid on Peloton? 

Dilip Reddy:

So, the good thing about that is we can track which keyword is causing our app to show for Peloton. Usually, I just go in there and tone it down to really stupidly low that either we get really, really cheap like installs our app, we force Apple to not show our add. I'm not a big fan of removing keywords too much often these are dynamic trends, right? Let's say I have a yoga for woman and Apple is showing Peloton search term for my app this week. It doesn't mean that it's going to keep doing there forever, maybe next month that might be my most profitable keyword. If I just remove that keyword, I just lost my opportunity forever. 

But if I bid low and I can retry later two, three months and see how is it performing. I think yoga for women is not a great example for broad, but you can think of as a, let's take weather since you're app, hopefully I'm not giving away weather app secrets. Imagine we are bidding, we just don't know anything about Apple Search Ads, but we know our app is about weather. So, you can run on weather as exact, it means that any variation whether we show the ad and we pay high for that. Let's assume we did not do any keyword research, we don't have any keyword list. If you also bid on broad of the same, so we are running two keywords basically Apple may show with their map, with their information, with their alerts. 

As far as we concerned, those are also relevant for us and you can go to the multiple levels. Now, with their alerts, we can bid on weather alerts because Apple showed, oh there's a weather alerts keyword. Now, we do the same thing, we add weather alerts as exact and also broad and our hope is that if someone is searching for whether alerts Philadelphia, it's still relevant because Matt still relevant and if someone is searching whether alerts Los Angeles, same thing. So, that's why I'm a big fan of maintaining a relevant keyword list because this is one-time action. Once we add whether alerts, we are not going to, unless the app changes center basis model and the name and the functionality, it's really rare that the keyword is going to be applicable forever. 

David Barnard:

Yeah. That's a good point though because for my app it would actually be the opposite. So, we would start with weather as a broad match and we would probably see that weather alerts is a very low intent keyword because we actually don't have alerts in the app. It's something we actually want to add to the app. So, then what you're saying is what we would do then is put weather as broad match when we see weather alerts come in and it's not creating a great ROI, even if maybe it is driving downloads because people don't realize the app doesn't do it. But then they get into the app, it doesn't do it, they leave a bad review, they don't buy or they cancel the free trial or whatever. 

So, then we make whether alerts a exact match and then we bid really low because maybe we still get impressions and maybe we still make some money, but we want that specific keyword to be super low price. But then when I add the feature, then we already have some relevancy for it and then I can bump the bids back up, right? 

Dilip Reddy:

In this case when we are sure a particular keyword is really badly performing and it's consistent over and over again, it can actually add it as a negative keyword. Basically, what we are saying to Apple is anything that says weather alerts, yeah take me out, I'm not in the game. So, this actually much better, but we need to be absolutely sure that this keyword is never going to be a good fit. I noticed some advice that just remove negative, any keywords that don't generate revenue in first week, that is really not great a strategy in my opinion because everything is dynamic. 

Some competitors may be bidding high this week, they may not drop off tomorrow, it's all dynamic. That's why I'm a big fan of Apple Search Ads. It gives us a lot of data to validate and test our hypothesis. In this case, if you know this keyword is a bad news, which is added this negative keyword, one-time thing and we never have to worry about it again. 

David Barnard:

So, is that how exact and broad match work? A lot of people who work in the industry are constantly managing these things will know that, but I'm sure there's a lot of folks listening who don't and then honestly even the people who work in this day in and day out could probably learn some of the intricacies about how it does and doesn't work from you as much as you've managed this and work with so many different folks over the years. So yeah, how does broad match work? 

Dilip Reddy:

Yeah. The goal for us when we are using broad match is to capture exponential number of search terms that are still relevant to our app. And added benefit is usually for these keywords these are so dynamic, they change every week. They may be low end popularity so our competitors may not be bidding really. 

David Barnard:

So, we've talked a bit about fluctuation, but one of the biggest fluctuations is seasonality is I'm sure January 1, all the search terms related to health and fitness are bid to crazy and that's to be expected. But any other things you think people should be watching out for seasonality wise or how does that impact CPAs and how should you be preparing and structuring your campaigns for seasonality so you don't end up paying too much. But also, so you don't miss out on the opportunities when there maybe are higher, higher ROI searches when maybe you are bidding lower even when there are higher intent people coming in? 

Dilip Reddy:

Yeah. For US storefront, the Thanksgiving is usually the highest competitive space. It's a similar phenomenon, a way of lifting up all the boat kind of thing. So, even if our app, it has nothing to do with shopping, we may see our CPA going up bit high. Even summer time is usually for, it depends on kind of apps, for some categories the summary is actually the cheapest CPA for some reason. You can think of this as cool if I'm education app and usually the best time period to test any new app is, if someone is thinking of trying out Apple Search Ads is during February till summer, this weirdly low energy in the competition competitive space because the January high is gone and now things are coming back. 

But it depends on the app too. If my app is dealing with something with, for example, flights, flight tracker, obviously I can benefit much because that's not peak travel season. Those are the seasonality trends.

David Barnard:

And some of them are counterintuitive too. I was actually talking to Ryan Jones from Flighty recently and he was saying how surprised he was that holiday travel is such a, I think the conversation was related to me saying, "Oh you must do really good business in the summer." He is like, "No, actually the summer's fine, but the holidays is where people are way higher intent, traveling more, care more about delays and stuff like that." And so, some of these seasonal trends are probably counterintuitive and even you're running your app thinking, let's take AllTrails for example. It's like, "Oh, the summer's going to be peak season," maybe. 

But are there ways that you, pretty much you just need to be watching week in, week out, but are there specific ways that you track these campaigns to make sure that you're not missing out on either potentially bidding lower when CPAs maybe drift lower or bidding higher when there's actually higher intent people coming in? How do you manage that on a day in day out basis? 

Dilip Reddy:

It depends on how our app category, how the competition is. For some apps, it depends on what our competitors are doing because when I'm saying competitors, I'm not talking about exact close competitors. I'm talking about any app that are bidding heavily and for a long time for our primary keywords, high intent keywords, it depends on that. For day in, day out usually we should not, we just keep in mind that hey, this is expected the good thing is it should correlate with organic traffic as well. So, we can just come back to App Store Connect and check last two years, what's our trend? 

And we can expect similar in Apple Search Ads, I'm not a big fan of changing bids by the seasonality because like I said, it's like stock market, everything's dynamic. We cannot predict that we'll get a percent share of this keyword, but we can see the results and act on it. There's a pitch thing about data, we don't have to make any assumptions. We can see the data and act accordingly. 

David Barnard:

So, what you're saying is the better tactic is just to watch the keywords. And so, if you're getting a higher ROI on a specific keyword, no matter the seasonality, that's when you would experiment with increasing the bid. And then if you see ROI drop, then that's when you experiment with lowering the bid and you should just be doing this as standard practice anyway, not necessarily focused on the seasonality. 

Dilip Reddy:

Even as I say watch keywords, it means it's like constant process. Imagine I am running an Apple Search Ads campaigns for my company and I just go in every day, every week once in a while and I just be according to what I'm seeing. It doesn't matter, maybe Germany is the least performing campaign. It doesn't matter what season, I just, "Yeah, you know what? I'm not going to spend too much budget, so let me bid lower for the entire art group." Maybe US is least performing, but Germany is best performing in December for some reason, I don't know. Because these are all external factors, we cannot analyze why it happened we could just react what we are seeing based on past 15 days or past seven days. 

David Barnard:

And speaking of ROI, how do you think about ROI because you know you're not always going to be able to get, the magic is plug in a keyword and 150% ROI in seven days and then you spend $1 million a month and you're just making money hand in foot. But that's not how it actually works in the real world. So, how do you first measure ROI and then how do you think about ROI? 

Dilip Reddy:

We'll take an example of a regular generic app with seven-day free trial and some- 

David Barnard:

Let's talk about my app because I am going to start running app store search ads on different weather terms and so I'll probably be talking about this over time on the podcast, what's worked and what hasn't for me. So yeah, a weather app. So, I start running ads for my weather app, so we'll use that as a hypothetical. 

Dilip Reddy:

What is the free trial? Is there free trial? 

David Barnard:

Seven days.

Dilip Reddy:

Seven days, perfect. Here's the thing about revenue in Apple Search Ads. Apple Search Ads by default, doesn't give revenue, doesn't show revenue unlike Meta, Google Ads. But there's some ways that you can use MMPs like other platforms and get the information and link it up. 

David Barnard:

And RevenueCat. 

Dilip Reddy:

Yes. Yeah, RevenueCat does a great job like linking up keywords and ad groups and you can use that information to analyze. My recommendation is do not go in there expecting a hard number as revenue goal and if it doesn't hit, take the entire channel down because these are really dynamic and there's so many things that we can improve on. For example, it's a totally different of revenue. If I Google right now, what is a good ROAS the usual wisdom is 400%. What they mean is, okay, if you spend $100, you expect $400 coming back. But that is not for app, that wisdom is not for app marketing because for simple reason. We have subscriptions and also all this 400% ROAS golden metric is for example, if I'm selling T-shirts, yeah, there's a metric I can follow on. 

Because if I'm a person who buys my T-shirt will never buy same T-shirt next month, but there's a important metric called time to break even or it means is it's fine even if you don't get 100% ROAS in the first week. Obviously, if we get that that's good, but even if we get like 40% or 50% ROAS it's actually not bad result because we can still improve on add more keywords, we can fine tune and also, we can check our LTV. And because we are bidding for high intent, we know that these users are perfect users of our app so we can expect, we can improve in-app experience to make sure they don't churn. 

And if you have monthly subscription, hopefully, because if you think about it, cohorted always mean we only spend that, let's say we spent $3 for that install for that user who started this free trial and they did convert. In that case we only spent $3 one time for that user and that's it. But the revenue is hopefully our app is better and this not huge churn we can keep on the revenues keep coming in so there's a time delay, we just have to keep track of it. 

David Barnard:

And so, then every app developer has to decide from a cashflow perspective, from a long-term business perspective. I was talking to one person who, because you hear so much in the industry of like, "Oh, what's your seven-day ROAS? What's your thirty-day ROAS? Are you getting ROAS positive in 90 days? Well, what's a metric?" And I was talking to somebody who has really great organic traffic, so their business is just spinning off cash. And so, their metric was 365-day ROAS. So, their metric was can I spend today and at least get to 100% by the first annual renewal. 

And for a lot of businesses that can make sense, if the cashflow makes sense, if you're wanting to really grow. And then the cool thing like you were saying is those people who renewed that first year, you get to ROAS positive in 365 days, which again is crazy, it's hard to float that for a lot of companies. But then a bunch of those people are going to renew the next year and the next year after that. And so, it's an arbitrage opportunity to grow faster now if you have the cash and you have some level of confidence in those renewals that you're going to start stacking those cohorts sooner. 

That's actually why I want to start advertising for my weather app as soon as possible and I'm kicking myself for not starting. It's like should at least be picking up the low-hanging fruit and maybe this is the message for those of you who aren't currently doing it, even for those of you who are and maybe could even spend more than you currently are is that, this is magic of the subscription app model is that that LTV is not capped. Even if you have an average LTV for most apps who are reasonably retentive, whatever average you calculate today is likely to just keep going up year after year after year after year. 

I was looking at my own app and seven years in, I still have 20% of people who subscribed within the first month. So, seven years still have 20% of people and so, if I can stack more of those cohorts now versus stacking them later, even if I'm not getting 100% ROAS on day seven, if I can afford the cash to do that and expand the business, it's going to pay dividends in the long run. But you just got to make sure you have some level of retention and confidence in that and then you have the cash and the risk tolerance to go ahead and invest more heavily. 

Dilip Reddy:

Yeah. Obviously, the caveat is that we have all the basics sorted out. What I mean is the install to free trial ratio, the free trial to conversion ratio, but this is nothing to do with Apple Search Ads. The goal of Apple Search Ads is bring in new users. We have to improve those in-app conversion rates anyway ourselves regardless if we run ASA or not. And one more thing about ROAS is ROAS should not be the primary metric, it should be a combination of revenue and ROAS. For example, we can think of two apps, let's take meditation, let's take headspace and comm. 

These are imaginary numbers, so don't sue me. Imagine headspace and comm both are running Apple Search Ads and headspace is spending let's say $1 million and their ROAS is 75% ROAS, but comm is spending $2 million with 65% ROAS. On the paper it looks like headspace is doing much better, because it's getting 70% ROAS within first 30 days or whatever. But in long term the comm is actually better positioned to lead because at the end of the one year they both are going to have 100% ROAS anyway according to the LTV metrics. But comm generated double the amount of revenue that headspace, even though there a ROAS is not great for the comm app. 

So, it is a combination of revenue and ROAS and if you don't consider revenue, our competition will and they will eventually bid higher of whatever the ROAS we are getting currently, it's going to suffer because now they're going to bid higher for our positive ROAS keywords. So yeah, revenue is one of the major number as well because that's what we want. 

David Barnard:

Well, and the other thing that revenue takes into account that's hard to factor in is that there is some organic lift when you're running ASO. I know this is another controversial topic and I know you have some hypothesis around the impact, but the reality is there's some impact. Whether your hypothesis is correct or tons of people in industry are like, "Oh, it's going to be magic run ASA and then your ASO is going to be amazing." So, what's your hypothesis on the impact of ASA on ASO? 

Dilip Reddy:

I would call it conspiracy theory. I would not even call hypothesis because that's such a take it with a bucket of salt. 

David Barnard:

This what I love having you on versus like a growth hacker because a growth hacker would've just been like, "Oh yeah, 100% this is how it works." You're like, you see the day to day in and day out. You're like this isn't even a hypothesis because you want to have a level of confidence in the answer. But yeah, what do you think? Because honestly your conspiracy theory is probably better than a lot of folks who speak confidently about this. 

Dilip Reddy:

My theory is that download volume correlates with organic ranking. It's one of those trend that it doesn't matter what app, it's always correlates with such a strong efficiency. The second factor I think is this is where the conspiracy theory report come in. Even at the keyword level, imagine we are having the same amount of downloads, there's not much huge change in the donor volume. But you start running for whether app, Apple Search Ads, let's say whether radar maybe is one of the feature of your app. 

David Barnard:

Yup, that's a great keyword for my app, yeah. 

Dilip Reddy:

So, whether radar, imagine your current organic ranking is maybe Sony or something. As soon as you store running Apple Search Ads for that particular weather radar search system and our app generates some installs, you may see the organic ranking go really high, maybe even reaching top 10 or top 20, just the skyrocketing within. But the download volume is same. That's my conspiracy theory that maybe app store considers Apple search as keyword as part of app store. And now even the app store algorithm can see that, "Oh, you know what? This weather app is totally relevant to weather radar because users are clicking on it and using it and download." 

David Barnard:

But then why wouldn't it change the download volume? If I start advertising for weather radar and I go from, I'm probably like 200 in that search currently. So, I go from 200 to 20, why would download volume not change? 

Dilip Reddy:

I'm only talking about that particular keyword. It's an imaginary hypothesis that we are running Apple Search Ads for only that keyword. Let's assume our daily download is like 100 units in US and we started running that keyword for two installs a day. I'm giving you an example, maybe over end of the week the download volume is not that high, like 100 versus 102 installs. But the organic ranking for every other keyword may remain same, but for this particular keyword, what we are running in Apple Search Ads, we may see an organic ranking uplift, but like I said, take it with bucket of salt, it may not happen. I don't recommend spending a lot of budget just for that purpose because it may happen, it may not happen. 

David Barnard:

Yeah. It does seem like Apple's search algorithm, although it's still frustrating, it doesn't feel like they're still especially great at it, but it does seem like it's gotten more and more sophisticated over the years. So, the people who think there's going to be a one-to-one response in taking a certain action might be surprised. And like you said, yeah, it's a lot of money to put on the line when any kind of algorithm like this, there's not a level of certainty on it. 

Dilip Reddy:

While we are on the topic, the dollar volume organ rank, the theory is that it is like a chicken and egg issue, especially for new apps or apps that are actually best in the class, better than all the competitors but the competitors will be always number one. In that case, do not treat Apple Search Ads as like there's a channel we should never run on because this may be the better option because as long as your app is better than competitors and they're ranking number one. Even then you can actually get high intent as long as we talked about the basic our in-app conversion ratios are basic, this all great. If that's the case, you don't even need to run million install campaigns, you can just target those high intent. 

We are not going to get quantity, but we'll get the quality users and as long as we scaling up our revenue, why does it matter how much if you're ranking number one or not? I'm talking about Apple Search Ads, obviously it matters that you rank number one, but sometimes it's an extreme impossible for some keywords, we can never rank number one. 

David Barnard:

Yeah, no, that makes sense. And speaking of scaling, that's actually something I wanted to talk to you about. And that's the challenge with app store search ads is that even though it can be an incredible source of really high intent downloads, what everybody in the industry talks about, it's like, "Well that's great, but it's not scalable." So, how do you think about scalability in app store search ads? 

Dilip Reddy:

Scalability, obviously we cannot compare Apple Search Ads with Meta campaigns or TikTok campaigns because it's apples versus oranges. We are looking quality versus quantity. In my opinion, the best way to scale is to adding more keywords, high intent, low intent, segmented by it and adding broad match keywords that will literally be exponentially increasing our awareness and more countries by default. That's a easy way to scale and just monitor campaigns and employ some kind of a workflow like campaign structure should be in such a way that it's easy to maintain them. 

I'm not a big fan of creating hundreds of campaigns for one country because you only have one computer set of keywords. We cannot have multiple competitors, we don't need to run different ad groups. The bottom line is adding more keywords, relevant keywords, the more relevant, the better they are. And adding more there's always competitor share that we can improve. And brand is something, as we talked about it is its own thing. We do not consider brand as our, because brand is steady flow, right? It's a good thing the brand impressions go up, but that's how I scale. 

David Barnard:

Yeah. But then do you see, I guess it's just going to be so contextual, right? It's like for a health and fitness app, it's like if they're willing to spend the money, it can be a massive channel because there's ton of high intent people coming and searching relevant keywords. But then if you're in a niche, a really small niche where you're really only relevant to very small number of searches, then you're going to hit the wall sooner. So, it's just very contextual of when you're going to hit the wall, but then you don't know that you've really hit the wall until you've done the experiments with broad match and experimented with the bidding and tried stuff, right? 

Dilip Reddy:

Yeah. And Apple Search is one of those things where the competitors can dictate what we can do next. For example, imagine there's a weather app and imagine there's for some weird reason, there's no weather app in Apple Search Ads ever running Apple Search Ads. Your app is the only one that has Apple Search Ads, in that case your CPA will be probably in sense in US. Your ROAS can skyrocket, but there's dream scenario. But imagine you have 25 with competitors, whether app competitors in United States unfortunately means that for the same number of installs your CPA probably like is 10 times now. But the good thing is competitor's also dynamic, just because they're bidding for a keyword doesn't mean that Apple considers them relevant.

Same phenomenal brand as well. If your competitor running same with a radar keyword and for some reason your app is performing better and Apple can see that even if you're bidding half of your competitor, apple will still prefer your ad because Apple has to make money too. They can just show other app if it's not getting too much of traction.

David Barnard:

Yeah, that's interesting. And so, then part of what you're saying is even if you are super niche and you're like, "Oh, it's going to tap me out really quick," go get those keywords because you're going to be so cheap. So, if you're at a niche and you're not running Apple Search Ads, you're probably leaving money on the table because there probably are some high intent low traffic keywords that still you're just printing money if you can pick those up.

Dilip Reddy:

Yeah. We just need to adjust our expectations. We can have $1 million-day one Meta kind of blitz, but if we have our expectations set, and, yeah. 

David Barnard:

The interesting thing too is even though Meta you can theoretically scale even for a niche app, you'll saturate that kind of ideal customer profile. And so yeah, you can scale near infinite on Meta, but you might not scale profitably near infinite because it's similar as you start saturating your market and those high-intent users that you're going to be able to get to, then your cost air are going to go up too because now you're ending up with medium intent people and then low-intent people. So, nothing's truly infinitely scalable unless you're, I guess these Temu and stuff like that. When anybody is truly a potential customer, you can't scale to a level that a lot of apps can't, but most apps aren't that. 

Dilip Reddy:

And one of the interesting thing is that how ad channels can behave differently for the same app. Obviously, there any mature app company will be using multiple ad channels at the same time that Apple Search Ads, Google, what not. The difference is that Apple Search Ads is relatively easier channel to maintain. The reason is we don't have to create this influencer videos, make it quirky and there's a thing called fitting. Imagine you probably noticed that if you're using Instagram and you're seeing same ad every day, something will sit in and will never, it actually may create aversion to the app. But Apple Search Ads, we don't have that problem because we are targeting new user looking exactly for that keyword.

So, the ad fitting is not a really huge thing, there is some display formats like search tabs, today tabs, but it's not that extent of Instagram where we see 10 ads of the same app in a row and it's not that great experience. And it's easier to scale because we are not dealing with creatives, Apple Search Ads creatives is our whatever the screenshots we have, Meta dataset. And the keywords, we don't change keywords, it's always the same keywords. We may add it, we may pause them or we may be lower, but a weather radar until your app exists in the app store, the way the radar is always be relevant. We don't need to think creatively and it's great for people like us who are developers because we don't have to worry about creative videos that catch people's attention. 

David Barnard:

Yeah. Speaking of creative and app store screenshots and stuff though, how has product pages and being able to send specific keywords to specific product pages started to change the game? Because there is a certain level now where you can experiment a little more with, "creative on the app store," because you can create a separate product page for these different kind of use cases. Have you seen that? Help accounts scale and get to better ROAS? 

Dilip Reddy:

Those are good problems to have because if you're considering a keyword-specific customer hot pages, it means that we already maximized what are the impression share and we are reaching. Let's assume there's a shopping app and they maximized 80% of impression share for all generic terms. Where custom product pages can move the needle is, let's assume there is a demo that's running ad for Thanksgiving and it's like a month-long feature. But they won't say Thanksgiving in the title and subtitles and screenshots, but they can create each entire custom power page just for Thanksgiving and have Thanksgiving all across only for United States. 

And they can use it to scale up those Thanksgiving, people who are creating, why I'm saying Thanksgiving, it should be Black Friday, right? Black Friday deals and yeah, that way a demo can scale up those keywords. And in a indirect way, I have also custom product pages for A-B testing. This is a one easy way to simply implement A-B testing and see, you can even use similar search terms. For example, bid for half the United States for our default product page and utilize the new custom product page. We want to test and see which one performs better and just compare each. There shouldn't be much difference between East Coast and West Coast anyway. 

Or you can also try add scheduling, which means during the first half of the day you can use it. But I'm not a big fan because the time of the day may dictate, for example, if I'm a yoga app, I may be looking in the morning, right? So, geographical separation might be a good place to test. 

David Barnard:

Yeah. Interesting. All right, well it's been so fun chatting today, but we do need to wrap up as we do. Is there anything else you wanted to share or anything you wanted to tell the audience?

Dilip Reddy:

Yeah. For new apps, use Apple Search Ads as customer discovery thing and you don't have to spend daily budgets, like hundreds of dollars. You can even get five, six installs a day, spending $10, $20 but use that as high intent and use them as because you want to target high intent users and improve your app and check us out. Our platform is designed to scale Apple Search Ads profitably, it's called searchadsoptimization.com. And find me on LinkedIn and ask me any questions and I'm always happy to talk about Apple Search Ads. 

David Barnard:

Awesome. And as I've been wrapping up these podcasts too, if you have specific questions instead of hitting them up on LinkedIn where only the two of you could talk, feel free to do it on the Sub Club community. So, we will post this podcast on Sub Club community, you can respond to that or just tag Dilip so that your questions can be beneficial to the rest of the community as well. But yeah, I was checking out your company and again it's really cool that you started this from a data perspective and it's clear that you've built the algorithms to make it so much easier for folks to do a lot of things that we talked about today. 

To manage those bids, to automate some of that as well. So yeah, it looks like a really cool platform and something I'm going to be checking out myself. 

Dilip Reddy:

Yeah. Sounds great. 

David Barnard:

Awesome. Thanks for joining me and thanks for sharing all these great nuggets with folks. 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.