In this episode of Mobile Growth & Pancakes, Esther Shatz is joined by Stefana Pesko, User Acquisition Manager at Product Madness. They discuss effective and creative UA strategies, tools, and future trends.
Check out all the other episodes of Mobile Growth & Pancakes here.
To connect with Stefana:
Listen to the full episode:
You can also listen on:
And please – rate & review Mobile Growth & Pancakes:
“Listen to different people who have ideas that are different from your own. You can learn, change your mind, or work with that person to achieve something different.”
Stefana Pesko
Key takeaways:
- Every channel can positively contribute to your marketing mix. So testing multiple channels is essential. Though many marketers rely more on Facebook and Google, Stefana talks about how her team uses Reddit as a brand awareness platform.
- She also mentions how they use category takeover so that users can see their best performing categories ads for twenty-four hours and trending takeover, which aligns the ads with the most relevant and popular trends on Reddit.
- Last year brought many changes in the UA space, especially when iOS 14.5 was released. For example, measuring and effectively optimizing campaigns became more challenging. So UA specialists needed to adopt a more creative, strategic, and highly-technical approach.
- Looking for high-quality users may be difficult. Stefana suggests coming up with targeted offers, analyzing the results, and then determining how to capitalize on them. You can also find some motivators to attract your users, segment them into different categories, and use your creatives to get the best type of users.
Full transcript:
Esther Shatz: Welcome to Mobile Growth & Pancakes, a podcast by Storemaven. We break down how and why mobile apps grow. In each episode, we invite a mobile growth expert onto the show to break down a specific mobile growth strategy: how it worked, why it worked, and what they would do differently. I’m your host Esther Shatz. All right, welcome to Mobile Growth & Pancakes. I’m super excited to be joined today by Stefana Pesko from Product Madness. Do you want to introduce yourself, Stefana?
Stefana: Yes, thank you so much for having me on. I’m Stefana. I work at Product Madness for close to two years now. I work in UA and there’s a little bit of retargeting. Most of my channels are self attributing networks like Facebook and Reddit. Before that, I used to work in an agency. I had a lot of different verticals to look after, and it’s nice to be totally focused on gaming now, which I do quite enjoy. Before that, I used to work mostly in branding for big brands like Marriott and other
Esther: Wow. Big shift over from Marriott to Heart of Vegas and the world of gaming.
Stefana: Yes, it’s kind of been.
Esther: I agree gaming is more fun. I got to go straight to it because you mentioned when we talk about self attributing networks, I don’t think Reddit is the first one on everyone’s mind. I would love to hear a little bit more about how UA’s strategy goes on the Reddit site and how you’ve seen success there.
Stefana: Yes, of course. For me personally, testing is very important, and testing different channels is also important. A lot of us go straight to Facebook, and Google, and that obviously makes sense the scale, all the past learnings that they have, the algorithm is well-tuned, but just like with TikTok where most advertising students see a lot of sense and we scaled so high, we are the top gaming advertiser on TikTok right now. Might not be the top, but there is also promise on Reddit. Again, not every channel is there for some users, for the same quality of users or for the same strategy.
For instance, on Reddit, we found that it’s possible for us to use that as a brand awareness platform. There are a lot of other options, not just auctions on Reddit. We’ve seen great success with takeovers, category takeovers, and trending takeovers. I’m giving here a big shout-out to the Reddit team who are wonderful. We also had a TV activity at the same time for two of our apps and this really helps us to tie this in together with the brand activity and support the TV activity that way. The takeovers really helped us not just with a paid activity that obviously lowered our costs overall on the auction, but it also helped us increase our organic presence that was already influenced by TV.
In terms of, brand awareness, downloading tend and all those different points that we got on the brand side from the TV activity and from the Reddit that was done simultaneously, and the studies that Reddit helped us work with Kantar on obviously helped us a lot to understand what kind of users we’re bringing. Are those users the same users? Are they different types of users? What kind of value do those users bring? That was the approach that we found on Reddit. I’m sure there are a multitude of other approaches you can also find. It just shows you that it’s not necessarily that every channel is simply revenue-based. There are other things that each channel can contribute to your marketing mix.
Esther: I think that’s super important. I have so many things I want to follow up on here, but first, let’s talk about the goals because I think it’s a good point with Reddit. We’ve heard it with TikTok and even YouTube that if you’re measuring it by click to install or any kind of download activity, you probably won’t see the same cost-effectivity that you might see elsewhere, but you do have this branding effect. How do you create the right KPIs? How do you create a measurement structure to understand what your performance should be and if you’re hitting those goals?
Stefana: Sure. While we were doing the TV activity we had a specific data science team that was working specifically on measuring how our TV activity was doing. Alongside that Reddit was helping us work alongside Kantar which has their own point-based system on how they attribute the increase in your awareness, in your top-funnel metrics, in your intent to download, and also helps you see which users from which other genres are interested and what kind of users those are supposed to be. Because for us Reddit was the channel that we used just to get new types of audiences because we have an established audience.
Many of our apps are already a few years old and have their own biases, so we know where to go for the high-value users that we usually buy from tier one countries. Reddit was for us a way to find new types of users that are not based on lookalike audiences that we normally use on our biggest channel for instance. We know now with iOS that is a bit difficult. Just like everyone else, we were attempting to find new pockets. Maybe new geos’, maybe new channels.
Esther: I love turning to Reddit because I can tell you not from the UA side, but when I’m doing research into a new industry or a new genre, or I’m trying to get a sense of competition, I always go to Reddit threads. That’s like one of my favorite places to find what users are actually saying because it’s such a huge community, and especially for gamers everywhere, there’s so much there. I want to also follow up. You’ve mentioned that you use this concept takeover, can you describe that a little bit more?
Stefana: Yes, of course. We did two different types of takeovers that our Reddit team advised us on. That is also based on the stand that you are willing to commit to with Reddit. We used category takeovers, which are based on the best performing categories that we had in our auction. That’s how we selected which categories were going to in a way own for 24 hours. The only ads that they would see on those categories would be our ads for a whole 24 hours. Then we also used the trending takeovers, which basically take over the main Reddit page, the takeover page, with different sorts of ads. It can be your statics, your video ads, and expose and everyone who goes on the trending page, which is most of Reddit’s users, will only see your ads.
Esther: That’s very, very cool. It also sounds like the Reddit team really has a lot of inside info to give and make it more effective. You mentioned that you had TV campaigns going and I’m sure you had plenty of other efforts going on as well. Can you talk to me a bit about that mix? Is it important to have that media mix? How do you select which channels go together? Do they cannibalize each other? I know you mentioned you had data science working on maybe the TV side, do you find that you’re able to attribute different efforts, or do you not look at it that way and you look at it as the overall push and its reaction? Just any light you can shed there would be amazing.
Stefana: Yes, of course, and it is a bit of a complicated topic. We had a whole specific team just working on a TV activity that was from the beginning with creatives because they have to be very specific and it’s really important what kind of creatives you use. It took a long time to finally find and create something that will definitely work. Then there was a data science team that was only working on measuring the effect of the TV activity on the overall brand, organic, paid, and everything else that comes with it.
I don’t know the details because I was mostly on the P channel supporting the TV activity, but what we do see from the paid side, you can compare how your KPIs and your main metrics react to the TV activity. Sometimes you see a spike in your costs. Sometimes you see that you’re getting more deposits or your deposit percentage is increasing or ARP or PPU is going on. That combined with the fact that you can see on paid can give you an overall picture.
Esther: Yes, I feel like every team that doesn’t work directly with the standard UA channels is so jealous of the amount of data that UAs had until now, but I do think it’s worth it, I feel like this is what I’ve been talking about every day for a year now. Obviously, UA has changed a lot in the last year, a year and a half, and is continuing to change. Google has recently announced the changes with their custom store listings and everything going on there, and they have the privacy sandbox that they’re working on. How do you feel the industry has changed already? Do you feel like there’s further to go? You as somebody who’s been managing UA for such a heavy spender, how do you feel like things have changed?
Stefana: I don’t know where to start, everything has changed to be fair. Just on the iOS side, I think it’s important to never lose track of the fact that the auction is still there and people will still advertise on iOS. I know some companies after the Apple changes in April last year pulled out of the iOS auction together, which I don’t believe was a great idea considering now that they’ve lost market share. It’s incredibly hard to get back into the market. Obviously, costs have risen. We have seen them come down after the initial first few months and after the implementation of the correct CV schema, whichever schema does work for you if you’re on SRN or however you manage, probabilistic if you’re on DSPs or any other networks.
I don’t even know where to start to be fair. If you have managed to find a CV schema that does work for you and again, that is still ongoing. I believe in the next year and a half, there will be something that works a lot better than what we currently have. I’m assuming just like us, most of the advertisers base their CV schema on revenue or revenue plus engagement events, but there are also things like we tested, which is the predictive CV schema that isn’t necessarily probably the best tune right now.
For some time I know, big advertisers like Facebook are looking to integrate that into their revenue-based schema. I do think that will then definitely help crack whatever’s left of our current iOS strategy. If you do combine revenue and predictive modeling with that lot of learnings that you have now in Android, you just have to transfer an iOS, because your creative testing isn’t best done on iOS at the moment. A lot of those learnings thankfully are transferable.
Depending on how you do your creative testing, that can really help you a lot. Also, different channels have more presence on iOS versus Android and that’s just based on our free-to-play social casino genre. For instance a lot of the time you see that TikTok works really well for iOS users. Maybe just a little bit less for Android users. That helps you fine-tune your strategy. On top of that, I think working with partners that use automation or machine learning services can really help you.
You can see that in the push from Facebook, for instance, in using AAAs that has been really fine-tuned and uses broad audiences, which helps you balance with the look-alikes that you’re using in your normal campaigns to attract A-list users and would that is more or less going to be the focus of advertising. You can see all the algorithms and how they’re looking, how many prospects, and events you’re getting.
Working with a machine learning partner, we’ll definitely help you save time, be more cost-effective, and efficient and get different learnings that you might not have gotten initially.
Esther: For sure. I think other than the measurement side, which is, I mean if you haven’t changed what you’re doing in measurement, you’re not optimizing because there’s nothing left that you could have possibly been doing the same. Do you look at kind of that even before you get to the measurement side at the creative journey at audiences, are you looking at those any differently since these changes came into play?
Stefana: Definitely. That isn’t necessarily to do with the changes that came into play. We just optimized our creative testing, which I’m actually very proud of. I think we have a wonderful system now that we use for creative testing. We segmented into different phases and our data science team and our AA partner did help us find thresholds that we need to pass for each phase in order to move on to the next.
In the first phase, we test our upper funnel metrics and all the basics of your CTRs, CVRs, and IPMs to see how many impressions you’re getting. What’s the number of installs you’re getting for that obviously TPI also comes into play. Then whichever creatives are best on those metrics. We move to our lower final metrics test, which is what we all care about the most the revenue-based metrics.
You look at your ROAS, your APU, or PPU also your retention, and depending on how you look at that maybe day seven day 30, whichever for creative testing, I would suggest shorter-term metrics we don’t want to test for too long.
Esther: You got to optimize fast. Yes.
Stefana: Yes, exactly. We all have so many creatives and so many new concepts are hitting and competitors are doing such a great job on testing different types of creatives, not just playable, but also tutorials are creatives, influencer creatives, name style creatives. There are so many options right now. You need to test fast and you need to be optimized in your testing strategy. Then we move on to phase three where we inject our best creatives based on those two phrases into RB.
Esther: Very nice. I’d love to touch a little bit more about creative types because I think that’s one place where we’ve really started to see a lot of innovation. Just that kind of how do we create in the absence of having maybe the lookalike audiences that we were used to, or the targets that we’re used to, we have creatives that we’re using to speak to different people, have anything super interesting that you’ve maybe new formats, new ways of looking at that you’ve tried in the last year or so?
Stefana: We’re trying to learn as much as we can into playable. I do think that’s a wonderful way to get users to experience the app without actual commitment. That’s what most users are actually scared of. Totally understand your usual gameplay with rewards-type ads works best most of the time. We found out that those innovative ads are great to test and they really help with your upper funnel metrics but when it comes down to it.
This may be just because of the genre that we are in, your iterations of your best creatives are the most cost-efficient, effective creatives you can ever have. That doesn’t mean that you shouldn’t be testing other types. We’ve seen a lot of success recently with influencer creatives or the TikTok style creatives, which work amazingly across Facebook, and TikTok other networks.
There are definitely a lot of other things you can and should be doing because relying just on one creative, no matter how many times you iterated, the algorithm still recognizes certain pieces of a creative and it will prioritize the longer standard creative over others. You don’t want that to happen to you. Test as much as you can.
Esther: You need your safety net over here and then you’re testing over here and you never take this one out. You just have that little testing subgroup.
Stefana: Exactly. You want balance at the end of the day, right? You want to have new creatives, but you also want to make sure you have the creatives that perform really well and you know they will perform.
Esther: Sure. I think with social casino, I mean it’s true obviously of other game genres as well, but the social casino, especially there’s that challenge of excellent top-of-the-funnel metrics that might even decrease your metrics further down. How do you go through that process of really finding the quality users without losing the critical mass and without the platforms are still looking at who’s downloading, so those metrics still matter, but how do you match that balance and not lose out on those whales, the high-quality users in exchange for that free coin mass who are coming in for that side?
Stefana: Yes, that’s actually funny because whenever we look at our upper funnel metrics, our CPR CTRs, some creators perform really, really well. Then you don’t really see them perform that well on your lower-funnel metrics which isn’t normally what you would expect, because if the user clicks and installs, you would expect them to be interested. Then again, there are other things that come into play from the product side of how they retain users and so on.
One of the good things that you can do is work with your product. They can help you work through events, particularly events, some are aimed at new users, some are at existing, some are at both, there are live ops, and there are so many offers that you can get into your creatives. Then the product team can also help you with a postmortem analysis to see how it influenced which type of users and how you can capitalize on that.
Another thing you can do is your partners for us Facebook at large it’s they really have a lot of options and ways to sponsor you and help you. Recently we did motivators testing. We try to find which motivator is the best to attract our users. We found that different types of users, of different ages and gender and different values in the short and long term, are attracted by different things within your app.
Once you’ve done the test, you can do so many iterations in order to use those motivators that worked best for you and then segment your users into different categories and use those creatives to get the best type of users that you’re looking for.
Esther: I have to say, it seems like you’re very on top of utilizing other teams in other departments, and really it’s something that we don’t see all the time because it’s so easy to fall into your own KPIs and the things that you control and just focus on hammering that out. It sounds like you’ve seen, which makes total sense, a huge amount of success, not just working with your teams internally at other stages, but also working with the teams outside who know their area and are willing to help. That’s awesome. I’d love to go back to the in-app event live op side because live ops are expanding, now they’re beta and I’m sure it will be in a matter of time available all across the board.
Metrics have been really interesting. How do you, from the UA side, look at in-app events and live ops, and how do you optimize from your side while still understanding that you’re completely tied to the product team as well?
Stefana: For us, most of the live ops and events are just utilized in creatives. Some of them run for the short term and we dedicate specific campaigns for them. Just so we can get more new users into that specific user cycle. And then they hopefully retain as well. We did have a lot of event-specific creatives and live ops-specific creatives for all our apps and we tested how each of those events works for our audience. Then we found out that there’s a specific timeframe in which we need to run those creatives. We get enough learnings, those creatives also do tend to be more expensive. Then you need to talk to the product and understand if they’re going to rerun the events if there was a specific offer you can’t use anymore.
There’s a lot of work back and forth between the product and marketing in order to optimize the way that you do use what they have in the product. The product also can help you run specific events that are potentially aimed more at new users, which you are trying to attract rather than their long spending users, which they just want to capitalize more and more of. Definitely work with your product. They have a ton of events that can help you. Not all of the events will work. That doesn’t mean you shouldn’t test.
Esther: I think it’s such even when events don’t work, it’s, there’s so much to gain especially live ops and they just released some numbers on the Google side, got to take it with a grain of salt but just insane increase in day 28, increase in immediate downloads. There’s so much potential there on every side that I think it’s a new feature and not everyone has put it as part of their routine, but it’s such a critical thing to start looking at and learning how to play around with. Speaking of the connection between product and UA when you look at an, or a lot of your games are pretty long-standing, but when you look at earlier stages in an app’s life cycle which typically tend to fall entirely under the product domain, where do you feel that UA can come in and assist in the world of pre-launch soft launch whatever you have going on there?
Stefana: I actually have worked on a soft launch, not so long ago and there is possibly other news coming around about our other launches so soft launch was a really exciting time for me. It was going from advertising at a large scale and looking mostly at revenue metrics to looking mostly at retention into your three markets and basing your advertising spend on which countries, your partners help you identify as the potentials for benchmarking your day one day, day seven retentions, your long term retention, and then moving on from that to something completely different. Then look at your monetization whether it’s year three or two, one market, whichever you are trying to expand that. That was a lot more work with a product than you do on a daily basis.
I think it’s good learning for probably every marketer or everyone who works in the marketing or advertising space. That helped, I think the whole UA team establish more of a connection to the product, which is a good thing long-term. Again, the soft launch and any other tests that you do afterward are not entirely under UA control cause the creatives, the game economy, the user life cycle, most of that is just decided by-product and they have the specific APIs that they need to hit. Whichever market you are in and with specific benchmarks, your partners can help you a lot with that. For us, it was mostly looking at the basic metrics of CPI and retention which is what you started with for soft launch, but it was a really good interesting learning.
Esther: Awesome. I think with soft launch sometimes, first of all, the KPIs are so dramatically different. You mentioned CPI, which maybe is a metric that doesn’t usually matter, but when a product isn’t fully baked, then, you got to know where is it you’re bringing in the wrong user and where is it that the user doesn’t have what they need which they will, once the product is developed a bit further, but I like that you said it’s almost a cheat to make sure product and UA work together because you have to, when it’s a soft launch, it’s a nice hack for co-teamwork bonding. I think we’ll go into the rapid-fire round with just one more question, from the side, what would you say is the next big thing coming up in the world of UA? So we’ve had a massive shift and we understand that there’s privacy and changes in measurement. What’s coming up. What’s the next wave?
Stefana: Well, the most obvious one would probably be the different shift that Android is going to go through but there are a lot of things you can do to mitigate that. Whatever learnings you had from your iOS gains or pains you can transfer to your potential learnings into what’s coming up for Android. In the meantime, I’d say, definitely do a lot of testing and different channels of creative testing don’t forget machine learning and automation that will all help you in the long run. It’s easier and more cost-effective. It just needs a bit of optimization in the beginning and then you’re good to go.
Esther: You also hit the nail on the head with that. There’s that fear where you pull back spending and you pull back because you think it’s changed and it’s unknown and you mentioned it before, but it’s totally right. You put yourself at such a, first of all, you get lower bids because everyone else is chickening out. Then you have less competition, but the loss that you have to make up for is so monumental and realistically you’re not going to stop advertising forever. It’s not something that’s going to happen. I think that’s super important whatever that wave is is to be able to take that risk and say, I’m not going to be able to measure the way I did and I’m not going to be able to do what I did, but I can’t stop. I don’t get to retire or think about it and go back in because I have such a disadvantage to start from there ready for the rapid-fire round?
Stefana: Yes.
Esther: All right, you can stretch out, get ready. First of all, if you could give just one tip to an aspiring growth market or what would it be?
Stefana: Oh, just one listen to different people who have ideas that are different from your own. You can learn from that and you might even change your mind or work with that person in order to achieve something different.
Esther: Very nice. Your favorite mobile growth resource.
Stefana: Mobile Dev Memo probably but I also do E-marketer a lot and I spend a lot of time on Sensor Tower and Marketing Week and Ad Week. I don’t know if there are a lot of them.
Esther: It’s good, but Mobile Dev Memo it’s the Bible basically.
Stefana: Yes, that’s what I hear from everyone, for sure.
Esther: Definitely, who is the person in the mobile industry that you’d most want to take for lunch and why?
Stefana: Okay,if I had to take someone from marketing, I’d like to even just have lunch with me, someone from Apple, I really want to know what and how they do their things, but if it didn’t have to be a marketing person I’d like to meet the creators of mobile games that I was obsessed with as a teen which were the Angry Birds and the Fruit Ninja. Those were so addictive I’ve never played something for so long.
Esther: Fruit Ninja took up such a huge percent of my time when I had no time to give to it, but wow. That was my first. That was also the first mobile game that I really stuck with. That’s a good answer.
Stefana: I know I don’t even remember the progression on that game, but I was just addicted to doing it.
Esther: I just remember slicing the slicing that’s all I remember. The most important question of course is what is your favorite flavor of pancake?
Stefana: I have two, I recently tried Baklava pancakes and there was so much flavor there. I know I couldn’t finish the portion. It was extremely filling, but that’s something I still think about over the weekend. If not that, I have a basic pancake that I like with blueberries.
Esther: Blueberry is a good choice but we’ve asked a lot of people about pancakes and that is by far, I think, the most unique type of pancake that I heard, amazing. Thank you so much for joining me today and sharing so much. Where can people find you if they just want to learn more, and see what you’re up to?
Stefana: Yes, for sure you can find me on LinkedIn. I take it semi-regularly so we can probably connect there.
Esther: Amazing Stefana, thank you so so much. This was awesome.
Stefana: Thank you for having me.
