ABOUT THE PANELISTS: Diane Germann | Ekaterina Gamsriegler | Gaurav Bhattacharya | Nataliia Drozd

ABOUT ROCKETSHIP HQ: Website | LinkedIn  | Twitter | YouTube


KEY HIGHLIGHTS

😼 Data-driven approaches and the impact of small creative changes to combat creative fatigue.
💯 Struggles faced by subscription-based apps due to the complexity of SKAN data.
🗯️ Data interpretation and the impact on traffic post-ATT.
🦾 Google vs. ASA.
👁️ Borrowing practices from product and engineering teams.
🧞‍♂️ Strategies using AI tools, from design to AI-generated VOs.

FULL TRANSCRIPT BELOW

Shamanth 

Welcome to the first live recording session of the Mobile User Acquisition Show. 

A couple of you have been guests already on the synchronous non-live version. I’m excited to have this elite panel here today. Please join me in welcoming our amazing guests – Diane Germann, Lead Performance Marketing Manager at Blinkist, Ekaterina Gamsriegler, Head of Marketing and Growth at MIMO, Natalia Drozd, Marketing and UA Lead at Fabulous, Gaurav Bhattacharya, Performance Marketing Lead at ZenJob. 

I’m your host for the day, Shamanth Rao, CEO at RocketShip HQ. 

Let’s jump in Gaurav, one of your big wins lately has been combating creative fatigue on UAC and Meta. Talk us through what you’ve been trying, and what you’ve seen work. 

Gaurav 

We usually have a lot of strategies which we employ but I would dial it down to mainly two strategies which worked well for us.

The first thing was to have a very data-driven approach to analyze our creatives and give feedback to a creative team. One of the things we do is to analyze the metrics of our creatives on both the upper funnel level and the lower funnel level. So upper funnel metrics like clicks, CTR, we check those things, but we also go to the lower level. So we look at in-app events like signups and so and when we look at those things we try to see where exactly the creative fatigue is coming from.

Is it due to upper funnel creatives which we use, which are not working for us? Is it the use of certain color schemes which are not working for us? We go ahead and try to make sure that we can give the right feedback to the creative team. And for that, we do a lot of testing. We do a lot of tests on Meta and on UAC to figure out how we can find the right balance.

We also take the wins and the losses from different channels, which is important because if you stick to one channel that might not be the right approach.

The second approach we take is to make sure we make very small changes in our creatives, which always sees a really good reaction to the user-centric behavior we see of our users. Let’s say, for a specific CTA, we change the color scheme a little bit, or we can change the stock image, those small changes which we might think might not work that well, actually work quite well and it turns out to be quite beneficial for us.

Shamanth 

Interesting. If I might double-click on an aspect of UAC. Because UAC is very black box-ish. And you’re not getting clear asset-level metrics. So how, if at all, are you triaging fatigue data from UAC.

Gaurav 

Yeah, that’s a good question. I would say because the kind of data which Google gives us is not conclusive, but you can always use scripts or you can use rules and figure out what is happening on the creative level as well.

Like taking the upper funnel metrics, as I said, you have three different columns where you can see if the creative is good, best or it’s not performing. So if you’re doing that, that’s fine, but you don’t get the real message.

So, for example, let’s say there could be a creative which has low CTR, but it’s giving you signups. So you can try to check those matrices. You can talk to the Google team, they can give you this sort of data. You can get that data and then you can decide what you need to go for. 

One of the things we did from that data was to personalize our user-centric approach to ads. What we did was that we saw the kind of creatives which we are using, reaching creative fatigue quite early. So we started mixing it up. We do a mix of performance-oriented creatives. We do a mix of UGC creatives, which always works for us.

Shamanth 

That also ties in with what Nat, you had described in one of the editions of the Mobile UA show, which to this day is one of our most popular episodes. You described an approach in that episode, which very much mirrors what Gaurav talked about, which is diversifying the asset mix. How has that approach changed in the last two or three years? How has that approach changed post-ATT? 

Nataliia 

So first thing, Google ads are killing it. It’s a little bit scary, but the truth is, that the amount of ads being shown on Google’s different placements is continuously increasing. This is why Google does this to increase its revenue. And it’s actually very, very profitable for them.

It was one thing that made Google better than Apple because, on ASA, there are ads under every keyword if you want to bid on it. On Google Play, brands were protected for a very long time, so brand searches were protected by the brands. It’s not a thing anymore, so no matter if you’re a Duolingo, Mimo or Blinkist.

Watch out for competitors. If you don’t invest in UAC, your traffic may be stolen. Of course, there is cannibalization and incrementality, but still, many people click on the thing which is the first.

So one, it is more placements on Google Play.

It’s not the answer about creatives, but it’s more the answer about what Google does. Google knows the placements that work best, and it’s Google-owned placements, and it’s capitalizing heavily on them. Talking about iOS, from my personal experience and from the experience of talking to Google people, there is very, very little success for subscription-based apps because of the lack of signal, and because of the pure messiness of data.

The way Google interprets SKAN, the way MMP interprets SKAN, it makes it just super hard to measure. So It just makes it hard to use Google as an ever-going, continuous source of traffic. I’ve seen some people who do it, but they mainly do it on YouTube. 

Google knows how to make money, and it always makes it harder for us to analyze things by not unlocking the black box, but the main rules that were two years ago are still here. 

One more thing I want to mention is that you can slightly improve user experience by using custom store listings. This is this beautiful synergy between UAC and ASO by creating themed topics on your Apple, on your UAC campaign for a specific ad group and creating better screenshots on your Google Play Store that can potentially lead to better results. 

Shamanth 

Sure, to follow up for those in the audience who may not be aware, can you elaborate on what you mean by lack of signal and broken data on iOS?

Nataliia 

What I mean by this is that Google cares about data points. Yes, creative is king, content is king. But Google is built by engineers, not designers. This is why if you don’t have a good SKAN schema And if you’re not willing to invest billions to feed the system with enough data points which are being obfuscated by SKAN then it’s very hard. 

So basically the ones who suffer are small and medium businesses that are not willing to invest millions and wait for two or three years to get their money back

Shamanth 

And something I’ve noticed is Google very much privileges Firebase.

Nataliia 

Always! Google reps get a bonus every time they sign a client that switches from MMP to Firebase. Why? Because Firebase is their own system, and what they are trying to do is to enrich the quality of signals. But what it also means is that we are giving Google all our data for free.

Shamanth 

Yeah, yeah, and I have heard it said that Firebase may be in some way looking at device signals that may or may not be forbidden by certain other companies.

Nataliia 

Never. Trillion-dollar companies always follow all the compliances.

Shamanth

To switch gears a bit, so we talked about UAC, we talked about Meta. Diane, you work and you focus a lot on channels that some would consider not as mainstream primarily Taboola, Outbrain etc.. For folks that are not as plugged in, can you talk about what the typical user flow is on these channels are and what are some of the parameters or elements that you see contributing to your success on these channels?

Diane

We’re probably not the biggest ones on Taboola and Outbrain, but definitely within educational verticals, or also within apps. But to speak on them specifically, native advertisements are what they do.

So, it’s a lot like display ads, but it’s always going to be on a news website. What you’re trying to achieve is having it be as similar to a news article as what people are reading. They click on it. It’s kind of a seamless experience.

The biggest metrics that we focus on, I would say is, making sure that above the fold image, everything, the first paragraph, the CTAs and I think all the content that goes into that, that’s where we’re optimizing the most. So for us, we have, what we call TPAs, Top Performing Articles. We try to make it like an article format. And we have MPAs, like mid-performing ones. So those ones we’re doing the most optimizing and testing and bringing them up to the best performers.

Shamanth 

Of course, the Top Performing Articles, they’re at the top of the funnel. It’s my understanding that a couple of possible flows that can happen. You can send users to a landing page. You can send users to the app store. What in your experience do you see most effective? And maybe that’s not a single answer.

Diane

Yeah, I mean, for us,

we always kind of built our own magazine and so it’s a landing page that we send people to. They have that experience. And then for us, the main thing we’re doing is sending people straight to the app store.That only helped us for, you know, mobile, but now we’re doing also the onboarding funnel for desktop.

Shamanth

And for folks that are not Blinkist, and if this is something they want to be testing, are there considerations you’d recommend keeping in mind for them to decide whether to send people to the app store? Let’s just say they wanted to set up an MVP plan. What would you recommend that MVP plan to be to test a big content channels? 

Diane

It’s a good question. I mean, the most important thing to do, in my opinion, if you’re going to start a channel like this and potentially any channel, I guess, would be obviously doing competitive research first so finding out what people in that area or that within native ads or on news websites have similar to you feel. And then I would go off replicating that. So for us, obviously it’s books and it’s focusing on that. But if you are more in a fitness area, maybe you would have made those tweaks based off of that.

Shamanth 

Katya to go over to you. A lot of your work is focused on marketing and growth holistically, not purely just top of the funnel marketing per se. I think your perspective is particularly interesting because you have a view of what’s happening on the product and growth side as well, and I understand you’re able to borrow some practices from the product and engineering teams to help marketing teams operate. Are you able to share what those look like?

Ekaterina

Yeah, of course, the thing is that I think it’s quite the common story that when you have some initiative that you want an engineering or a product team to work on and you go to them and suggest that, very often you will get an answer that they have no time, no resources or that it’s already in the backlog.

Sadly, it very rarely gets out of there. What I see is that with lean marketing teams, like ours at Mimo, it’s also quite common that we have scarce resources and we have to prioritize ruthlessly. So what really helps us is to have this so-called agile marketing approach.

So we have bi-weekly sprints during which we plan our work. It’s not too long, to get stuck in and to not be able to make changes and adapt if needed. We also borrowed some of the rituals from the engineering teams. I don’t think that to be successful you have to borrow all of them. I personally find doing the retrospectives after every sprint fairly useful because this way you can again learn from your mistakes, figure out why things did not go the way they were planned. 

Did we plan for too much and under deliver? This helps you learn. Another thing that we are definitely borrowing is to decrease the amount of time wasted as much as possible. Asynchronous communication helps a lot. That’s why we have a set of various templates for various tasks, which means when we are planning the sprint and coming up with the sprint stories we try to fill in most importantly, the why section and the potential outcome, for the whole company to see, and for us to be very transparent, why we’re actually working on this instead of a hundred other things we could be possibly working on.

Shamanth:

What does a marketing sprint look like? What would it contain? 

Ekaterina:

Every team member makes a plan of what they will be working on, and we have a mix of different stories for app store optimization, experimentation, with visual assets, and improving our potential keyword rankings.

Maybe pitching for features, then we have the paid user acquisition section where we also plan all the gross experiments in this area. 

I also document a lot of learnings. So we have this very neat board, which is also visible and transparent for the whole company where you can see the things moving throughout the sprint.

The very interesting column for me personally is the results section because it’s exciting when something moves from waiting for results to being done. This is when we also share the learnings with the rest of the company of what worked and what did not.

Shamanth:

What I liked was what you briefly mentioned, the word “asynchronously”, and I think that’s a very critical part of all of this, to be able to execute this without being swamped in meetings, all along. The way I’ve thought about this is, it is so important to have a writing-heavy culture, which you also expressed in terms of documenting a lot of these sprints, which can unlock a lot of efficiency. So thank you for sharing.

Ekaterina:

Yeah, I feel like you really need to try to strike the balance between pushing for this over bureaucracy, making your team document everything versus documenting the stuff that really needs to be shared with the rest of the company to explain why we’re really doing all this.

Shamanth:

And Nat I know a lot of your work is with your product company and your product teams. Are you able to share what the collaboration looks like cross-company and across marketing and other functions on your teams?

Nataliia:

What I think is the most important thing is to over-communicate rather than under-communicate. There are two things – I hate meetings and I hate repeating meetings. But at the same time, when there is an issue, I would rather call a person. Let’s talk and figure it out. And it actually works better than long slack threads.

So, it’s all about balance, like everywhere. The balance of enough meetings, or at least being approachable by other people, and not getting stuck in this over communication. The other thing is when you have beautiful documentation, but when you start doing things step by step, it turns out the documentation is right, but the implementation is bad.

So it’s always good to spend some time seeing if actually things that you wrote don’t work the way that you expect them to work. 

Shamanth:

I’m not sure that very many people here love meetings. 

Nataliia

Meetings do not have to be boring. We have a really beautiful team culture. When we have marketing all-hands meetings, we always start with either a Kahoot or Scribble just to make it fun because we are humans and it’s always so fun to share. And well, especially in the remote companies and Fabulous has been remote since day one.

So for the last nine years, we have a meme culture, we have a background meme culture. So it doesn’t sound like we are distant, quite the opposite. And by that, we try to make meetings informative, but also fun. 

Shamanth

For sure icebreakers definitely make a huge difference. And also to add to what you said in terms of over-communication, I think asynchronous over-communication actually doesn’t take so much time in my humble experience.

At least on our team, if I need to even yell at somebody, we’ll just do it via Loom. I think I have learned to be a lot more patient with asynchronous because it’s my experience that just, we get a lot more done if I just send a loom, and wait for the other person to get back in 30 to 45 minutes or so. And it’s just better for everybody’s mental health and sanity. 

Nataliia

The best thing about Looms is when you receive a notification somebody watched your video. 

Shamanth:

To switch gears a bit, Gaurav to touch on growth marketing aspects again, I know you’ve seen some very significant wins from first-party data. Maybe it’s just me you know, I hear the phrase first party data used a lot. Unpack that for us and tell us what that means, what that meant for you and the way you use it on a day-to-day tactical level. 

Gaurav

When we started looking at our marketing funnel, we wanted to see how we can declutter the data which we have, data is messy and that’s how we play around with it. But we wanted to set up a framework to see how we can make that better, how we can get data which works for us. And also, in the coming year, we have Privacy Sandbox, we already have SKAN. So all these things are making life difficult for everyone, but we need to find ways around it to make it better.

So that’s what we did. We like we got a CDP And when we got the CDP we realized that we can play around with a lot of first-party data, especially because we have all the sources since we can connect it to everything, from our data warehouse, data from our platform data, like Google, Facebook, TikTok and everything in one place.

Once we have all that in one place, we know we can build audiences. Building audiences is something which everyone has done a lot of times. But this is something which we wanted to take a different approach on. One was that we did a dial-down on which particular users we wanted to target.

So we. took a funnel-based approach. We took users who, let’s say, signed up but did not take the next step. We targeted them, we took that data, we fed it back to the networks, and we created lookalike audiences as well. Recently, we were doing this test with the new campaign template Google has for lookalike audiences based on that first-party data.

So the way we look at it is to make full use of whatever we have in-house rather than depending on whatever is coming from the MMP, from Google ads or Facebook or whichever channels we have. That also helps us to make sure that we know which channel is working or at which stage of our funnels.

One channel could be really up a funnel and we know that we can use that data and leverage it to show ads on another platform. So these are the kind of strategies which have really worked for us. 

Shamanth:

Certainly, and you’re making the most of the data that you do have available, and you need your product and that’s uniquely distinctive to help you guys. 

Diane, you spoke about how you guys have seen a lot of events from using AI tools for content and idea generation. Talk to us about what that process looks like. And if possible, if you can share any non obvious wins that you’ve seen,

Diane:

I think that when using AI tools there’s the simple ways that we choose to use them where I think it’s like just fixing creative ad text just slightly and populate this five times, I think that’s the good way of using it. But then feeding it the information that you already have, like we have 27 different categories, for example like science, history, politics, things like that, feeding that kind of information. Now give me different ad text with this information. 

And take that to another level, just doing the visualization, whether you’re doing an image, or whether you’re doing a motion ad, or a video. I think it’s easy enough when we use AI and we’re just like, “hey, give me a woman who is in a hallway.” We all know that doesn’t work very well. It’s a struggle, and the eyes are all weird. 

But I think the more you’re using these tools you should give better input, for example – so this is my best creative. I put this one in, and now I just want you to change the background color. I just want you to apply it for a older generation, or I want you to make this person look slightly older, slightly younger.

We all know that that is the best way to test is to have the very minor changes. And that I think is something that we’ve definitely learned, and we are capitalizing on more and more. 

Shamanth:

Yeah, it ties into what Gaurav said in terms of using your own secret sauce. Somebody I know described this very well. They’re like, in any AI tool if you input garbage, you’re going to get garbage. You are the secret ingredient, and if you don’t put you into the input, it’s just not going to be helpful or effective. 

Katya, are you able to speak to any of the ways you’re using AI tools on your team across the funnel?

Ekaterina:

We tried stuff at the top of the funnel for tweaking the copy, double checking the copies a very nice use case especially if you have non-native speakers like this really helps us avoid a lot of mistakes. Also, we used AI to add voiceover to some of our video creatives together with like we also designed an animation.

To my surprise, this worked well. I did not expect to see a change in click through rates, but it happened. It worked quite well. Speaking of the use cases, our CEO, Johannes actually created a version of the bot which is trained on our internal style guide and on our internal communication so when you create some new piece of content whether it’s for the app store or for the ads, you can run a quick check, and it will give you advice on how you can fix it. For example, to avoid the use of emojis, or some other some other types of input. So this has been quite helpful for us at the top of the funnel. 

Shamanth:

I know we are almost coming up on time. So let’s do a quick rapid fire round of questions. Since AI is very top of mind for everyone, what are some non obvious strategies you’ve used for AI tools or a non obvious tool that you have used. Gaurav, let’s start with you. 

Gaurav:

In terms of applications which we have, Microsoft Designer has been working quite well, which I personally use. It’s a very good alternative to Canva and I love using it. I don’t need to bug my creative team to make this small change. I can do it by myself with simple prompts and it works. 

Nataliia

I use AI when I need to write a long email saying no in a polite way.

Ekaterina 

I would still say using it for voicing over our creatives was a non-obvious, but interesting win.

Diane

I would double down on that. When I have a really detailed thing that I know so much about and I want to give an update to the company, I will say – please condense this into two sentences.

Even as a native English speaker, that can improve and make it more readable for everybody else. 

Shamanth

I think that we are very close to time. So thank you so much for the amazing insights and inputs today. Thank you everyone for being a part of the first live recording of the Mobile User Acquisition Show.

A REQUEST BEFORE YOU GO

I have a very important favor to ask, which as those of you who know me know I don’t do often. If you get any pleasure or inspiration from this episode, could you PLEASE leave a review on your favorite podcasting platform – be it iTunes, Overcast, Spotify, or wherever you get your podcast fix. This podcast is very much a labor of love – and each episode takes many many hours to put together. When you write a review, it will not only be a great deal of encouragement to us, but it will also support getting the word out about the Mobile User Acquisition Show.

Constructive criticism and suggestions for improvement are welcome, whether on podcasting platforms – or by email to shamanth@rocketshiphq.com. We read all the reviews & I want to make this podcast better.

Thank you – and I look forward to seeing you with the next episode!