In this episode of Mobile Growth & Pancakes, Esther Shatz is joined by Offer Yehudai, President at Fyber. Offer details the changes we will see in data sharing and contextual ad targeting for mobile ads after the IDFA deprecation. He also lays out his predictions about if/when we will see a recovery in mobile advertising during the current fiscal year.
Check out all the other episodes of Mobile Growth & Pancakes here.
To connect with Offer:
Timestamps:
00:57 – Introduction to Offer and Inneractive (now acquired by Fyber)
01:52 – In-app Mobile Advertising today vs 15 years ago
03:24 – How to optimize efforts and redefine advertising success without IDFA?
08:38 – Contextual advertising and data sharing after IDFA deprecation
13:34 – Advertising on pre-loaded apps without IDFA
19:37 – KPIs for success in native-app advertising
20:33 – Predictions for performance marketing for mobile advertisers
24:50 – Quickfire round
Listen to the full episode here:
Or listen on:
Apple Podcasts
Spotify
Google Podcasts
“For pre-loaded apps, what’s really unique is the scale because, for the right advertiser, you can basically put yourself on millions of devices very very fast”
Offer Yehudai
Key Takeaways:
- Offer has spent 15 years in mobile advertising. His previous company Inneractive (now acquired by Fyber) was among the first companies to provide in-game advertising on Nokia devices.
- With the launch of iOS 14.5 and IDFA deprecation, the important considerations for brands are how to measure privacy and performance.
- Offer predicts that after users upgrade their iOS devices, the majority of traffic will come from users without IDFA consent.
- To overcome this, advertisers and marketers will have to implement context-based advertising without violating user privacy in any way.
- Publishers will segment first-party data and control ad exposure to specific user categories by putting a premium price tag on this data.
- Publishers will engage with mediation platforms to create a “garden of data” to facilitate cross-platform data sharing.
- For pre-loaded apps, the focus will shift to evergreen apps that tend to get used/played for a long time. Notifications and cross-app retargeting can help increase engagement on these apps.
- For native-app advertising, the KPI measurement will be specific to each advertiser but it will have to be measured in a wider attribution window.
- Offer predicts adoption of new channels and recovery in performance marketing by the end of the current fiscal year.
- Offer is not a pancake fan but if he had to have one, he would choose Nutella Banana pancake.
- Offer would love to take his team out for lunch because he hasn’t met them in person in a long time.
- Offer’s advice to all aspiring growth marketers is to become friends with their CEO to ensure coordination among the highest-level teams.
- Offer’s favourite resource for Mobile Growth is Mobile Debt Memo (MDM)
Full Transcript:
Esther: Welcome to Mobile Growth & Pancakes. I am very excited to be joined today by… I’m going to say it the Israeli way, so bear with me, Offer Yehudai is president of Fyber. Yes. Offer, can you introduce yourself quickly to everyone listening?
Offer: Yes. Hi, Esther. My name is Offer. I’m president at Fyber, the mobile and mediation company recently acquired by Digital Turbine. I spent the last 10 years in the Bay Area, in the last 15 years, if you can believe it, in mobile advertising. I’m also the co-founder of Inneractive that Fyber acquired, that I think was one of the very first companies to do in-game advertising on Nokia devices. That pretty much gives away my age.
Esther: [chuckles] It’s crazy to think that mobile advertising has been around for 15 years.
Offer: Unbelievable.
Esther: Yes, hard to hear out loud. You’ve been in since the beginning. I think one of the questions I have to ask, obviously, a lot has changed, because I’d imagine you’re not overly focused on Nokia devices today. [chuckles]
Offer: Completely.
Esther: What do you think is the biggest shift between how you’re looking at in-app mobile advertising today versus what you were doing 15 years ago?
Offer: Oh, wow. I think one of the biggest changes, and if we look at the context of 10 years for a second, 10 years ago, I remember visiting Mobile World Congress in Barcelona, and people asked, “When will brands, like the [unintelligible 00:02:18] advertising will jump and come to mobile?” Every year, it was like, “This is the year of mobile advertising.” Now it’s the year where we see brands finally coming to mobile.
We still have this vision and this dream of brands coming to mobile. Every year, it’s getting better and better. Nevertheless, what we see right now is most advertisers, most marketers are very much ROAS-driven. We all ask ourselves, “How can we get the best results?” It’s getting even more lighted up now with iOS 14.5. Depending when this will be, [unintelligible 00:02:54] before or after. [chuckles]
Esther: I think it’s coming on Monday, so it’s probably going to be after. We’re coming from the past. Yes, it’s about to drop.
Offer: Now that we have iOS 14.5, and IDFA went away, brings up more and more questions, how do you balance privacy, and performance, and measurability? Because that’s why we are here. Those questions are [unintelligible 00:03:19] for me.
Esther: Let’s go straight to it, because we’re talking about a shift that– At least, since I haven’t been in for 15 years, but in, let’s say, eight years since I’ve been in this space, I think the idea of deprecation is probably the bigger change that I can think of that’s happened lately, and a huge revolution. Making the move from user-level ROAS, real-time data to understanding what to do. How do you even begin to tackle the idea of creating value and structuring efforts in a way where you’re making sure that you’re not losing retention?
Because I think with mobile, the big challenge is, you don’t have that immediacy, the download usually is free, the first login to the app isn’t necessarily the conversion that you’re optimizing for. How are you even structuring that mental side of, “Okay, we’re about to get in a world where we don’t have this user-level data, how do we look at ad structures? How do we look at success in our efforts?”
Offer: Yes. First, I start by reminding myself that user acquisition is not going away. The notion of, “Oh, so maybe I should shift my budgets from iOS to Android,” those were some of the preliminary thoughts when this was announced. User acquisition is key for every business. With that in mind, then you’re asking, so what do we do? What are the alternatives? We will have some user-level targeting for [unintelligible 00:04:55] traffic. I personally think that those numbers would be very, very low, but very soon we’ll see exactly where they land.
For some traffic, where you do get concerned, and you do get the IDFA, you can use some of the data and maybe extrapolate from it. For the vast majority of traffic without IDFA, I think we need to start employing a more of a contextual view to targeting. Contextual doesn’t mean what you remember from 2014 desktop contextual targeting that, “Oh, this banner is alongside an article about whatever.” No, it’s about different game genres and being very specific about the genre of the game.
I see more and more marketers cataloging the games and the apps in a very [unintelligible 00:05:48] way. Marketers cataloging games and apps in a very [unintelligible 00:05:54] way that allows you to better target against other games or other apps. For instance, if I look at hyper-casual, which is one of those area when you say, “Oh, what will happen to hyper-casual?” I think that you have a very high level of affiliation between, “Oh, this is a Tap Tap Jump game, so I can find similar games to attach to it.” It’s less about the user. It’s more about the content. That’s one point.
Esther: I want to make sure that I’m following this, it’s basically saying, “Okay, we’re losing the way we’re used to targeting users, which is to ensure that we’re creating [unintelligible 00:06:35] groups or are using whatever data we have for that,” and it’s more saying, “Okay, we’re assuming that user base X is located here and this is the area that we’re trying to conquer. Let’s make sure that our app becomes visible here, where this user base is hanging out.”
Offer: Yes, absolutely. Look, I separate between targeting and attribution, because once IDs are going away, those are the two areas; how do I target and then how do I attribute what happened in the game? It’s almost like a separate discussion. There are so many new tools and new ways to look at attribution, different MMPs are pushing for it. I really like what AlgoLift is doing as part of Vungle now trying to look at, “How do I predict that?” That’s one thing.
On the targeting side, again, you have the context of the game or the app, and then you have the context of the user. Again, to be completely privacy safe and without storing any data, I can still give the marketer clues about the impression. For instance, what is the average click-through rate of the user in this game? Nothing ID, nothing identifiable here, but this user tend to click and engage with rewarded video. This user typically watches rewarded videos to the end or not.
There is a lot of information that can inform the advertiser in real-time, about, “Okay, so is it a [unintelligible 00:08:12] campaign or maybe I should look for creatives that are more sensitive to clicks?” For instance, out of the contextual signals that we share are, “What are the last ads, the last bundles or games the user saw and engaged with? This user is engaging with match 3 games. I can use this type of information.”
Esther: It’s essentially shifting the view of– Instead of looking at things like in-app purchases and specific engagements that you lose that ability to share, you’re looking at contextual engagement, “This is how you interact with an ad, this is what happens when it comes in place, this is what we’re looking at,” and then you’re creating different signals, certainly, but you’re still saying, basically, “This is my proxy for understanding value,” the way somebody engages if somebody is watching till the end and skipping back through a video. “We’ve hit something, we’ve gotten some level of engagement. We have to assume that that’s going to lead to the yield that we’re looking for afterwards.”
Offer: Yes. It’s going to be interesting for publishers with a healthy in-app purchase business and activity. Some of them may start saying, “You know what? I as the publisher, I know if this user right now is a top spender or not.” The question is, do I want to share this information with the advertiser? Because why would I let this usually go away? You can put a price tag and you can say, “You know what? I can use my first-party data and segment spenders from non-spenders in such a way,” and say, “You want access to my top spenders, that’s fine, but you will have to pay a premium versus my non-spenders.” This is a way to use first-party data in the way you monetize your audience.
Esther: I need more info here, because these sound super interesting and I’m not sure that I followed all the way through.
Offer: Again, it’s not a matter of how the publisher is looking at the first-party data in the audience space. I think the evolution of how publishers monetize in iOS 14 and above and how mediation platforms and different tools will come in place is to allow you to segment the audience, based on your first-party data, without sharing IDs, but still pricing them differently.
Again, as the publisher, I can tell, “Oh, Esther just came in and she’s a top spender. I want to show no ads to Esther whatsoever to protect her,” or “I want to let only my non-competitive advertisers be in front of Esther,” and they will have to pay a premium because you are a top spender. All of those decisions will be on the publisher’s side.
Esther: That’s super interesting. I like that. Basically, what you’re saying is, I can’t control my targeting necessarily, but I can control who my users are exposed to and I know who my whales are, I know who my casual players are. I basically decide, what am I willing to share at what price and play? Interesting. That’s actually, I think, the first– I’ve been talking about IDFA a lot and I’m pretty sure that’s the first time I’m hitting this and it’s super, super interesting.
Offer: Thank you.
Esther: I’m wondering how that works, if we look at the developers who are multi-titled developers and have their cross-promotional network. I have to say, I expect more and more companies to– We talked about where hyper-casual is going. I expect a lot of people to acquire hyper-casual studios just to get access to a mass audience group and put them in their networks, sooner or later. How would that strategy work out in terms of your own internal advertising? How do you look at your cannibalization versus your ability to grow within the network of people you’ve already acquired?
Offer: Look, you’re right. We’ve seen it. We’ve seen companies out there acquiring early, because as long it’s the same company, you can share IDs. Of course, [unintelligible 00:12:31] can actually work very well. I definitely see how publishers are now managing monetization, cost promotion, in-app purchase, special offers, and deciding exactly [unintelligible 00:12:46] users what to do.
The only point I would highlight is that, where it gets tricky, this data is worth gold. How exactly do I as the publisher share this information with the platform and what will the platform do with the data? If this person also has other businesses that may compete with my own business, this is where it may get a bit tricky. I’m guessing that’s where we see more publishers looking at either owning mediation, building mediation, or working with mediation platforms in such a way that creates a world garden around the data. In the next year or so, I believe we’ll see more developments there as well.
Esther: Agreed. Okay. Anyway, we were talking about preloaded apps. That’s what I was about to ask you about before we were so rudely interrupted. I think it’s a really interesting avenue to understand. We talked so much about acquisition and getting users to download, what happens when the app just exits from the second you power up?
Offer: This is a fascinating area. I’m still learning myself. Preloads are so different than just buying on the open web. The experience difference. When you buy a new device and you put it up and you set it up, you have a wizard where you can basically choose personalized content that will be ready for your device from the get-go, but from this point onward, the discovery, playing for the first time, that’s something we need to definitely educate yourself, but it’s a massive opportunity, especially without IDs.
Esther: Exactly, that’s what I’m thinking. If we don’t need to create that audience group, it becomes a really interesting opportunity. Is it possible yet in iOS devices or is it only Google Play?
Offer: It’s only Android devices at the moment. Here in the US at least, iOS is a very big piece, and we talk a lot about iOS, especially now with IDFA going away, but Android is a massive part of the market. We are talking about, I think, more than 60 million new devices every quarter or so. It’s a big opportunity.
Esther: Then, I guess, the question becomes, how do you get somebody– First of all, obviously, you have to be somewhat selective in a preload. If we start and you open your phone and there’s 600 apps, you’re not going to get the power that you want. We’ve got to be selective there. How do we look at getting users to notice that content? Sometimes you’ll have it on your home screen and you still need that context to push them in where you’re looking to get them. Any insights on that and how that works?
Offer: Very two good points, Esther. One, a very limited supply of the device. We don’t want to overflow the user with options, also, the experience has to be smooth. They’re just a handful of apps available in the device when you put it up. Typically, the recommendation is for those evergreen games and apps. Think about the game that is more likely to be played for a long period of time. If it’s a solitaire game or a branded match 3 game, probably more likely than a short-lived hyper-casual type.
Then you asked about how we make sure that people notice? That’s the interesting part, where I still learn everyday about Digital Turbine. There are a suite of tools to bring that game into attention. If those are notifications on the device, if it’s about the remarketing and retargeting, so you play some other game and the ad will tell you, “Hey, did you know you have this game on the device?” and deep linking to automatically open it up. There are a suite of tools that together create very high engagement levels, and from what I’ve seen so far, much, much higher conversion rates, then open web.
Esther: Makes sense. I’m wondering also, is there a way to use demographics there? If I’m imagining solitaire, for example, my mom, when she opens her device for the first time, she’s gonna play solitaire. She’s used to having solitaire on her digital devices. My younger sister, maybe not so much, maybe there’s a better app [crosstalk] for her to see. Yes, [chuckles] but she’s married. She was a very young bride. I guess, I don’t know if it’s based on a login or based maybe on the newness of the device, is there any way to customize and tailor and say, “I don’t want to be on every device, I want to hit this demographic”?
Offer: I’ll answer with the question, because I don’t know yet enough, but what we are seeing now with iOS, and look, I think Android will follow suit again, with Android ID is the balance of power between the carriers OEMs and the OS. We’ll have to see how much data will become available from within the platform or from the carrier and then what we as an industry can use again in the privacy, where we respect privacy, we respect opt-in and opt-outs. Again, that’s something that we’ll see in the next couple of years. This is not an overnight change, but targeting right now, again, you can target just like you do on any other platform.
Esther: Okay. That makes a lot of sense. I think it would be unrealistic to not say flexible, because I totally agree, Apple and Google, they’re not too far behind each other. One makes one change, the other one’s coming up within the next year or so to make sure that they’re on the ball as well.
Offer: Esther, one point to remember, we talk about preloading, what’s really unique is the scale, because for the right advertiser, you can basically put yourself on literally millions of devices very, very fast, like you are there, you are on the device. It’s very hard to replicate that scale when you buy open web, when you buy a network. You can, it’s a matter of money, but here, the math works very well. You can just preload on millions of devices. You can tell your devices that you want, the regions that you want, and you have a huge audience. That’s quite remarkable.
Esther: Crazy scale. I guess, the question is, your success metric has to change. What’s the KPI you’re looking at to understand if you’ve had the right success in a native app implementation?
Offer: Well, I would say, it does change from advertiser to advertiser. We do recommend to have a slightly wider attribution window. Just when you buy open web and there is a view-through of a day or not, here, from the moment the device was shipped and booted for the first time, we should expect more than seven days until the user may open the game or the app for the first time. That’s the only recommendation, but from this point, the KPIs are the same. The KPIs, whatever, you have [unintelligible 00:20:22] target or CPA-
Esther: Retention, engagement. Yes, everything comes down to.
Offer: Yes, it’s the same audience.
Esther: I’m going to make you be a fortune-teller for a minute. We think back this time one year from now, what do you think performance marketing looks like for mobile advertisers?
Offer: I think in a year from now, we will laugh about the third quarter of 2021, how people freaked out after the launch and how it did recover in Q4, and how Q1 is much stronger. I do see this [unintelligible 00:21:06] from starting now when iOS went away, until we recover. We’ll probably also talk about all the new channels that performance marketers adapted. If it’s influencers, if it’s reloading, just the media mix will change.
Esther: Totally agree, because the second you’ve had, I think, performance marketing, the networks that have been successful until now have benefited off that immediacy, that real tangible ROI. Now we give such an interesting case to the alternative channels that you felt like measurement wasn’t as immediate or wasn’t as solid as before.
Many developers, I think, have missed massive markets because they’ve been so reluctant to lose that granular piece of data, and now you’re giving that same chance, you can discover all of a sudden, there’s networks that are performing much better for you, you just have to be able to analyze it in a more aggregative way, instead of incrementality, instead of that immediacy and that user by user feedback you were looking for before.
Offer: Absolutely. After that point, I think iOS will be a big part of it. You see also part of iOS 14.5 more placements in the search page and the recommendation page, you can trade. That will be a big part as well.
Esther: Yes. I think one of the things that we’re also– We’ve always seen the connection between paid and organic on the ASO side, the more you spend, the more your rankings tend to benefit and the more you’re improving. I think now when we’re looking at– You tend to acquire much more broadly or acquire a different set of users, the connection between those two pathways becomes even more important.
Understanding that pure impact, not just what you’ve been looking at until now, which is, “I’m in my Facebook dashboard and I’m looking at my ROAS per ad,” you’ve got to start to force yourself into true performance, holistic performance behavior, which is awesome, because now you’re getting, in a way, more accurate, even if you’re losing that level of granularity.
Offer: Absolutely. We talked before about measurements and attributions. To your point, how do we measure? I can see how we are moving into more and more incrementality testing because we won’t be able to point to a specific user, but I just opened a new channel, just [unintelligible 00:23:37], so what does it mean for me? How can I see the impact on my business throughout incrementality a year from now? Also a big part.
Esther: Now you’re committed to coming back a year from now, so we can assess how accurate you went. Amazing. Before we move on, we have a quickfire round, where you’re going to have to answer questions very quickly, but before we move on to there, is there anything else that you think is really important to be said as we’re at the cusp of this massive change that’s coming into our industry?
Offer: Yes. I’m also intrigued about global expansion and how different markets operate. We talk a lot about North America back, LATAM and APAC, those markets in 2020 go like crazy. I also think user acquisition stoppages, obviously, must be localized. Again, a year from now, I would love to see also how we adapt and are successful also in those markets, both on the monetization side to fuel the business and, of course, also acquisition. That’s very interesting.
Esther: Definitely agreed. Are you ready for quickfire?
Offer: No, but let’s do it.
Esther: It’s not so bad, I promise. If you could give one tip to an aspiring growth marketer, what would it be?
Offer: Become good friends with the CEO and make sure you have some side budgets for a rainy day.
Esther: [laughs] I like it. Your favorite resource on mobile growth?
Offer: MDM, Mobile Dev Memo. Love it.
Esther: The best. In California, you’re catching up on your vaccine rollout, so you’re about to be able to go out to lunch again and meet people. Who in the industry is the person you’d most want to take to lunch and why?
Offer: I have so many people I want to meet. I want to meet my team. I haven’t seen my team for such a long time.
Esther: True.
Offer: Giving out one name, this is too hard, pass.
Esther: I want to create a little drama in your day today, who can we piss off the most that you didn’t put them first?
Offer: You know what? I’ll have a drink with Adam [unintelligible 00:25:55] congrats him on taking up [unintelligible 00:25:57] Public. Definitely been an amazing journey. These guy started small, raised just a bit of money, and against all odds, built a remarkable company. [crosstalk] drink and dinner. I just miss the guy.
Esther: Yes. I feel like he should maybe treat you because he’s had a good quarter, but yes, agreed. [laughs] The most important question. Are you ready? What is your favorite type of pancake?
Offer: [laughs] I am not a pancake guy.
Esther: No.
Offer: Yes, I know. Look, my wife and kids, we always have a debate, but no. I will make a Nutella banana pancake if I become a pancake guy, for sure.
Esther: It’s correct. Objectively, it’s correct. I can’t accept that you’re not a pancake guy, but I can’t accept your answer is good. Offer, where can people find you if they want to hear more, see what you’re up to, stock your professional life?
Offer: Hit me up on LinkedIn. I have a very catchy name, Offer Yehudai. Good luck with that. [laughs] [unintelligible 00:27:09] Fyber and Offer. [laughs] If you’re here in the Bay Area, I’m vaccinated, would love to get back to normality and meet. Esther, I hope to see you in Tel Aviv soon as well.
Esther: I was going to say, I’m not going to be flying for a little while, but I very much expect you to come to meet us for a drink when you’re here, hopefully, very soon.
Offer: I want to thank you for having me and for the little one to accommodate us [unintelligible 00:27:35]. It was fun.
Esther: Nava, you want to say, “You’re welcome”? She does not, but I want to thank you as well. You stole the words out of my mouth. It was a lot of fun and I’m looking forward to next year when we talk about our space and how silly it was that we all freaked out so very much. Amazing. [laughs]
Offer: It’s a date.
