Prepping for iOS 14 and the loss of the IDFA: What Nebo Radovic, Growth Lead at N3TWORK is doing
If growth is your world and your mission is acquiring new users, the September launch date of iOS 14 might seem like doomsday. And it’s getting closer and closer.
But is that really true?
To find out, I spent some time with Nebo Radovic, who leads growth at the N3TWORK Scale Platform. N3TWORK, of course, is the publisher of Legendary: Game of Heroes, a perennial top-grossing game. Perhaps even more importantly, N3TWORK is one of a very exclusive club of game publishers that have become so good at finding, acquiring, and retaining new players that they’ve built a platform to grow other developers’ games as well — investing $50 million to make it happen.
Long story short: Radovic knows his stuff.
We chatted about September and iOS 14, ranging from tech N3TWORK is building right now to code they’ll write after the IDFA dust settles. And from this is “the death of mobile growth” to “working off of aggregated data is not necessarily bad.” We also chat about whether Google will follow suit and deprecate the GAID, plus whether losing look-alike audiences is actually a big deal. And we also discuss what mobile marketers will need from the MMPs in the post-IDFA world.
Here’s the transcript, slightly edited for clarity:
John Koetsier: Let’s start here … how are you feeling right now about iOS 14 and IDFA?
Nebo Radovic: It’s been a month … I feel like I don’t know much more now than a month ago.
It’s probably best to prepare things that are obviously necessary, and wait for more complex projects when we know we need them. I did learn a few things about consent … which should be available.
We won’t be able to get full consent … but that’s not something we have right now either, and we’re still operating.
One possible scenario is that everything goes to mobile web. Developers move to the browser where they have more control, and you can do more. This is highly unlikely, but it is an option, as the app environment will be more restricted.
John Koetsier: You have a pretty unique thought about getting tracking permission in iOS 14 …
Nebo Radovic: Well, I was wondering … how can we get the most out of the current set-up?
We’ll probably get only 10% of people to give consent, but if we get the right 10%, maybe we don’t need more. I mean, by day 7 you lost 80-90% of users anyways. What you need to learn is where those 10% are coming from … if you could get consent from all the people who pay, then you’d be able to map where they come from and optimize towards those placements.
John Koetsier: What would tracking just paying players look like in your apps? How would you approach it with users?
Nebo Radovic: The example I like to use is Facebook Messenger location permissions.
For example, you come to San Francisco, and you want to meet a friend at the Ferry Building. So both of you want to share your locations. When you start, the app doesn’t have location permissions, but when you hit the Share location button, the app asks: do you want to give Messenger this permission?
So it’s opportunistic. Done at the right time, just like push notifications. You need to provide value in exchange for that permission.
You could also do what Twitter did: started to ask users permission for personalized ads. Twitter won’t let you see tweets chronologically before giving permission, so you get a better experience of the software in exchange for the permission. Developers could start testing this and learning when is the best point to ask.
For example, if you’re in idle games, maybe users get 2X more coins if they give tracking permission?
John Koetsier: So incentivizing the permission is OK?
Nebo Radovic: Well, hard incentivizing is not OK, just like you can’t incentivize people to rate your app, for example. But this is more like: to get extra value from the app or game, please give this permission. It makes the app experience better.
Maybe this won’t work, but it is one solution to try. And if you get really engaged users, you solve the problem of understanding where high-quality users are coming from.
John Koetsier: And data you’d get would be aggregated and non-personal …
Nebo Radovic: Yes, it would be aggregated: campaign and publisher data.
There’s an unspoken truth here: most developers work with aggregated data anyways.
You need more signals; you need a statistically significant amount of data. And you get to know who’s going to be a payer fairly quickly. Most payers convert in the first 7 days. That’s long enough for developers to understand who’s in the app.
At least, that’s my proposal. It’s not too intrusive, as long as Apple is OK with it.
John Koetsier: But what about finding more people like your best users via look-alike audiences?
Nebo Radovic: Well, look-alike audiences are becoming a thing of the past anyway. It’s actually not a big deal: the biggest ad network on the planet, Google, never really used them. And Facebook, the second biggest, is moving away from them.
They’re moving to App Acquisition Ads, and those don’t leverage look-alikes.
Look-alikes might not be as necessary as knowing where the best users are coming from in terms of geos, creatives, publishers, and campaigns. It’s not as big as a deal as not being able to attribute users to a specific campaign.
If look-alike audiences actually do become a thing, then many developers will move towards first-party logins, like Supercell ID, Machine Zone log-in, and so on. Tons of developers with multiple games that gamers like to play on multiple devices have them.
Ultimately, Google’s Game Center and Apple’s similar tech were not great ways to transfer games between devices. Big game devs have means of tracking users across platforms: email, phone numbers. So there are workarounds.
John Koetsier: What do you think about an IDFV strategy … acquiring other games and apps so a single publisher has a vast number of addressable users for promotion and cross-promotion?
Nebo Radovic: I predicted that people might go after hyper-casual games or build hub apps. The strategy is to acquire highly converting apps (conversion to install), drive users there cheaply, and then send those users to the better monetizing products.
This assumes, however, that there’s an overlap between hyper-casual and hard-core games. We will have to see if that works.
And If it does … Apple might go after those developers and ban IDFV for promoting apps.
What is possible is that you could use IDFV to target those users … it’s a pretty good strategy to retarget users. You could use an in-house DSP to do that, especially if you have multiple apps in the same category, like casino apps. In fact, it doesn’t have to be a gaming app: any app or a utility app could work as long as you have a valid IDFV.
John Koetsier: How much work will it take to get your UA and marketing teams set up for iOS 14?
Nebo Radovic: Internally, there are a few things we need to do, starting with the biggest: building the ability to work off aggregated data, not precise user-level data. All our tools need to work off that. Working off of aggregated data is not necessarily bad because, with more data, your machine learning algorithms learn faster and ultimately start performing better.
We also need a campaign mapping tool to map campaign IDs to geos, apps, etc. This is very important and will be something that each developer will need after the iOS14 launch.
But we’re also waiting to see what happens before we build something more advanced. We don’t know what exactly will be allowed, so it’s worth waiting, and thinking, and planning.
John Koetsier: Are you pretty comfortable with the guidelines from Apple?
Nebo Radovic: I’m not 100% sure, mainly because they haven’t covered many use cases in their documentation, which is why I think the best strategy is to wait and see before building any complex solutions.
Ultimately, most developers will try to come up with workarounds, and Apple will try to police everything and weed out the unethical strategies. But that’s why only tracking paying players and getting consent from them makes sense. Apple will be okay with that: we’d only be tracking 5-10% of users, with full consent, and those would be the users we care about the most.
John Koetsier: What are your thoughts about SKAN, Singular’s open proposal for common standards in using SKAdNetwork?
Nebo Radovic: It definitely makes sense, and I like how they selflessly shared it with the industry.
One of the ways to make this period as painless as possible is to work together. I think the only way MMPs can survive is by doing something like this. They’re building the probabilistic solutions, and that’s what the MMPs of the future will do.
What you need from your MMP is:
- Getting the right tech and data set-up
- Getting the right aggregation
- Serving as a postback solution for sharing that data with 3rd parties (FB, Google)
- Helping publishers optimize
- Helping publishers make algorithmic decisions on where to spend and not to spend
John Koetsier: What about Google? Do you think Google will make GAID opt-in as well?
Nebo Radovic: I actually don’t think Google will follow suit, or at least their implementation won’t be as strict as Apple’s. Google is, first and foremost, an ad network. Performance advertising is one of their core products, so I think they will be way more cautious about how they do this. If you look at how they approached getting rid of third-party cookies, that might give you a better idea of how Google will handle it. They will likely give developers a year or two to build all the necessary systems before they go live with the update.
Another reason why I don’t think Google will follow suit so quickly is the fact that Google doesn’t mind the hyper-casual developers:
They’re making money there, and they are one of the biggest ad monetization publishers …
… that’s why they might be a little less strict. It’s in their interest to keep the app ecosystem growing.
John Koetsier: And Facebook?
Nebo Radovic: Well, Facebook will likely want you to use their conversion lift tools. I wouldn’t be surprised if Google and Facebook start working together. That’s just a guess. They’re definitely the most affected; this could have a massive impact on both companies.
John Koetsier: Thank you for your time!