Joseph Kim (CEO of LILA Games) talks about the real-world data and the implications of IDFA deprecation with Brian Bowman (CEO of Consumer Acquisition - Creative Studio and UA services), Matej Lancaric (Independent UA & Marketing Consultant) and Warren Woodward (Co-founder and Chief Growth Officer at Upptic - Growth Services).
Consumer Acquisition is seeing iOS revenue down 15-20% across their portfolio of companies, with different impacts for subgenres: the most impacted are games targeting narrow audiences (hunting whales) and apps monetizing with in-app ads.
In their earnings, Facebook is reporting that impressions are up 6% and CPMs are up 47%. This means that click-to-install quality is down.
In their earnings report, Apple is projecting their services revenue will be down for the second half of the year. They’ve started realizing what the data means in terms of the impact on their own business.
There was lots of talk about how to best operate in the new environment (structure, conversion scheme, etc.), but the reality is that a lot of advertisers have just paused or decreased spend.
AppAnnie came out with research saying social casino and strategy games are being hit harder. Customer Acquisition also believes hardcore and hypercasual games are also heavily impacted (hypercasual because they rely on low CPMs).
If people want to play a game and there are no ads, they will look for a game they can remember. So games that have more presence of mind (i.e. a stronger brand) therefore have an advantage.
Some brands have a tendency to buy on a blended model, combining organic + paid. Smaller studios tend to focus on direct UA performance, so their KPIs will look worse because a large percentage of downloads are not tracked as well.
The spend has been shifted to Android and right now iOS looks like the whole UA space last year when Covid started: very low competition. It’s probably just a phase. This would explain why Supercell, Clash Royale, etc. have seen spikes.
With the loss of IDFA, the performance of lookalike audiences is half what it was before May 26 (based on a large spectrum of clients). With a limited number of people opting-in to be tracked, the custom audiences are very small and lookalike audiences don’t last long. This is true for Facebook (AEO, VO) as well as other ad networks with algorithms using app events. [Brian, Consumer Acquisition]
For Consumer Acquisition’s portfolio, Google is less negatively impacted than other platforms from both a ROAS and a spend perspective. One reason might be that Google was already probabilistic in nature.
Both ironSource and Unity have been using the term “contextual targeting” and feel like they’re well positioned to take advantage of that. Both are bullish on the stability of their network, and where they’re headed. Unity captures and analyzes 50 billion in-app events per day.
Upptic sees that on iOS, for certain apps, SDK Networks have started to represent a bigger proportion of the profitable spend than FB/Google. However a possible reason is that major networks are still relying on probabilistic attribution through MMPs, and this might not be a short-sighted approach.
Consumer Acquisition is seeing a 15-20% ROAS loss overall, but with some categories much more impacted (e.g. apps whale hunting vs. match-3 games).
If you’re a network and you’ve lost your ability to deliver high-value users, you’re going to put more lower-quality traffic into the mix.
Good UA has a halo effect, especially for large titles. So when Upptic is still working off old prediction models for an app and UA is not as good, they’re also seeing organics trend down.
Any team is going to have to get into the game of prediction, whether they’re comfortable with it or not. If you’re not yet tracking opt-in and opt-out events yet, start working with your product team because you will need that data in the future.
Decreasing spend on iOS can help you better identify the campaign performance by looking at blended ROAS. You can see the baseline of organics drop as you decrease spend from a source.
Things with the best performance right now: top countries, localized ad copy with AEO (ideally with broad targeting) and interest groups (formulating contextual mapping).
On paid social, if you don’t get your marketing and product teams working together to come up with a new way of working (including to get ahead of creative fatigue), you’re going to be a dinosaur.
A persona-led approach throughout the funnel (interest groups, ad creative, onboarding flow) is hard but seems like a sustainable long-term strategy.
You gotta treat your ASO as you treat your ads portfolio: testing, innovating, refreshing. It will also benefit your organic conversions too.
Warren from Upptic thinks that SKAN is not the long-term solution for a viable UA. Either it has to evolve or the industry has to build a superior approach.
Creative testing on iOS is dead. Don’t waste your time. Creative testing is all moved to Android.
For creative testing, Consumer Acquisition uses Singapore and India on Android as a proxy for the US. It works really well (no issue with cultural sensitivity) and it’s a tenth of the cost.
If nothing material changes, then media mix modeling and targeting clusters of users will be what we have. It won’t target the individual and will be substantially less efficient. It’s a very likely path.
In iOS 15, there will be cross-network attribution and it will normalize the way traffic looks across all platforms. It will put the networks on parity.
What will be very important moving forward: onboarding flows, early monetization signals and a deep understanding of your personas (who you’re building for, what do they read/watch/listen to and their motivations in the external world). This will allow you to cluster creative and interest. Example of campaigns: targeting people that like to kill each other vs. people that like to relax vs. people that like to complete levels.
Winners: large companies with large first-party data reservoirs, midcore games with broad monetization profiting from other games not able to target whales, IP games. Losers: companies spending less than a million a month on advertising, companies primarily relying on ad monetization, companies that don’t have the cash flow to roll-out an 18-month LTV window and sustain it.
Consumer Acquisition is seeing iOS revenue down 15-20% across their portfolio of companies, with different impacts for subgenres: the most impacted are games targeting narrow audiences (hunting whales) and apps monetizing with in-app ads.
In their earnings, Facebook is reporting that impressions are up 6% and CPMs are up 47%. This means that click-to-install quality is down.
In their earnings report, Apple is projecting their services revenue will be down for the second half of the year. They’ve started realizing what the data means in terms of the impact on their own business.
There was lots of talk about how to best operate in the new environment (structure, conversion scheme, etc.), but the reality is that a lot of advertisers have just paused or decreased spend.
AppAnnie came out with research saying social casino and strategy games are being hit harder. Customer Acquisition also believes hardcore and hypercasual games are also heavily impacted (hypercasual because they rely on low CPMs).
If people want to play a game and there are no ads, they will look for a game they can remember. So games that have more presence of mind (i.e. a stronger brand) therefore have an advantage.
Some brands have a tendency to buy on a blended model, combining organic + paid. Smaller studios tend to focus on direct UA performance, so their KPIs will look worse because a large percentage of downloads are not tracked as well.
The spend has been shifted to Android and right now iOS looks like the whole UA space last year when Covid started: very low competition. It’s probably just a phase. This would explain why Supercell, Clash Royale, etc. have seen spikes.
With the loss of IDFA, the performance of lookalike audiences is half what it was before May 26 (based on a large spectrum of clients). With a limited number of people opting-in to be tracked, the custom audiences are very small and lookalike audiences don’t last long. This is true for Facebook (AEO, VO) as well as other ad networks with algorithms using app events. [Brian, Consumer Acquisition]
For Consumer Acquisition’s portfolio, Google is less negatively impacted than other platforms from both a ROAS and a spend perspective. One reason might be that Google was already probabilistic in nature.
Both ironSource and Unity have been using the term “contextual targeting” and feel like they’re well positioned to take advantage of that. Both are bullish on the stability of their network, and where they’re headed. Unity captures and analyzes 50 billion in-app events per day.
Upptic sees that on iOS, for certain apps, SDK Networks have started to represent a bigger proportion of the profitable spend than FB/Google. However a possible reason is that major networks are still relying on probabilistic attribution through MMPs, and this might not be a short-sighted approach.
Consumer Acquisition is seeing a 15-20% ROAS loss overall, but with some categories much more impacted (e.g. apps whale hunting vs. match-3 games).
If you’re a network and you’ve lost your ability to deliver high-value users, you’re going to put more lower-quality traffic into the mix.
Good UA has a halo effect, especially for large titles. So when Upptic is still working off old prediction models for an app and UA is not as good, they’re also seeing organics trend down.
Any team is going to have to get into the game of prediction, whether they’re comfortable with it or not. If you’re not yet tracking opt-in and opt-out events yet, start working with your product team because you will need that data in the future.
Decreasing spend on iOS can help you better identify the campaign performance by looking at blended ROAS. You can see the baseline of organics drop as you decrease spend from a source.
Things with the best performance right now: top countries, localized ad copy with AEO (ideally with broad targeting) and interest groups (formulating contextual mapping).
On paid social, if you don’t get your marketing and product teams working together to come up with a new way of working (including to get ahead of creative fatigue), you’re going to be a dinosaur.
A persona-led approach throughout the funnel (interest groups, ad creative, onboarding flow) is hard but seems like a sustainable long-term strategy.
You gotta treat your ASO as you treat your ads portfolio: testing, innovating, refreshing. It will also benefit your organic conversions too.
Warren from Upptic thinks that SKAN is not the long-term solution for a viable UA. Either it has to evolve or the industry has to build a superior approach.
Creative testing on iOS is dead. Don’t waste your time. Creative testing is all moved to Android.
For creative testing, Consumer Acquisition uses Singapore and India on Android as a proxy for the US. It works really well (no issue with cultural sensitivity) and it’s a tenth of the cost.
If nothing material changes, then media mix modeling and targeting clusters of users will be what we have. It won’t target the individual and will be substantially less efficient. It’s a very likely path.
In iOS 15, there will be cross-network attribution and it will normalize the way traffic looks across all platforms. It will put the networks on parity.
What will be very important moving forward: onboarding flows, early monetization signals and a deep understanding of your personas (who you’re building for, what do they read/watch/listen to and their motivations in the external world). This will allow you to cluster creative and interest. Example of campaigns: targeting people that like to kill each other vs. people that like to relax vs. people that like to complete levels.
Winners: large companies with large first-party data reservoirs, midcore games with broad monetization profiting from other games not able to target whales, IP games. Losers: companies spending less than a million a month on advertising, companies primarily relying on ad monetization, companies that don’t have the cash flow to roll-out an 18-month LTV window and sustain it.
Consumer Acquisition is seeing iOS revenue down 15-20% across their portfolio of companies, with different impacts for subgenres: the most impacted are games targeting narrow audiences (hunting whales) and apps monetizing with in-app ads.
In their earnings, Facebook is reporting that impressions are up 6% and CPMs are up 47%. This means that click-to-install quality is down.
In their earnings report, Apple is projecting their services revenue will be down for the second half of the year. They’ve started realizing what the data means in terms of the impact on their own business.
There was lots of talk about how to best operate in the new environment (structure, conversion scheme, etc.), but the reality is that a lot of advertisers have just paused or decreased spend.
AppAnnie came out with research saying social casino and strategy games are being hit harder. Customer Acquisition also believes hardcore and hypercasual games are also heavily impacted (hypercasual because they rely on low CPMs).
If people want to play a game and there are no ads, they will look for a game they can remember. So games that have more presence of mind (i.e. a stronger brand) therefore have an advantage.
Some brands have a tendency to buy on a blended model, combining organic + paid. Smaller studios tend to focus on direct UA performance, so their KPIs will look worse because a large percentage of downloads are not tracked as well.
The spend has been shifted to Android and right now iOS looks like the whole UA space last year when Covid started: very low competition. It’s probably just a phase. This would explain why Supercell, Clash Royale, etc. have seen spikes.
With the loss of IDFA, the performance of lookalike audiences is half what it was before May 26 (based on a large spectrum of clients). With a limited number of people opting-in to be tracked, the custom audiences are very small and lookalike audiences don’t last long. This is true for Facebook (AEO, VO) as well as other ad networks with algorithms using app events. [Brian, Consumer Acquisition]
For Consumer Acquisition’s portfolio, Google is less negatively impacted than other platforms from both a ROAS and a spend perspective. One reason might be that Google was already probabilistic in nature.
Both ironSource and Unity have been using the term “contextual targeting” and feel like they’re well positioned to take advantage of that. Both are bullish on the stability of their network, and where they’re headed. Unity captures and analyzes 50 billion in-app events per day.
Upptic sees that on iOS, for certain apps, SDK Networks have started to represent a bigger proportion of the profitable spend than FB/Google. However a possible reason is that major networks are still relying on probabilistic attribution through MMPs, and this might not be a short-sighted approach.
Consumer Acquisition is seeing a 15-20% ROAS loss overall, but with some categories much more impacted (e.g. apps whale hunting vs. match-3 games).
If you’re a network and you’ve lost your ability to deliver high-value users, you’re going to put more lower-quality traffic into the mix.
Good UA has a halo effect, especially for large titles. So when Upptic is still working off old prediction models for an app and UA is not as good, they’re also seeing organics trend down.
Any team is going to have to get into the game of prediction, whether they’re comfortable with it or not. If you’re not yet tracking opt-in and opt-out events yet, start working with your product team because you will need that data in the future.
Decreasing spend on iOS can help you better identify the campaign performance by looking at blended ROAS. You can see the baseline of organics drop as you decrease spend from a source.
Things with the best performance right now: top countries, localized ad copy with AEO (ideally with broad targeting) and interest groups (formulating contextual mapping).
On paid social, if you don’t get your marketing and product teams working together to come up with a new way of working (including to get ahead of creative fatigue), you’re going to be a dinosaur.
A persona-led approach throughout the funnel (interest groups, ad creative, onboarding flow) is hard but seems like a sustainable long-term strategy.
You gotta treat your ASO as you treat your ads portfolio: testing, innovating, refreshing. It will also benefit your organic conversions too.
Warren from Upptic thinks that SKAN is not the long-term solution for a viable UA. Either it has to evolve or the industry has to build a superior approach.
Creative testing on iOS is dead. Don’t waste your time. Creative testing is all moved to Android.
For creative testing, Consumer Acquisition uses Singapore and India on Android as a proxy for the US. It works really well (no issue with cultural sensitivity) and it’s a tenth of the cost.
If nothing material changes, then media mix modeling and targeting clusters of users will be what we have. It won’t target the individual and will be substantially less efficient. It’s a very likely path.
In iOS 15, there will be cross-network attribution and it will normalize the way traffic looks across all platforms. It will put the networks on parity.
What will be very important moving forward: onboarding flows, early monetization signals and a deep understanding of your personas (who you’re building for, what do they read/watch/listen to and their motivations in the external world). This will allow you to cluster creative and interest. Example of campaigns: targeting people that like to kill each other vs. people that like to relax vs. people that like to complete levels.
Winners: large companies with large first-party data reservoirs, midcore games with broad monetization profiting from other games not able to target whales, IP games. Losers: companies spending less than a million a month on advertising, companies primarily relying on ad monetization, companies that don’t have the cash flow to roll-out an 18-month LTV window and sustain it.
Notes for this resource are currently being transferred and will be available soon.
Brian
[💎@02:56] Consumer Acquisition is seeing iOS revenue down 15-20% across their portfolio of companies, with different impacts for subgenres: the most impacted are games targeting narrow audiences (hunting whales) and apps monetizing with in-app ads.
From public companies reporting earnings, 2 things that have impact:
1. Uncertainty of Covid-19 and the fact that people are going back to work (things are going back to normal)
2. Apple’s IDFA changes and the potential for a wide range of outcomes (positive and negative)
ironSource says the industry thought there would be impact in Q1 (didn’t happen), then in Q2 (happened to some extent) and is now monitoring for short-term effects.
Some stats from AppLovin: 80% of iOS devices are updated to iOS 14.5+, consent rates vary from 20-25% to 60% with an average of 35-40% opting in.
Some stats from Zynga: IDFA is having a material impact. They’ve warned revenue could go down, scaled back UA on iOS and are adjusting their LTV calculations to compensate.
[💎@05:25] In their earnings, Facebook is reporting that impressions are up 6% and CPMs are up 47%. This means that click-to-install quality is down.
Warren
[💎@05:53] In their earnings report, Apple is projecting their services revenue will be down for the second half of the year. They’ve started realizing what the data means in terms of the impact on their own business.
Matej
Some developers saw even bigger revenue drop on iOS. Budget has already been shifted to Android while they figure out which conversion scheme works.
Warren
[💎@07:23] There was lots of talk about how to best operate in the new environment (structure, conversion scheme, etc.), but the reality is that a lot of advertisers have just paused or decreased spend.
Brian
[💎@09:30] AppAnnie came out with research saying social casino and strategy games are being hit harder. Customer Acquisition also believes hardcore and hypercasual games are also heavily impacted (hypercasual because they rely on low CPMs).
Joseph
Looked at the portfolio of social casino games. There is an impact on downloads, and we can expect an impact on revenue coming up.
Brian
Things will continue getting worse at least until the end of August as more people adopt iOS 14.5 and we see the consequences.
Warren
[💎@12:57] If people want to play a game and there are no ads, they will look for a game they can remember. So games that have more presence of mind (i.e. a stronger brand) therefore have an advantage.
[💎@13:18] Some brands have a tendency to buy on a blended model, combining organic + paid. Smaller studios tend to focus on direct UA performance, so their KPIs will look worse because a large percentage of downloads are not tracked as well.
The model on how you can measure organic uplift itself is very important. Upptic had something built before, but now they’ve had to rebuild everything from scratch. They now need more data to test it.
Matej
[💎@15:04] The spend has been shifted to Android and right now iOS looks like the whole UA space last year when Covid started: very low competition. It’s probably just a phase. This would explain why Supercell, Clash Royale, etc. have seen spikes.
In the US for midcore games, Matej is seeing $4 CPI vs. $12-15 on Android!
Several companies are preparing iOS campaigns so that they can get the data to better optimize.
Brian
Mentioned Facebook’s article Privacy-Enhancing Technologies and Building for the Future.
CPMs are out of whack, which is a natural reaction because of the shift to Android.
Traffic quality is down, there is a loss in efficiency, so ad networks work on optimizing their algorithms.
Things are going to take time, it’s not going to change in the short term.
[💎@19:18] With the loss of IDFA, the performance of lookalike audiences is half what it was before May 26 (based on a large spectrum of clients). With a limited number of people opting-in to be tracked, the custom audiences are very small and lookalike audiences don’t last long. This is true for Facebook (AEO, VO) as well as other ad networks with algorithms using app events. [Brian, Consumer Acquisition]
Matej
Facebook performance is decreasing quite heavily and is getting worse.
Matej is not seeing the same thing with the lookalike audiences. It’s hard to build them and you need to refresh them more often, but on some games these lookalike audiences are working better. [Matej Lancaric]
Matej
Google UAC is performing better than Facebook on Android, but not on iOS.
Warren
Of the big players, Google has been the worst at adapting. They’ve always reported on Google-modeled installs but the modeling they use for UAC changed drastically after the SKAN changes. Numbers on the platform might look good, but don't match what is seen on MMPs on iOS.
Matej
Google was late with the SKAdNetwork on iOS. There are also 5-days delays.
Brian
TikTok seems to be doing better.
[💎@23:10] For Consumer Acquisition’s portfolio, Google is less negatively impacted than other platforms from both a ROAS and a spend perspective. One reason might be that Google was already probabilistic in nature.
Matej
[💎@23:37] Unity is a clear winner of the last couple of months.
Brian
Google’s earnings don’t even mention IDFA once.
[💎@24:08] Both ironSource and Unity have been using the term “contextual targeting” and feel like they’re well positioned to take advantage of that. Both are bullish on the stability of their network, and where they’re headed. Unity captures and analyzes 50 billion in-app events per day.
Matej
Last year, the consensus was that Facebook and Google were enough to manage half a million of spend. Now you need to go further.
Warren
[💎@25:29] Upptic sees that on iOS, for certain apps, SDK Networks have started to represent a bigger proportion of the profitable spend than FB/Google. However a possible reason is that major networks are still relying on probabilistic attribution through MMPs, and this might not be a short-sighted approach.
Brian
[💎@26:49] Consumer Acquisition is seeing a 15-20% ROAS loss overall, but with some categories much more impacted (e.g. apps whale hunting vs. match-3 games).
Some apps like match-3 games are doing ok, and the main challenge for them is really attribution and figuring out where their organics come from.
[💎@28:06] If you’re a network and you’ve lost your ability to deliver high-value users, you’re going to put more lower-quality traffic into the mix.
Warren
[💎@29:05] Good UA has a halo effect, especially for large titles. So when Upptic is still working off old prediction models for an app and UA is not as good, they’re also seeing organics trend down.
Warren
“We’ve had it easy the last several years in mobile UA”
[💎@30:55] Any team is going to have to get into the game of prediction, whether they’re comfortable with it or not. If you’re not yet tracking opt-in and opt-out events yet, start working with your product team because you will need that data in the future.
Matej
Working with predictions and using multipliers for each cohort (from the opt-ins) to get a sense of performance.
[💎@32:53] Decreasing spend on iOS can help you better identify the campaign performance by looking at blended ROAS. You can see the baseline of organics drop as you decrease spend from a source.
Warren
There is an opportunity to figure out your real baseline.
Brian
Not worried about modeling.
[💎@34:00] Things with the best performance right now: top countries, localized ad copy with AEO (ideally with broad targeting) and interest groups (formulating contextual mapping).
With UAC and AAA, FB/Google has been removing levers from media buyers in order to make UA more accessible. Consumer Acquisition believes the creatives are now the one real lever, and the rate of success is 15%. It’s really hard to stay ahead of creative fatigue.
If you’re not focused on creative fatigue and your marketing and product team are kept separate, advertising assets are often behind.
[💎@36:09] On paid social, if you don’t get your marketing and product teams working together to come up with a new way of working (including to get ahead of creative fatigue), you’re going to be a dinosaur.
[💎@36:17] A persona-led approach throughout the funnel (interest groups, ad creative, onboarding flow) is hard but seems like a sustainable long-term strategy.
Matej
Using SKAN data only to leverage it in the prediction models.
Brian
Always focus on the bottom of the funnel. But on Facebook you can have only 1 account, and usually the advertisers keep that one internal. So they’ve been focusing on Android and older operating systems on iOS.
Warren
The first thing they started was an ASO tool: the leverage you have in UA is decreasing, so they wanted to build a robust ASO tool to be able to rapidly test app store creatives. This is now paying off.
[💎@39:42] You gotta treat your ASO as you treat your ads portfolio: testing, innovating, refreshing. It will also benefit your organic conversions too.
Matej
Performance marketing is not dead: now there is more work than ever! More help is needed.
Joseph
SKAN probably advantages short-ARPU games: games where the ARPU curve flattens out relatively quickly.
Warren
[💎@42:10] Warren from Upptic thinks that SKAN is not the long-term solution for a viable UA. Either it has to evolve or the industry has to build a superior approach.
Brian
You have to completely change onboarding flows to be able to push users to monetize early on. But if it was easy it would have been done years ago.
Brian
[💎@43:30] Creative testing on iOS is dead. Don’t waste your time. Creative testing is all moved to Android.
“Creative testing on iOS is dead”
In general, you can’t test 2 ads that were introduced at different times because they all drag historical data.
[💎@44:13] For creative testing, Consumer Acquisition uses Singapore and India on Android as a proxy for the US. It works really well (no issue with cultural sensitivity) and it’s a tenth of the cost.
Brian
Reactivation campaigns were already limited before, and now they are dead. You can’t do it because device-ids are gone.
“Reactivation campaigns are dead”
Brian
We’re all in the same boat together. People that can invest now to put together pieces will have an unfair advantage.
Facebook has more data than anyone and builds up the walls of the walled garden + metaverse + html5 games. More content fortresses.
Warren
Might be a catalyst for FB to deprioritize their ad business and diversify the revenue streams.
Brian
[💎@49:35] If nothing material changes, then media mix modeling and targeting clusters of users will be what we have. It won’t target the individual and will be substantially less efficient. It’s a very likely path.
Warren
It’s hard to measure the contribution of each channel when you can’t measure them individually.
Brian
FB/Google always had an advantage in how they claim traffic, because there was no arbitror.
[💎@51:37] In iOS 15, there will be postbacks from SAN allowing cross-network attribution and it will normalize the way traffic looks across all platforms. It will put the networks on parity.
Matej
Matej is using this opportunity to run incrementality tests to get some sense of how channels well channels work.
Brian
CA has a lot of clients doing Facebook and paid search. Facebook is often treated as top of funnel, aided branding. When they turn off Facebook, paid search’s efficiency tanks.
Brian
[💎@53:40] What will be very important moving forward: onboarding flows, early monetization signals and a deep understanding of your personas (who you’re building for, what do they read/watch/listen to and their motivations in the external world). This will allow you to cluster creative and interest. Example of campaigns: targeting people that like to kill each other vs. people that like to relax vs. people that like to complete levels.
Warren
More meetings than ever with the product teams. Example: leads for financing loans are a very rare signal. You still need a lot of data to get a day-1 signal. It’s making marketing teams and product teams work together more closely.
“Now you MUST be friend with the product team”
Warren
Google is a winner. It’s also the hour to shine for all the smaller ad networks.
Brian & Others
[💎@57:30] Winners: large companies with large first-party data reservoirs, midcore games with broad monetization profiting from other games not able to target whales, IP games. Losers: companies spending less than a million a month on advertising, companies primarily relying on ad monetization, companies that don’t have the cash flow to roll-out an 18-month LTV window and sustain it.
Netflix is also a winner, with all the data they have and their game offering.
“Whatever you’re doing with creative testing today, 10x it.” - Brian
“Do your own research. Not everything is going to work for you.” - Matej
“Don’t be in a hurry to spend your money. Save your bullets.” - Warren
The revenue loss that’s currently seen in gaming is one thing, but you remove the ability to remarket/retarget ecommerce you really negatively impact black friday to christmas efficiency.
Brian
[💎@02:56] Consumer Acquisition is seeing iOS revenue down 15-20% across their portfolio of companies, with different impacts for subgenres: the most impacted are games targeting narrow audiences (hunting whales) and apps monetizing with in-app ads.
From public companies reporting earnings, 2 things that have impact:
1. Uncertainty of Covid-19 and the fact that people are going back to work (things are going back to normal)
2. Apple’s IDFA changes and the potential for a wide range of outcomes (positive and negative)
ironSource says the industry thought there would be impact in Q1 (didn’t happen), then in Q2 (happened to some extent) and is now monitoring for short-term effects.
Some stats from AppLovin: 80% of iOS devices are updated to iOS 14.5+, consent rates vary from 20-25% to 60% with an average of 35-40% opting in.
Some stats from Zynga: IDFA is having a material impact. They’ve warned revenue could go down, scaled back UA on iOS and are adjusting their LTV calculations to compensate.
[💎@05:25] In their earnings, Facebook is reporting that impressions are up 6% and CPMs are up 47%. This means that click-to-install quality is down.
Warren
[💎@05:53] In their earnings report, Apple is projecting their services revenue will be down for the second half of the year. They’ve started realizing what the data means in terms of the impact on their own business.
Matej
Some developers saw even bigger revenue drop on iOS. Budget has already been shifted to Android while they figure out which conversion scheme works.
Warren
[💎@07:23] There was lots of talk about how to best operate in the new environment (structure, conversion scheme, etc.), but the reality is that a lot of advertisers have just paused or decreased spend.
Brian
[💎@09:30] AppAnnie came out with research saying social casino and strategy games are being hit harder. Customer Acquisition also believes hardcore and hypercasual games are also heavily impacted (hypercasual because they rely on low CPMs).
Joseph
Looked at the portfolio of social casino games. There is an impact on downloads, and we can expect an impact on revenue coming up.
Brian
Things will continue getting worse at least until the end of August as more people adopt iOS 14.5 and we see the consequences.
Warren
[💎@12:57] If people want to play a game and there are no ads, they will look for a game they can remember. So games that have more presence of mind (i.e. a stronger brand) therefore have an advantage.
[💎@13:18] Some brands have a tendency to buy on a blended model, combining organic + paid. Smaller studios tend to focus on direct UA performance, so their KPIs will look worse because a large percentage of downloads are not tracked as well.
The model on how you can measure organic uplift itself is very important. Upptic had something built before, but now they’ve had to rebuild everything from scratch. They now need more data to test it.
Matej
[💎@15:04] The spend has been shifted to Android and right now iOS looks like the whole UA space last year when Covid started: very low competition. It’s probably just a phase. This would explain why Supercell, Clash Royale, etc. have seen spikes.
In the US for midcore games, Matej is seeing $4 CPI vs. $12-15 on Android!
Several companies are preparing iOS campaigns so that they can get the data to better optimize.
Brian
Mentioned Facebook’s article Privacy-Enhancing Technologies and Building for the Future.
CPMs are out of whack, which is a natural reaction because of the shift to Android.
Traffic quality is down, there is a loss in efficiency, so ad networks work on optimizing their algorithms.
Things are going to take time, it’s not going to change in the short term.
[💎@19:18] With the loss of IDFA, the performance of lookalike audiences is half what it was before May 26 (based on a large spectrum of clients). With a limited number of people opting-in to be tracked, the custom audiences are very small and lookalike audiences don’t last long. This is true for Facebook (AEO, VO) as well as other ad networks with algorithms using app events. [Brian, Consumer Acquisition]
Matej
Facebook performance is decreasing quite heavily and is getting worse.
Matej is not seeing the same thing with the lookalike audiences. It’s hard to build them and you need to refresh them more often, but on some games these lookalike audiences are working better. [Matej Lancaric]
Matej
Google UAC is performing better than Facebook on Android, but not on iOS.
Warren
Of the big players, Google has been the worst at adapting. They’ve always reported on Google-modeled installs but the modeling they use for UAC changed drastically after the SKAN changes. Numbers on the platform might look good, but don't match what is seen on MMPs on iOS.
Matej
Google was late with the SKAdNetwork on iOS. There are also 5-days delays.
Brian
TikTok seems to be doing better.
[💎@23:10] For Consumer Acquisition’s portfolio, Google is less negatively impacted than other platforms from both a ROAS and a spend perspective. One reason might be that Google was already probabilistic in nature.
Matej
[💎@23:37] Unity is a clear winner of the last couple of months.
Brian
Google’s earnings don’t even mention IDFA once.
[💎@24:08] Both ironSource and Unity have been using the term “contextual targeting” and feel like they’re well positioned to take advantage of that. Both are bullish on the stability of their network, and where they’re headed. Unity captures and analyzes 50 billion in-app events per day.
Matej
Last year, the consensus was that Facebook and Google were enough to manage half a million of spend. Now you need to go further.
Warren
[💎@25:29] Upptic sees that on iOS, for certain apps, SDK Networks have started to represent a bigger proportion of the profitable spend than FB/Google. However a possible reason is that major networks are still relying on probabilistic attribution through MMPs, and this might not be a short-sighted approach.
Brian
[💎@26:49] Consumer Acquisition is seeing a 15-20% ROAS loss overall, but with some categories much more impacted (e.g. apps whale hunting vs. match-3 games).
Some apps like match-3 games are doing ok, and the main challenge for them is really attribution and figuring out where their organics come from.
[💎@28:06] If you’re a network and you’ve lost your ability to deliver high-value users, you’re going to put more lower-quality traffic into the mix.
Warren
[💎@29:05] Good UA has a halo effect, especially for large titles. So when Upptic is still working off old prediction models for an app and UA is not as good, they’re also seeing organics trend down.
Warren
“We’ve had it easy the last several years in mobile UA”
[💎@30:55] Any team is going to have to get into the game of prediction, whether they’re comfortable with it or not. If you’re not yet tracking opt-in and opt-out events yet, start working with your product team because you will need that data in the future.
Matej
Working with predictions and using multipliers for each cohort (from the opt-ins) to get a sense of performance.
[💎@32:53] Decreasing spend on iOS can help you better identify the campaign performance by looking at blended ROAS. You can see the baseline of organics drop as you decrease spend from a source.
Warren
There is an opportunity to figure out your real baseline.
Brian
Not worried about modeling.
[💎@34:00] Things with the best performance right now: top countries, localized ad copy with AEO (ideally with broad targeting) and interest groups (formulating contextual mapping).
With UAC and AAA, FB/Google has been removing levers from media buyers in order to make UA more accessible. Consumer Acquisition believes the creatives are now the one real lever, and the rate of success is 15%. It’s really hard to stay ahead of creative fatigue.
If you’re not focused on creative fatigue and your marketing and product team are kept separate, advertising assets are often behind.
[💎@36:09] On paid social, if you don’t get your marketing and product teams working together to come up with a new way of working (including to get ahead of creative fatigue), you’re going to be a dinosaur.
[💎@36:17] A persona-led approach throughout the funnel (interest groups, ad creative, onboarding flow) is hard but seems like a sustainable long-term strategy.
Matej
Using SKAN data only to leverage it in the prediction models.
Brian
Always focus on the bottom of the funnel. But on Facebook you can have only 1 account, and usually the advertisers keep that one internal. So they’ve been focusing on Android and older operating systems on iOS.
Warren
The first thing they started was an ASO tool: the leverage you have in UA is decreasing, so they wanted to build a robust ASO tool to be able to rapidly test app store creatives. This is now paying off.
[💎@39:42] You gotta treat your ASO as you treat your ads portfolio: testing, innovating, refreshing. It will also benefit your organic conversions too.
Matej
Performance marketing is not dead: now there is more work than ever! More help is needed.
Joseph
SKAN probably advantages short-ARPU games: games where the ARPU curve flattens out relatively quickly.
Warren
[💎@42:10] Warren from Upptic thinks that SKAN is not the long-term solution for a viable UA. Either it has to evolve or the industry has to build a superior approach.
Brian
You have to completely change onboarding flows to be able to push users to monetize early on. But if it was easy it would have been done years ago.
Brian
[💎@43:30] Creative testing on iOS is dead. Don’t waste your time. Creative testing is all moved to Android.
“Creative testing on iOS is dead”
In general, you can’t test 2 ads that were introduced at different times because they all drag historical data.
[💎@44:13] For creative testing, Consumer Acquisition uses Singapore and India on Android as a proxy for the US. It works really well (no issue with cultural sensitivity) and it’s a tenth of the cost.
Brian
Reactivation campaigns were already limited before, and now they are dead. You can’t do it because device-ids are gone.
“Reactivation campaigns are dead”
Brian
We’re all in the same boat together. People that can invest now to put together pieces will have an unfair advantage.
Facebook has more data than anyone and builds up the walls of the walled garden + metaverse + html5 games. More content fortresses.
Warren
Might be a catalyst for FB to deprioritize their ad business and diversify the revenue streams.
Brian
[💎@49:35] If nothing material changes, then media mix modeling and targeting clusters of users will be what we have. It won’t target the individual and will be substantially less efficient. It’s a very likely path.
Warren
It’s hard to measure the contribution of each channel when you can’t measure them individually.
Brian
FB/Google always had an advantage in how they claim traffic, because there was no arbitror.
[💎@51:37] In iOS 15, there will be postbacks from SAN allowing cross-network attribution and it will normalize the way traffic looks across all platforms. It will put the networks on parity.
Matej
Matej is using this opportunity to run incrementality tests to get some sense of how channels well channels work.
Brian
CA has a lot of clients doing Facebook and paid search. Facebook is often treated as top of funnel, aided branding. When they turn off Facebook, paid search’s efficiency tanks.
Brian
[💎@53:40] What will be very important moving forward: onboarding flows, early monetization signals and a deep understanding of your personas (who you’re building for, what do they read/watch/listen to and their motivations in the external world). This will allow you to cluster creative and interest. Example of campaigns: targeting people that like to kill each other vs. people that like to relax vs. people that like to complete levels.
Warren
More meetings than ever with the product teams. Example: leads for financing loans are a very rare signal. You still need a lot of data to get a day-1 signal. It’s making marketing teams and product teams work together more closely.
“Now you MUST be friend with the product team”
Warren
Google is a winner. It’s also the hour to shine for all the smaller ad networks.
Brian & Others
[💎@57:30] Winners: large companies with large first-party data reservoirs, midcore games with broad monetization profiting from other games not able to target whales, IP games. Losers: companies spending less than a million a month on advertising, companies primarily relying on ad monetization, companies that don’t have the cash flow to roll-out an 18-month LTV window and sustain it.
Netflix is also a winner, with all the data they have and their game offering.
“Whatever you’re doing with creative testing today, 10x it.” - Brian
“Do your own research. Not everything is going to work for you.” - Matej
“Don’t be in a hurry to spend your money. Save your bullets.” - Warren
The revenue loss that’s currently seen in gaming is one thing, but you remove the ability to remarket/retarget ecommerce you really negatively impact black friday to christmas efficiency.
Brian
[💎@02:56] Consumer Acquisition is seeing iOS revenue down 15-20% across their portfolio of companies, with different impacts for subgenres: the most impacted are games targeting narrow audiences (hunting whales) and apps monetizing with in-app ads.
From public companies reporting earnings, 2 things that have impact:
1. Uncertainty of Covid-19 and the fact that people are going back to work (things are going back to normal)
2. Apple’s IDFA changes and the potential for a wide range of outcomes (positive and negative)
ironSource says the industry thought there would be impact in Q1 (didn’t happen), then in Q2 (happened to some extent) and is now monitoring for short-term effects.
Some stats from AppLovin: 80% of iOS devices are updated to iOS 14.5+, consent rates vary from 20-25% to 60% with an average of 35-40% opting in.
Some stats from Zynga: IDFA is having a material impact. They’ve warned revenue could go down, scaled back UA on iOS and are adjusting their LTV calculations to compensate.
[💎@05:25] In their earnings, Facebook is reporting that impressions are up 6% and CPMs are up 47%. This means that click-to-install quality is down.
Warren
[💎@05:53] In their earnings report, Apple is projecting their services revenue will be down for the second half of the year. They’ve started realizing what the data means in terms of the impact on their own business.
Matej
Some developers saw even bigger revenue drop on iOS. Budget has already been shifted to Android while they figure out which conversion scheme works.
Warren
[💎@07:23] There was lots of talk about how to best operate in the new environment (structure, conversion scheme, etc.), but the reality is that a lot of advertisers have just paused or decreased spend.
Brian
[💎@09:30] AppAnnie came out with research saying social casino and strategy games are being hit harder. Customer Acquisition also believes hardcore and hypercasual games are also heavily impacted (hypercasual because they rely on low CPMs).
Joseph
Looked at the portfolio of social casino games. There is an impact on downloads, and we can expect an impact on revenue coming up.
Brian
Things will continue getting worse at least until the end of August as more people adopt iOS 14.5 and we see the consequences.
Warren
[💎@12:57] If people want to play a game and there are no ads, they will look for a game they can remember. So games that have more presence of mind (i.e. a stronger brand) therefore have an advantage.
[💎@13:18] Some brands have a tendency to buy on a blended model, combining organic + paid. Smaller studios tend to focus on direct UA performance, so their KPIs will look worse because a large percentage of downloads are not tracked as well.
The model on how you can measure organic uplift itself is very important. Upptic had something built before, but now they’ve had to rebuild everything from scratch. They now need more data to test it.
Matej
[💎@15:04] The spend has been shifted to Android and right now iOS looks like the whole UA space last year when Covid started: very low competition. It’s probably just a phase. This would explain why Supercell, Clash Royale, etc. have seen spikes.
In the US for midcore games, Matej is seeing $4 CPI vs. $12-15 on Android!
Several companies are preparing iOS campaigns so that they can get the data to better optimize.
Brian
Mentioned Facebook’s article Privacy-Enhancing Technologies and Building for the Future.
CPMs are out of whack, which is a natural reaction because of the shift to Android.
Traffic quality is down, there is a loss in efficiency, so ad networks work on optimizing their algorithms.
Things are going to take time, it’s not going to change in the short term.
[💎@19:18] With the loss of IDFA, the performance of lookalike audiences is half what it was before May 26 (based on a large spectrum of clients). With a limited number of people opting-in to be tracked, the custom audiences are very small and lookalike audiences don’t last long. This is true for Facebook (AEO, VO) as well as other ad networks with algorithms using app events. [Brian, Consumer Acquisition]
Matej
Facebook performance is decreasing quite heavily and is getting worse.
Matej is not seeing the same thing with the lookalike audiences. It’s hard to build them and you need to refresh them more often, but on some games these lookalike audiences are working better. [Matej Lancaric]
Matej
Google UAC is performing better than Facebook on Android, but not on iOS.
Warren
Of the big players, Google has been the worst at adapting. They’ve always reported on Google-modeled installs but the modeling they use for UAC changed drastically after the SKAN changes. Numbers on the platform might look good, but don't match what is seen on MMPs on iOS.
Matej
Google was late with the SKAdNetwork on iOS. There are also 5-days delays.
Brian
TikTok seems to be doing better.
[💎@23:10] For Consumer Acquisition’s portfolio, Google is less negatively impacted than other platforms from both a ROAS and a spend perspective. One reason might be that Google was already probabilistic in nature.
Matej
[💎@23:37] Unity is a clear winner of the last couple of months.
Brian
Google’s earnings don’t even mention IDFA once.
[💎@24:08] Both ironSource and Unity have been using the term “contextual targeting” and feel like they’re well positioned to take advantage of that. Both are bullish on the stability of their network, and where they’re headed. Unity captures and analyzes 50 billion in-app events per day.
Matej
Last year, the consensus was that Facebook and Google were enough to manage half a million of spend. Now you need to go further.
Warren
[💎@25:29] Upptic sees that on iOS, for certain apps, SDK Networks have started to represent a bigger proportion of the profitable spend than FB/Google. However a possible reason is that major networks are still relying on probabilistic attribution through MMPs, and this might not be a short-sighted approach.
Brian
[💎@26:49] Consumer Acquisition is seeing a 15-20% ROAS loss overall, but with some categories much more impacted (e.g. apps whale hunting vs. match-3 games).
Some apps like match-3 games are doing ok, and the main challenge for them is really attribution and figuring out where their organics come from.
[💎@28:06] If you’re a network and you’ve lost your ability to deliver high-value users, you’re going to put more lower-quality traffic into the mix.
Warren
[💎@29:05] Good UA has a halo effect, especially for large titles. So when Upptic is still working off old prediction models for an app and UA is not as good, they’re also seeing organics trend down.
Warren
“We’ve had it easy the last several years in mobile UA”
[💎@30:55] Any team is going to have to get into the game of prediction, whether they’re comfortable with it or not. If you’re not yet tracking opt-in and opt-out events yet, start working with your product team because you will need that data in the future.
Matej
Working with predictions and using multipliers for each cohort (from the opt-ins) to get a sense of performance.
[💎@32:53] Decreasing spend on iOS can help you better identify the campaign performance by looking at blended ROAS. You can see the baseline of organics drop as you decrease spend from a source.
Warren
There is an opportunity to figure out your real baseline.
Brian
Not worried about modeling.
[💎@34:00] Things with the best performance right now: top countries, localized ad copy with AEO (ideally with broad targeting) and interest groups (formulating contextual mapping).
With UAC and AAA, FB/Google has been removing levers from media buyers in order to make UA more accessible. Consumer Acquisition believes the creatives are now the one real lever, and the rate of success is 15%. It’s really hard to stay ahead of creative fatigue.
If you’re not focused on creative fatigue and your marketing and product team are kept separate, advertising assets are often behind.
[💎@36:09] On paid social, if you don’t get your marketing and product teams working together to come up with a new way of working (including to get ahead of creative fatigue), you’re going to be a dinosaur.
[💎@36:17] A persona-led approach throughout the funnel (interest groups, ad creative, onboarding flow) is hard but seems like a sustainable long-term strategy.
Matej
Using SKAN data only to leverage it in the prediction models.
Brian
Always focus on the bottom of the funnel. But on Facebook you can have only 1 account, and usually the advertisers keep that one internal. So they’ve been focusing on Android and older operating systems on iOS.
Warren
The first thing they started was an ASO tool: the leverage you have in UA is decreasing, so they wanted to build a robust ASO tool to be able to rapidly test app store creatives. This is now paying off.
[💎@39:42] You gotta treat your ASO as you treat your ads portfolio: testing, innovating, refreshing. It will also benefit your organic conversions too.
Matej
Performance marketing is not dead: now there is more work than ever! More help is needed.
Joseph
SKAN probably advantages short-ARPU games: games where the ARPU curve flattens out relatively quickly.
Warren
[💎@42:10] Warren from Upptic thinks that SKAN is not the long-term solution for a viable UA. Either it has to evolve or the industry has to build a superior approach.
Brian
You have to completely change onboarding flows to be able to push users to monetize early on. But if it was easy it would have been done years ago.
Brian
[💎@43:30] Creative testing on iOS is dead. Don’t waste your time. Creative testing is all moved to Android.
“Creative testing on iOS is dead”
In general, you can’t test 2 ads that were introduced at different times because they all drag historical data.
[💎@44:13] For creative testing, Consumer Acquisition uses Singapore and India on Android as a proxy for the US. It works really well (no issue with cultural sensitivity) and it’s a tenth of the cost.
Brian
Reactivation campaigns were already limited before, and now they are dead. You can’t do it because device-ids are gone.
“Reactivation campaigns are dead”
Brian
We’re all in the same boat together. People that can invest now to put together pieces will have an unfair advantage.
Facebook has more data than anyone and builds up the walls of the walled garden + metaverse + html5 games. More content fortresses.
Warren
Might be a catalyst for FB to deprioritize their ad business and diversify the revenue streams.
Brian
[💎@49:35] If nothing material changes, then media mix modeling and targeting clusters of users will be what we have. It won’t target the individual and will be substantially less efficient. It’s a very likely path.
Warren
It’s hard to measure the contribution of each channel when you can’t measure them individually.
Brian
FB/Google always had an advantage in how they claim traffic, because there was no arbitror.
[💎@51:37] In iOS 15, there will be postbacks from SAN allowing cross-network attribution and it will normalize the way traffic looks across all platforms. It will put the networks on parity.
Matej
Matej is using this opportunity to run incrementality tests to get some sense of how channels well channels work.
Brian
CA has a lot of clients doing Facebook and paid search. Facebook is often treated as top of funnel, aided branding. When they turn off Facebook, paid search’s efficiency tanks.
Brian
[💎@53:40] What will be very important moving forward: onboarding flows, early monetization signals and a deep understanding of your personas (who you’re building for, what do they read/watch/listen to and their motivations in the external world). This will allow you to cluster creative and interest. Example of campaigns: targeting people that like to kill each other vs. people that like to relax vs. people that like to complete levels.
Warren
More meetings than ever with the product teams. Example: leads for financing loans are a very rare signal. You still need a lot of data to get a day-1 signal. It’s making marketing teams and product teams work together more closely.
“Now you MUST be friend with the product team”
Warren
Google is a winner. It’s also the hour to shine for all the smaller ad networks.
Brian & Others
[💎@57:30] Winners: large companies with large first-party data reservoirs, midcore games with broad monetization profiting from other games not able to target whales, IP games. Losers: companies spending less than a million a month on advertising, companies primarily relying on ad monetization, companies that don’t have the cash flow to roll-out an 18-month LTV window and sustain it.
Netflix is also a winner, with all the data they have and their game offering.
“Whatever you’re doing with creative testing today, 10x it.” - Brian
“Do your own research. Not everything is going to work for you.” - Matej
“Don’t be in a hurry to spend your money. Save your bullets.” - Warren
The revenue loss that’s currently seen in gaming is one thing, but you remove the ability to remarket/retarget ecommerce you really negatively impact black friday to christmas efficiency.
