Post-IDFA Changes, Strategies and Trends

💎  x
16

Lina Danilchik (SplitMetrics/SearchAdsHQ) moderates a panel where Thomas Petit (Mobile Growth Consultant), Andrea Raggi (Performance Marketing Lead at Phiture - Mobile Growth Consultancy), Johannes von Cramon (Co-founder, Growfirst - Mobile Agency) and Ryan Goeden (VP of Sales at Kochava - MMP) discuss the biggest app marketing questions of 2020 and 2021.

Type:
Panel
Publication date:
November 30, 2020
Added to Growth Gems on:
December 1, 2020
💎 #
1

Start getting your legal team involved in the discussions because the context of this change is increased privacy. As you think about prompting for consent to share data, it's also a good time to think about consent management overall from a legislation perspective.

09:14
💎 #
2

Pilot some SKAdNetworks campaigns already: there is no reason to wait until the ATT framework is enforced before doing it. You should be able to see how things are going to look in your MMP.

10:03
💎 #
3

You actually don't need to require consent (depending on your app/model). This decision on requesting consent or not is the starting point before deciding on your data infrastructure, etc. 

12:25
💎 #
4

Already define which conversion values you are going to use. These conversion values are not going to be the same as the current postback you get from MMPs. You need to find the "magic events" you have that are indicators of good long term retention and monetization (run this analysis).

15:00
💎 #
5

If you do request consent, start testing different UX flows to find the best moment (e.g. at the beginning, at a happy moment, etc.).

17:25
💎 #
6

Also ask yourself how important iOS traffic is for you. Maybe you should switch budget to Android, to web traffic, etc.

19:13
💎 #
7

Do not underestimate the work needed to build your attribution with SKAdNetwork and setting up your conversion values. MMPs are building clever frameworks on top of the conversion value method that can be utilized to change these values server-side.

23:42
💎 #
8

With the ATT framework coming into effect, probabilistic attribution will become more important. There might be more fraud through an increase of click flooding to gain the attribution.

31:30
💎 #
9

Another potential form of fraud: bogus IDFAs might be created to inflate numbers of IDFA traffic, because that inventory will be premium.

32:40
💎 #
10

If retargeting was important for you before, it's time to double down on lifecycle/CRM: push, liveops, etc. You retarget people when you don't manage to re-engage them with your first party channels. 

37:20
💎 #
11

For a UA manager, optimizing is going to be much harder. But from a strategic standpoint, optimization is going to be done through measuring incrementality.

39:00
💎 #
12

Next year, Google and FB will strategically not be able to just show ads to only 60-70% of the inventory for app campaigns (i.e. not target IDFA-0 users). So if/when they start tracking users through SKAdNetwork the reach will go back to 100% of the inventory (i.e ads will target all users).

52:19
💎 #
13

Apple has not communicated yet how they will track ASA. Users will see ads whether they consent or not. It looks like ASA won't be tracked through SKAdNetwork, and Apple might go with their own system. 

53:40
💎 #
14

According to Kochava, Apple just released an API that allows them to claim attribution on users that did not opt-in through ATT and associate them with keywords.

55:30
💎 #
15

FB is not going to be more expensive on the inventory side, it's going to be cheaper. Especially if they start displaying ads to 100% of the inventory. Another factor that will decrease CPMs is that advertisers won't be willing to pay as much because high-converting users are harder to track. FB should get cheaper and it's an opportunity.

01:05:25
💎 #
16

Something you can work on if you're an advertiser is your ability to really know who your users are through your first-party data so you can go and get high-quality users for your app.

01:09:40
This is premium content. Upgrade your plan in your profile to get access.
💎 #
1

Start getting your legal team involved in the discussions because the context of this change is increased privacy. As you think about prompting for consent to share data, it's also a good time to think about consent management overall from a legislation perspective.

09:14
💎 #
2

Pilot some SKAdNetworks campaigns already: there is no reason to wait until the ATT framework is enforced before doing it. You should be able to see how things are going to look in your MMP.

10:03
💎 #
3

You actually don't need to require consent (depending on your app/model). This decision on requesting consent or not is the starting point before deciding on your data infrastructure, etc. 

12:25
💎 #
4

Already define which conversion values you are going to use. These conversion values are not going to be the same as the current postback you get from MMPs. You need to find the "magic events" you have that are indicators of good long term retention and monetization (run this analysis).

15:00
💎 #
5

If you do request consent, start testing different UX flows to find the best moment (e.g. at the beginning, at a happy moment, etc.).

17:25
💎 #
6

Also ask yourself how important iOS traffic is for you. Maybe you should switch budget to Android, to web traffic, etc.

19:13
💎 #
7

Do not underestimate the work needed to build your attribution with SKAdNetwork and setting up your conversion values. MMPs are building clever frameworks on top of the conversion value method that can be utilized to change these values server-side.

23:42
💎 #
8

With the ATT framework coming into effect, probabilistic attribution will become more important. There might be more fraud through an increase of click flooding to gain the attribution.

31:30
💎 #
9

Another potential form of fraud: bogus IDFAs might be created to inflate numbers of IDFA traffic, because that inventory will be premium.

32:40
💎 #
10

If retargeting was important for you before, it's time to double down on lifecycle/CRM: push, liveops, etc. You retarget people when you don't manage to re-engage them with your first party channels. 

37:20
💎 #
11

For a UA manager, optimizing is going to be much harder. But from a strategic standpoint, optimization is going to be done through measuring incrementality.

39:00
💎 #
12

Next year, Google and FB will strategically not be able to just show ads to only 60-70% of the inventory for app campaigns (i.e. not target IDFA-0 users). So if/when they start tracking users through SKAdNetwork the reach will go back to 100% of the inventory (i.e ads will target all users).

52:19
💎 #
13

Apple has not communicated yet how they will track ASA. Users will see ads whether they consent or not. It looks like ASA won't be tracked through SKAdNetwork, and Apple might go with their own system. 

53:40
💎 #
14

According to Kochava, Apple just released an API that allows them to claim attribution on users that did not opt-in through ATT and associate them with keywords.

55:30
💎 #
15

FB is not going to be more expensive on the inventory side, it's going to be cheaper. Especially if they start displaying ads to 100% of the inventory. Another factor that will decrease CPMs is that advertisers won't be willing to pay as much because high-converting users are harder to track. FB should get cheaper and it's an opportunity.

01:05:25
💎 #
16

Something you can work on if you're an advertiser is your ability to really know who your users are through your first-party data so you can go and get high-quality users for your app.

01:09:40
This content is available to Growth Gems premium members only.
Apply Now
💎 #
1

Start getting your legal team involved in the discussions because the context of this change is increased privacy. As you think about prompting for consent to share data, it's also a good time to think about consent management overall from a legislation perspective.

09:14
💎 #
2

Pilot some SKAdNetworks campaigns already: there is no reason to wait until the ATT framework is enforced before doing it. You should be able to see how things are going to look in your MMP.

10:03
💎 #
3

You actually don't need to require consent (depending on your app/model). This decision on requesting consent or not is the starting point before deciding on your data infrastructure, etc. 

12:25
💎 #
4

Already define which conversion values you are going to use. These conversion values are not going to be the same as the current postback you get from MMPs. You need to find the "magic events" you have that are indicators of good long term retention and monetization (run this analysis).

15:00
💎 #
5

If you do request consent, start testing different UX flows to find the best moment (e.g. at the beginning, at a happy moment, etc.).

17:25
💎 #
6

Also ask yourself how important iOS traffic is for you. Maybe you should switch budget to Android, to web traffic, etc.

19:13
💎 #
7

Do not underestimate the work needed to build your attribution with SKAdNetwork and setting up your conversion values. MMPs are building clever frameworks on top of the conversion value method that can be utilized to change these values server-side.

23:42
💎 #
8

With the ATT framework coming into effect, probabilistic attribution will become more important. There might be more fraud through an increase of click flooding to gain the attribution.

31:30
💎 #
9

Another potential form of fraud: bogus IDFAs might be created to inflate numbers of IDFA traffic, because that inventory will be premium.

32:40
💎 #
10

If retargeting was important for you before, it's time to double down on lifecycle/CRM: push, liveops, etc. You retarget people when you don't manage to re-engage them with your first party channels. 

37:20
💎 #
11

For a UA manager, optimizing is going to be much harder. But from a strategic standpoint, optimization is going to be done through measuring incrementality.

39:00
💎 #
12

Next year, Google and FB will strategically not be able to just show ads to only 60-70% of the inventory for app campaigns (i.e. not target IDFA-0 users). So if/when they start tracking users through SKAdNetwork the reach will go back to 100% of the inventory (i.e ads will target all users).

52:19
💎 #
13

Apple has not communicated yet how they will track ASA. Users will see ads whether they consent or not. It looks like ASA won't be tracked through SKAdNetwork, and Apple might go with their own system. 

53:40
💎 #
14

According to Kochava, Apple just released an API that allows them to claim attribution on users that did not opt-in through ATT and associate them with keywords.

55:30
💎 #
15

FB is not going to be more expensive on the inventory side, it's going to be cheaper. Especially if they start displaying ads to 100% of the inventory. Another factor that will decrease CPMs is that advertisers won't be willing to pay as much because high-converting users are harder to track. FB should get cheaper and it's an opportunity.

01:05:25
💎 #
16

Something you can work on if you're an advertiser is your ability to really know who your users are through your first-party data so you can go and get high-quality users for your app.

01:09:40
This is premium content. Upgrade your plan in your profile to get access.
This content is available to Growth Gems premium members only.
Apply Now
You need to be logged in the community to see the discussion below. You can also head over to
this topic in the community

iOS 14: what's new and how it already affects app marketing

Johannes

  • Giving your IDFA info as a user will be opt-in

Thomas

  • This has a lot of consequences: how to ask for consent, managing your campaigns, etc. A lot of confusions because there is an impact on a lot of things

Ryan

  • SKAdNetwork has been around since 2018 but never really used and only allowed to attribute installs (no post-install event).
  • Apple has now updated to a v2 which includes the possibility to define "conversion value" in order to get post-install performance.

IDFA preparations: what to do to get ready

Ryan

  • There are still a lot of unknowns, so the first thing to do is learn about ATT, SKAdNetwork. Then talk to your partners about it.
  • [💎 @09:14] Start getting your legal team involved in the discussions because the context of this change is increased privacy. As you think about prompting for consent to share data, it's also a good time to think about consent management overall from a legislation perspective.
  • [💎 @10:03] Pilot some SKAdNetworks campaigns already: there is no reason to wait until the ATT framework is enforced before doing it. You should be able to see how things are going to look in your MMP.
  • Tracking users at an individual level and sharing that data is going to go away. So advertisers and publishers should think more about their first party data to understand their users in order to find the right way to both sell and buy advertising.

Thomas

  • [💎 @12:25] You actually don't need to require consent, depending on your app. This decision on requesting consent or not is the starting point before deciding on your data infrastructure, etc.You can live with a group that gave consent and one that didn't but the datasets might be hard to reconcile. So in some cases (not big on UA, not really using the IDFA currently, etc.) you can also decide to not request consent at all in order to not lower the psych level of users during onboarding.
  • [💎 @15:00] Already define which conversion values you are going to use. These conversion values are not going to be the same as the current postback you get from MMPs. You need to find the "magic events" you have that are indicators of good long term retention and monetization (run this analysis).

Andrea

  • [💎 @17:25] If you do request consent, start testing different UX flows to find the best moment (e.g. at the beginning, at a happy moment, etc.).

Johannes

  • How much you're affected by this change depends on your category and your business model
  • [💎 @19:13] Also ask yourself how important iOS traffic is for you. Maybe you should switch budget to Android, to web traffic, etc.

Role of MMPs

Ryan

  • Back in 2012: some people underestimated how much work it was to build your own attribution tools
  • There will always be a need for a neutral partner to have a single source of truth: to ingest reporting data, to look at discrepancies, etc.
  • [💎 @23:42] Do not underestimate the work needed to build your attribution with SKAdNetwork and setting up your conversion values. MMPs are building clever frameworks on top of the conversion value method that can be utilized to be able to change these values server-side.

How ad networks will be affected?

Andrea - there will be losers and winners

  • Losers
  • FB will loose, especially for FAN and LAL. Inventories are already declining because they do not target IDFA-0 users.
  • IDFA-0 at 34%. Adoption of iOS 14 is fast: 50% in October, 70% in November.
  • FB had said that 50% of their revenue could drop without personalized ads
  • Winners
  • ASA
  • ASO more important
  • Also start thinking of new "blue ocean" strategies and channels, not the typical duopoly

Will fraud decrease for iOS?

Ryan

  • 2 potential areas for fraud
  • [💎 @31:30] With the ATT framework coming into effect, probabilistic attribution will become more important. There might be more fraud through an increase of click flooding to gain the attribution.
  • [💎 @32:40] Another potential form of fraud: bogus IDFAs might be created to inflate numbers of IDFA traffic, because that inventory will be premium.

What will happen to retargeting

Johannes

  • Retargeting on iOS will die. It is very tied to the UX of an app: deep linking, etc. It's already hard to do now, but it will be even harder.
  • You'll be able to retarget on a higher level only (e.g. users that just installed the app)

How to work in the post-IDFA reality

Thomas

  • [💎 @37:20] If retargeting was important for you before, it's time to double down on lifecycle/CRM: push, liveops, etc. You retarget people when you don't manage to re-engage them with your first party channels.
  • A lot of people are not prepared of what's ahead of us in terms of how the data will look like. You can already go to your MMP to see how your campaigns will look like.

A lot of people are not prepared of what's ahead of us in terms of how the data will look like

  • [💎 @39:00] For a UA manager, optimizing is going to be much harder. But from a strategic standpoint, optimization is going to be done through measuring incrementality.
  • UA focus has been evolving: CPI → ROAS → Payback Period. Now it is incrementality.
  • We'll have to be more strategic when it comes to optimizing. Having less precision will force us to become more strategic.

Ryan

  • If your business model relies on identifying whales (spending a lot on IAPs), it's going to be much harder

How SKAdNetwork works and what are the limitations

Andrea

  • We'll have aggregated data only (not user-level), moving away from deterministic and towards probabilistic.
  • Only 63 conversion events so you'll need to be strategic. Also moving from real time optimization to delayed data (without precise timestamps).
  • No cohorted data.
  • Ad fraud is another limitation. Event conversion values are unsigned and you can not verify when the events happened exactly.
  • SKAdNetwork does not support deep linking either

Apple Search Ads: changes and opportunities

Johannes

  • ASA might be the only way to reach users without consent. However search has limited scalability so it might not have a huge effect.
  • Unless Apple moves to something more display (Apple News, Apple Stock, etc.) and have more inventory.
  • "LAT On" users should be treated as a different user group: more privacy-focused, know how things work.

Thomas

  • Today, not showing ads to IDFA-0 users on FB/Google for app installs.
  • On ASA you can show the ads but you can not track them.
  • [💎 @52:19] Next year, Google and FB will strategically not be able to just show ads to only 60-70% of the inventory for app campaigns. So if/when they start tracking users through SKAdNetwork the reach will go back to 100% of the inventory (i.e ads will target all users).
  • We don't really know what will happen with Apple's "display" network. Apple News and Stock are still very limited, it's not at the scale of other ad networks.
  • [💎 @53:40] Apple has not communicated how they will track ASA. Users will see ads whether they consent or not. It looks like ASA won't be tracked through SKAdNetwork, and Apple might go with their own system.
  • From a legal and anticompetitive point of view it is a grey area. How will Apple get away with it?

Ryan

  • [💎 @55:30] Apple just released an API that allows them to claim attribution on users that did not opt-in through ATT and associate them with keywords.

What changes will happen in Apple's services and ads business

Ryan

  • A lot of developers might be rethinking their business model. There might be a switch to higher-value services, to subscriptions, etc.

Thomas

  • When Apple announced the change some people said it was for privacy, or to boost their ad service business. Apple doesn't get any money when developers monetize through ads. They are also forcing a shift in business model where more developers monetize through IAPs and subscriptions so they can get a cut.

Andrea

  • Apple trying to become the leader in tech, and the move hurts FB.
  • They will also force Google to take the same kind of steps on privacy.

Can ASO mitigate the consequences of IDFA deprecation?

Andrea

  • Budgets distributed more evenly, and there will be more attention to ASO. It's not a solution to the IDFA challenges. But maybe we'll see an increase in testing new assets, etc.

Johannes

  • iOS going away from strong ROAS focus and to awareness/branding.
  • Business goals might shift to getting more reach
  • Traffic might shift to Android and web

How will performance marketing change?

Thomas

  • [💎 @01:05:25] FB is not going to be more expensive on the inventory side, it's going to be cheaper. Especially if they start displaying ads to 100% of the inventory. Another factor that will decrease CPMs is that advertisers won't be willing to pay as much because high-converting users are harder to track. FB should get cheaper and it's an opportunity.

App Marketing: what works today and what will

Ryan

  • More importance on very good first party data.
  • On the publisher side to really be able to segment your users so you can monetize your audience
  • [💎 @01:09:40] Something you can work on if you're an advertiser is your ability to really know who your users are through your first-party data so you can go and get high-quality users for your app.

Thomas

  • Thinking about hacks is the wrong way. You need to think long term.
  • Some people have been focusing on the IDFV: growing the internal portfolio through acquisitions so they can keep tracking users. On the flip side with the app store fees lowered for smaller developers, there is less of an incentive to group apps.

Johannes

  • Invest early on in new channels: ASA, TikTok, etc. Don't wait until everyone does it.

Q&A

  • LATOn/LATOff
  • Andrea: Before iOS 14, users were split between LAT On and LAT Off. iOS 14 users are not in either of these categories anymore, just in total downloads.
  • How SKAdNetwork could evolve
  • Thomas: it is unlikely that you will get any conversion after day 1 because of how the timer is build for SKAdNetwork. That's the wrong direction to send developers on. Extending the window of time to get the postback of conversion value so apps don't become too aggressive on monetization.
  • Fingerprinting off the table for MMPs/DSPs?
  • Ryan: that's a big unknown. Struggles to see how probabilistic attribution is going to go away. It would destroy some big segments of the advertising landscape that would also impact Apple as well.
  • Price of traffic?
  • Johannes: might be cheaper on Facebook because advertisers are spending less (as Thomas mentioned above)
  • Getting rid of LAT On users?
  • You should keep targeting LAT On users, otherwise you're giving up a valuable inventory.
  • You can calculate how many iOS 14 users you're targeting by taking your total downloads and subtracting LAT On + LAT Off downloads.
  • IDFA gone?
  • Thomas: IDFA is not completely gone. It is gone as the standard for attribution. But you may want to ask for consent to get the IDFA, and that's what Google is pushing you to do so that they can keep feeding the algorithm.
  • IDFA is not completely gone. It is gone as the standard for attribution
  • Johannes: Google is in a pretty good spot to sit and wait
  • How to calculate conversion metrics?
  • Thomas: attribution is going to be different, and you won't be able to compare directly. You'll need to find your new target.

This is premium content. Upgrade your plan in your profile to get access.

iOS 14: what's new and how it already affects app marketing

Johannes

  • Giving your IDFA info as a user will be opt-in

Thomas

  • This has a lot of consequences: how to ask for consent, managing your campaigns, etc. A lot of confusions because there is an impact on a lot of things

Ryan

  • SKAdNetwork has been around since 2018 but never really used and only allowed to attribute installs (no post-install event).
  • Apple has now updated to a v2 which includes the possibility to define "conversion value" in order to get post-install performance.

IDFA preparations: what to do to get ready

Ryan

  • There are still a lot of unknowns, so the first thing to do is learn about ATT, SKAdNetwork. Then talk to your partners about it.
  • [💎 @09:14] Start getting your legal team involved in the discussions because the context of this change is increased privacy. As you think about prompting for consent to share data, it's also a good time to think about consent management overall from a legislation perspective.
  • [💎 @10:03] Pilot some SKAdNetworks campaigns already: there is no reason to wait until the ATT framework is enforced before doing it. You should be able to see how things are going to look in your MMP.
  • Tracking users at an individual level and sharing that data is going to go away. So advertisers and publishers should think more about their first party data to understand their users in order to find the right way to both sell and buy advertising.

Thomas

  • [💎 @12:25] You actually don't need to require consent, depending on your app. This decision on requesting consent or not is the starting point before deciding on your data infrastructure, etc.You can live with a group that gave consent and one that didn't but the datasets might be hard to reconcile. So in some cases (not big on UA, not really using the IDFA currently, etc.) you can also decide to not request consent at all in order to not lower the psych level of users during onboarding.
  • [💎 @15:00] Already define which conversion values you are going to use. These conversion values are not going to be the same as the current postback you get from MMPs. You need to find the "magic events" you have that are indicators of good long term retention and monetization (run this analysis).

Andrea

  • [💎 @17:25] If you do request consent, start testing different UX flows to find the best moment (e.g. at the beginning, at a happy moment, etc.).

Johannes

  • How much you're affected by this change depends on your category and your business model
  • [💎 @19:13] Also ask yourself how important iOS traffic is for you. Maybe you should switch budget to Android, to web traffic, etc.

Role of MMPs

Ryan

  • Back in 2012: some people underestimated how much work it was to build your own attribution tools
  • There will always be a need for a neutral partner to have a single source of truth: to ingest reporting data, to look at discrepancies, etc.
  • [💎 @23:42] Do not underestimate the work needed to build your attribution with SKAdNetwork and setting up your conversion values. MMPs are building clever frameworks on top of the conversion value method that can be utilized to be able to change these values server-side.

How ad networks will be affected?

Andrea - there will be losers and winners

  • Losers
  • FB will loose, especially for FAN and LAL. Inventories are already declining because they do not target IDFA-0 users.
  • IDFA-0 at 34%. Adoption of iOS 14 is fast: 50% in October, 70% in November.
  • FB had said that 50% of their revenue could drop without personalized ads
  • Winners
  • ASA
  • ASO more important
  • Also start thinking of new "blue ocean" strategies and channels, not the typical duopoly

Will fraud decrease for iOS?

Ryan

  • 2 potential areas for fraud
  • [💎 @31:30] With the ATT framework coming into effect, probabilistic attribution will become more important. There might be more fraud through an increase of click flooding to gain the attribution.
  • [💎 @32:40] Another potential form of fraud: bogus IDFAs might be created to inflate numbers of IDFA traffic, because that inventory will be premium.

What will happen to retargeting

Johannes

  • Retargeting on iOS will die. It is very tied to the UX of an app: deep linking, etc. It's already hard to do now, but it will be even harder.
  • You'll be able to retarget on a higher level only (e.g. users that just installed the app)

How to work in the post-IDFA reality

Thomas

  • [💎 @37:20] If retargeting was important for you before, it's time to double down on lifecycle/CRM: push, liveops, etc. You retarget people when you don't manage to re-engage them with your first party channels.
  • A lot of people are not prepared of what's ahead of us in terms of how the data will look like. You can already go to your MMP to see how your campaigns will look like.

A lot of people are not prepared of what's ahead of us in terms of how the data will look like

  • [💎 @39:00] For a UA manager, optimizing is going to be much harder. But from a strategic standpoint, optimization is going to be done through measuring incrementality.
  • UA focus has been evolving: CPI → ROAS → Payback Period. Now it is incrementality.
  • We'll have to be more strategic when it comes to optimizing. Having less precision will force us to become more strategic.

Ryan

  • If your business model relies on identifying whales (spending a lot on IAPs), it's going to be much harder

How SKAdNetwork works and what are the limitations

Andrea

  • We'll have aggregated data only (not user-level), moving away from deterministic and towards probabilistic.
  • Only 63 conversion events so you'll need to be strategic. Also moving from real time optimization to delayed data (without precise timestamps).
  • No cohorted data.
  • Ad fraud is another limitation. Event conversion values are unsigned and you can not verify when the events happened exactly.
  • SKAdNetwork does not support deep linking either

Apple Search Ads: changes and opportunities

Johannes

  • ASA might be the only way to reach users without consent. However search has limited scalability so it might not have a huge effect.
  • Unless Apple moves to something more display (Apple News, Apple Stock, etc.) and have more inventory.
  • "LAT On" users should be treated as a different user group: more privacy-focused, know how things work.

Thomas

  • Today, not showing ads to IDFA-0 users on FB/Google for app installs.
  • On ASA you can show the ads but you can not track them.
  • [💎 @52:19] Next year, Google and FB will strategically not be able to just show ads to only 60-70% of the inventory for app campaigns. So if/when they start tracking users through SKAdNetwork the reach will go back to 100% of the inventory (i.e ads will target all users).
  • We don't really know what will happen with Apple's "display" network. Apple News and Stock are still very limited, it's not at the scale of other ad networks.
  • [💎 @53:40] Apple has not communicated how they will track ASA. Users will see ads whether they consent or not. It looks like ASA won't be tracked through SKAdNetwork, and Apple might go with their own system.
  • From a legal and anticompetitive point of view it is a grey area. How will Apple get away with it?

Ryan

  • [💎 @55:30] Apple just released an API that allows them to claim attribution on users that did not opt-in through ATT and associate them with keywords.

What changes will happen in Apple's services and ads business

Ryan

  • A lot of developers might be rethinking their business model. There might be a switch to higher-value services, to subscriptions, etc.

Thomas

  • When Apple announced the change some people said it was for privacy, or to boost their ad service business. Apple doesn't get any money when developers monetize through ads. They are also forcing a shift in business model where more developers monetize through IAPs and subscriptions so they can get a cut.

Andrea

  • Apple trying to become the leader in tech, and the move hurts FB.
  • They will also force Google to take the same kind of steps on privacy.

Can ASO mitigate the consequences of IDFA deprecation?

Andrea

  • Budgets distributed more evenly, and there will be more attention to ASO. It's not a solution to the IDFA challenges. But maybe we'll see an increase in testing new assets, etc.

Johannes

  • iOS going away from strong ROAS focus and to awareness/branding.
  • Business goals might shift to getting more reach
  • Traffic might shift to Android and web

How will performance marketing change?

Thomas

  • [💎 @01:05:25] FB is not going to be more expensive on the inventory side, it's going to be cheaper. Especially if they start displaying ads to 100% of the inventory. Another factor that will decrease CPMs is that advertisers won't be willing to pay as much because high-converting users are harder to track. FB should get cheaper and it's an opportunity.

App Marketing: what works today and what will

Ryan

  • More importance on very good first party data.
  • On the publisher side to really be able to segment your users so you can monetize your audience
  • [💎 @01:09:40] Something you can work on if you're an advertiser is your ability to really know who your users are through your first-party data so you can go and get high-quality users for your app.

Thomas

  • Thinking about hacks is the wrong way. You need to think long term.
  • Some people have been focusing on the IDFV: growing the internal portfolio through acquisitions so they can keep tracking users. On the flip side with the app store fees lowered for smaller developers, there is less of an incentive to group apps.

Johannes

  • Invest early on in new channels: ASA, TikTok, etc. Don't wait until everyone does it.

Q&A

  • LATOn/LATOff
  • Andrea: Before iOS 14, users were split between LAT On and LAT Off. iOS 14 users are not in either of these categories anymore, just in total downloads.
  • How SKAdNetwork could evolve
  • Thomas: it is unlikely that you will get any conversion after day 1 because of how the timer is build for SKAdNetwork. That's the wrong direction to send developers on. Extending the window of time to get the postback of conversion value so apps don't become too aggressive on monetization.
  • Fingerprinting off the table for MMPs/DSPs?
  • Ryan: that's a big unknown. Struggles to see how probabilistic attribution is going to go away. It would destroy some big segments of the advertising landscape that would also impact Apple as well.
  • Price of traffic?
  • Johannes: might be cheaper on Facebook because advertisers are spending less (as Thomas mentioned above)
  • Getting rid of LAT On users?
  • You should keep targeting LAT On users, otherwise you're giving up a valuable inventory.
  • You can calculate how many iOS 14 users you're targeting by taking your total downloads and subtracting LAT On + LAT Off downloads.
  • IDFA gone?
  • Thomas: IDFA is not completely gone. It is gone as the standard for attribution. But you may want to ask for consent to get the IDFA, and that's what Google is pushing you to do so that they can keep feeding the algorithm.
  • IDFA is not completely gone. It is gone as the standard for attribution
  • Johannes: Google is in a pretty good spot to sit and wait
  • How to calculate conversion metrics?
  • Thomas: attribution is going to be different, and you won't be able to compare directly. You'll need to find your new target.

This content is available to Growth Gems premium members only.
Apply Now

iOS 14: what's new and how it already affects app marketing

Johannes

  • Giving your IDFA info as a user will be opt-in

Thomas

  • This has a lot of consequences: how to ask for consent, managing your campaigns, etc. A lot of confusions because there is an impact on a lot of things

Ryan

  • SKAdNetwork has been around since 2018 but never really used and only allowed to attribute installs (no post-install event).
  • Apple has now updated to a v2 which includes the possibility to define "conversion value" in order to get post-install performance.

IDFA preparations: what to do to get ready

Ryan

  • There are still a lot of unknowns, so the first thing to do is learn about ATT, SKAdNetwork. Then talk to your partners about it.
  • [💎 @09:14] Start getting your legal team involved in the discussions because the context of this change is increased privacy. As you think about prompting for consent to share data, it's also a good time to think about consent management overall from a legislation perspective.
  • [💎 @10:03] Pilot some SKAdNetworks campaigns already: there is no reason to wait until the ATT framework is enforced before doing it. You should be able to see how things are going to look in your MMP.
  • Tracking users at an individual level and sharing that data is going to go away. So advertisers and publishers should think more about their first party data to understand their users in order to find the right way to both sell and buy advertising.

Thomas

  • [💎 @12:25] You actually don't need to require consent, depending on your app. This decision on requesting consent or not is the starting point before deciding on your data infrastructure, etc.You can live with a group that gave consent and one that didn't but the datasets might be hard to reconcile. So in some cases (not big on UA, not really using the IDFA currently, etc.) you can also decide to not request consent at all in order to not lower the psych level of users during onboarding.
  • [💎 @15:00] Already define which conversion values you are going to use. These conversion values are not going to be the same as the current postback you get from MMPs. You need to find the "magic events" you have that are indicators of good long term retention and monetization (run this analysis).

Andrea

  • [💎 @17:25] If you do request consent, start testing different UX flows to find the best moment (e.g. at the beginning, at a happy moment, etc.).

Johannes

  • How much you're affected by this change depends on your category and your business model
  • [💎 @19:13] Also ask yourself how important iOS traffic is for you. Maybe you should switch budget to Android, to web traffic, etc.

Role of MMPs

Ryan

  • Back in 2012: some people underestimated how much work it was to build your own attribution tools
  • There will always be a need for a neutral partner to have a single source of truth: to ingest reporting data, to look at discrepancies, etc.
  • [💎 @23:42] Do not underestimate the work needed to build your attribution with SKAdNetwork and setting up your conversion values. MMPs are building clever frameworks on top of the conversion value method that can be utilized to be able to change these values server-side.

How ad networks will be affected?

Andrea - there will be losers and winners

  • Losers
  • FB will loose, especially for FAN and LAL. Inventories are already declining because they do not target IDFA-0 users.
  • IDFA-0 at 34%. Adoption of iOS 14 is fast: 50% in October, 70% in November.
  • FB had said that 50% of their revenue could drop without personalized ads
  • Winners
  • ASA
  • ASO more important
  • Also start thinking of new "blue ocean" strategies and channels, not the typical duopoly

Will fraud decrease for iOS?

Ryan

  • 2 potential areas for fraud
  • [💎 @31:30] With the ATT framework coming into effect, probabilistic attribution will become more important. There might be more fraud through an increase of click flooding to gain the attribution.
  • [💎 @32:40] Another potential form of fraud: bogus IDFAs might be created to inflate numbers of IDFA traffic, because that inventory will be premium.

What will happen to retargeting

Johannes

  • Retargeting on iOS will die. It is very tied to the UX of an app: deep linking, etc. It's already hard to do now, but it will be even harder.
  • You'll be able to retarget on a higher level only (e.g. users that just installed the app)

How to work in the post-IDFA reality

Thomas

  • [💎 @37:20] If retargeting was important for you before, it's time to double down on lifecycle/CRM: push, liveops, etc. You retarget people when you don't manage to re-engage them with your first party channels.
  • A lot of people are not prepared of what's ahead of us in terms of how the data will look like. You can already go to your MMP to see how your campaigns will look like.

A lot of people are not prepared of what's ahead of us in terms of how the data will look like

  • [💎 @39:00] For a UA manager, optimizing is going to be much harder. But from a strategic standpoint, optimization is going to be done through measuring incrementality.
  • UA focus has been evolving: CPI → ROAS → Payback Period. Now it is incrementality.
  • We'll have to be more strategic when it comes to optimizing. Having less precision will force us to become more strategic.

Ryan

  • If your business model relies on identifying whales (spending a lot on IAPs), it's going to be much harder

How SKAdNetwork works and what are the limitations

Andrea

  • We'll have aggregated data only (not user-level), moving away from deterministic and towards probabilistic.
  • Only 63 conversion events so you'll need to be strategic. Also moving from real time optimization to delayed data (without precise timestamps).
  • No cohorted data.
  • Ad fraud is another limitation. Event conversion values are unsigned and you can not verify when the events happened exactly.
  • SKAdNetwork does not support deep linking either

Apple Search Ads: changes and opportunities

Johannes

  • ASA might be the only way to reach users without consent. However search has limited scalability so it might not have a huge effect.
  • Unless Apple moves to something more display (Apple News, Apple Stock, etc.) and have more inventory.
  • "LAT On" users should be treated as a different user group: more privacy-focused, know how things work.

Thomas

  • Today, not showing ads to IDFA-0 users on FB/Google for app installs.
  • On ASA you can show the ads but you can not track them.
  • [💎 @52:19] Next year, Google and FB will strategically not be able to just show ads to only 60-70% of the inventory for app campaigns. So if/when they start tracking users through SKAdNetwork the reach will go back to 100% of the inventory (i.e ads will target all users).
  • We don't really know what will happen with Apple's "display" network. Apple News and Stock are still very limited, it's not at the scale of other ad networks.
  • [💎 @53:40] Apple has not communicated how they will track ASA. Users will see ads whether they consent or not. It looks like ASA won't be tracked through SKAdNetwork, and Apple might go with their own system.
  • From a legal and anticompetitive point of view it is a grey area. How will Apple get away with it?

Ryan

  • [💎 @55:30] Apple just released an API that allows them to claim attribution on users that did not opt-in through ATT and associate them with keywords.

What changes will happen in Apple's services and ads business

Ryan

  • A lot of developers might be rethinking their business model. There might be a switch to higher-value services, to subscriptions, etc.

Thomas

  • When Apple announced the change some people said it was for privacy, or to boost their ad service business. Apple doesn't get any money when developers monetize through ads. They are also forcing a shift in business model where more developers monetize through IAPs and subscriptions so they can get a cut.

Andrea

  • Apple trying to become the leader in tech, and the move hurts FB.
  • They will also force Google to take the same kind of steps on privacy.

Can ASO mitigate the consequences of IDFA deprecation?

Andrea

  • Budgets distributed more evenly, and there will be more attention to ASO. It's not a solution to the IDFA challenges. But maybe we'll see an increase in testing new assets, etc.

Johannes

  • iOS going away from strong ROAS focus and to awareness/branding.
  • Business goals might shift to getting more reach
  • Traffic might shift to Android and web

How will performance marketing change?

Thomas

  • [💎 @01:05:25] FB is not going to be more expensive on the inventory side, it's going to be cheaper. Especially if they start displaying ads to 100% of the inventory. Another factor that will decrease CPMs is that advertisers won't be willing to pay as much because high-converting users are harder to track. FB should get cheaper and it's an opportunity.

App Marketing: what works today and what will

Ryan

  • More importance on very good first party data.
  • On the publisher side to really be able to segment your users so you can monetize your audience
  • [💎 @01:09:40] Something you can work on if you're an advertiser is your ability to really know who your users are through your first-party data so you can go and get high-quality users for your app.

Thomas

  • Thinking about hacks is the wrong way. You need to think long term.
  • Some people have been focusing on the IDFV: growing the internal portfolio through acquisitions so they can keep tracking users. On the flip side with the app store fees lowered for smaller developers, there is less of an incentive to group apps.

Johannes

  • Invest early on in new channels: ASA, TikTok, etc. Don't wait until everyone does it.

Q&A

  • LATOn/LATOff
  • Andrea: Before iOS 14, users were split between LAT On and LAT Off. iOS 14 users are not in either of these categories anymore, just in total downloads.
  • How SKAdNetwork could evolve
  • Thomas: it is unlikely that you will get any conversion after day 1 because of how the timer is build for SKAdNetwork. That's the wrong direction to send developers on. Extending the window of time to get the postback of conversion value so apps don't become too aggressive on monetization.
  • Fingerprinting off the table for MMPs/DSPs?
  • Ryan: that's a big unknown. Struggles to see how probabilistic attribution is going to go away. It would destroy some big segments of the advertising landscape that would also impact Apple as well.
  • Price of traffic?
  • Johannes: might be cheaper on Facebook because advertisers are spending less (as Thomas mentioned above)
  • Getting rid of LAT On users?
  • You should keep targeting LAT On users, otherwise you're giving up a valuable inventory.
  • You can calculate how many iOS 14 users you're targeting by taking your total downloads and subtracting LAT On + LAT Off downloads.
  • IDFA gone?
  • Thomas: IDFA is not completely gone. It is gone as the standard for attribution. But you may want to ask for consent to get the IDFA, and that's what Google is pushing you to do so that they can keep feeding the algorithm.
  • IDFA is not completely gone. It is gone as the standard for attribution
  • Johannes: Google is in a pretty good spot to sit and wait
  • How to calculate conversion metrics?
  • Thomas: attribution is going to be different, and you won't be able to compare directly. You'll need to find your new target.