Building a Strong Marketing Stack, UA for Subscription Apps and Creatives

Item not bookmarked
Resource bookmarked
Bookmarking...
Review star
Review star
Review star
Review star
Review star
💎  x
10

Thomas Petit (Growth Consultant) talks about creativity in marketing, how to build a world-class marketing stack, challenges and consideration of subscription apps.

Source:
Building a Strong Marketing Stack, UA for Subscription Apps and Creatives
(no direct link to watch/listen)
(direct link to watch/listen)
Type:
Podcast
Publication date:
July 29, 2019
Added to the Vault on:
April 27, 2020
These insights were shared through the free Growth Gems newsletter.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
💎 #
1

Investing in data is usually an excellent choice, regardless if your build or buy (there are more solutions nowadays). 

06:00
💎 #
2

The big focus of the data stack at 8fit was how to predict revenue the earliest possible and in the most accurate way in order to know what to re-inject in acquisition. 

08:00
💎 #
3

If you focus too much on people converting to free trial you might attract more people who don't convert after. For example younger audiences are a lot more aware about how to cancel a free trial so you might get good CPAs but a cohort that converts poorly. 

09:23
💎 #
4

Constant testing on offers and pricing is the only way to survive for subscription apps. What you might have learned 3 years ago might be different today: audience you are approaching or overall market could be different, there might be subscription fatigue, etc.  

12:12
💎 #
5

Something that worked is putting a longer funnel before the install. Example: sending people from Outbrain/Taboola to a landing page (article form) where 8fit had a lot more space to tell people what makes the app different. 

21:20
💎 #
6

On the App Store there is a bottleneck: everybody sees the same thing no matter which USP you advertise. Via web you can create a more coherent journey so people understand why they are going to see the App Store. It created a virtuous cycle (better install conversion, better place in the auction, ASO, etc.). 

23:45
💎 #
7

Even though "inspirational" videos did not perform well on Facebook, what did work well was to retarget the video viewers with direct response ads and a more direct messaging.  

26:25
💎 #
8

Turns out Google has already developed internally predictive testing and can predict by 70% if a creative is going to outperform the current one or not before they spend money. 

31:00
💎 #
9

When you find a creative that really performs on the top of the funnel (clicks, installs, etc.) it is extremely fast to see it. The algorithm makes most of the delivery go to it, you can scale your campaigns, etc. It is obvious. But it doesn't mean that it's the best for the long term because the cohort needs to mature. 

32:45
💎 #
10

When using AEO and VO it's about which event you're feeding the algorithm/machine. Maybe you don't optimize for sign ups, or free trial, but for "free trial + this event completed within 24 hours". 

35:30
The "gems" from this resource are only available to premium members.
  • Unlock access to gems from over 130 mobile growth resources
  • Define your preferred categories and receive new relevant gems directly in your inbox
  • Discuss key insights (and any other mobile growth topic) in the members-only community.
Upgrade Your Plan
💎 #
1

Investing in data is usually an excellent choice, regardless if your build or buy (there are more solutions nowadays). 

06:00
💎 #
2

The big focus of the data stack at 8fit was how to predict revenue the earliest possible and in the most accurate way in order to know what to re-inject in acquisition. 

08:00
💎 #
3

If you focus too much on people converting to free trial you might attract more people who don't convert after. For example younger audiences are a lot more aware about how to cancel a free trial so you might get good CPAs but a cohort that converts poorly. 

09:23
💎 #
4

Constant testing on offers and pricing is the only way to survive for subscription apps. What you might have learned 3 years ago might be different today: audience you are approaching or overall market could be different, there might be subscription fatigue, etc.  

12:12
💎 #
5

Something that worked is putting a longer funnel before the install. Example: sending people from Outbrain/Taboola to a landing page (article form) where 8fit had a lot more space to tell people what makes the app different. 

21:20
💎 #
6

On the App Store there is a bottleneck: everybody sees the same thing no matter which USP you advertise. Via web you can create a more coherent journey so people understand why they are going to see the App Store. It created a virtuous cycle (better install conversion, better place in the auction, ASO, etc.). 

23:45
💎 #
7

Even though "inspirational" videos did not perform well on Facebook, what did work well was to retarget the video viewers with direct response ads and a more direct messaging.  

26:25
💎 #
8

Turns out Google has already developed internally predictive testing and can predict by 70% if a creative is going to outperform the current one or not before they spend money. 

31:00
💎 #
9

When you find a creative that really performs on the top of the funnel (clicks, installs, etc.) it is extremely fast to see it. The algorithm makes most of the delivery go to it, you can scale your campaigns, etc. It is obvious. But it doesn't mean that it's the best for the long term because the cohort needs to mature. 

32:45
💎 #
10

When using AEO and VO it's about which event you're feeding the algorithm/machine. Maybe you don't optimize for sign ups, or free trial, but for "free trial + this event completed within 24 hours". 

35:30
The "gems" from this resource are only available to premium members.

Gems are the key bite-size insights "mined" from a specific mobile marketing resource, like a webinar, a panel or a podcast.
They allow you to save time by grasping the most important information in a couple of minutes, and also each include the timestamp from the source.

Become a member to:
  • Unlock access to gems from over 130 mobile growth resources
  • Define your preferred categories and receive new relevant gems directly in your inbox
  • Discuss key insights (and any other mobile growth topic) in the members-only community.
Request Access
💎 #
1

Investing in data is usually an excellent choice, regardless if your build or buy (there are more solutions nowadays). 

06:00
💎 #
2

The big focus of the data stack at 8fit was how to predict revenue the earliest possible and in the most accurate way in order to know what to re-inject in acquisition. 

08:00
💎 #
3

If you focus too much on people converting to free trial you might attract more people who don't convert after. For example younger audiences are a lot more aware about how to cancel a free trial so you might get good CPAs but a cohort that converts poorly. 

09:23
💎 #
4

Constant testing on offers and pricing is the only way to survive for subscription apps. What you might have learned 3 years ago might be different today: audience you are approaching or overall market could be different, there might be subscription fatigue, etc.  

12:12
💎 #
5

Something that worked is putting a longer funnel before the install. Example: sending people from Outbrain/Taboola to a landing page (article form) where 8fit had a lot more space to tell people what makes the app different. 

21:20
💎 #
6

On the App Store there is a bottleneck: everybody sees the same thing no matter which USP you advertise. Via web you can create a more coherent journey so people understand why they are going to see the App Store. It created a virtuous cycle (better install conversion, better place in the auction, ASO, etc.). 

23:45
💎 #
7

Even though "inspirational" videos did not perform well on Facebook, what did work well was to retarget the video viewers with direct response ads and a more direct messaging.  

26:25
💎 #
8

Turns out Google has already developed internally predictive testing and can predict by 70% if a creative is going to outperform the current one or not before they spend money. 

31:00
💎 #
9

When you find a creative that really performs on the top of the funnel (clicks, installs, etc.) it is extremely fast to see it. The algorithm makes most of the delivery go to it, you can scale your campaigns, etc. It is obvious. But it doesn't mean that it's the best for the long term because the cohort needs to mature. 

32:45
💎 #
10

When using AEO and VO it's about which event you're feeding the algorithm/machine. Maybe you don't optimize for sign ups, or free trial, but for "free trial + this event completed within 24 hours". 

35:30
The access to discussions on each resource is only available to premium members.

Growth Gems members discuss resources and their key insights (as well as other mobile growth topics) in the community. It's the perfect way to dig deeper, ask questions and get additional perspectives.
Upgrade to premium to:
  • Unlock access to key insights from over 130 mobile growth resources
  • Define your preferred categories and receive new relevant gems directly in your inbox
  • Discuss key insights (and any other mobile growth topic) in the members-only community.
Upgrade Your Plan
The access to discussions on each resource is only available to premium members.

Growth Gems members discuss resources and their key insights (as well as other mobile growth topics) in the community. It's the perfect way to dig deeper, ask questions and get additional perspectives.
Become a member to:
  • Unlock access to gems from over 130 mobile growth resources
  • Define your preferred categories and receive new relevant gems directly in your inbox
  • Discuss key insights (and any other mobile growth topic) in the members-only community.
Request Access

You need to be logged in the community to be able to see the discussion below.
You can also head over directly to this topic in the community

Notes for this resource are currently being transferred and will be available soon.

Where to start if you're building out a team for UA

Back then, people with experience were in gaming and it was hard to attract them. So hired people based on mindset (even if they were in the web) rather than experience and not only UA people.


Biggest takeaway for 8fit

They understood very only on that they needed a solid data stack and accessibility to data. They built they own data stack, scalable and custom to what they needed.

[💎@06:00] Investing in data is usually an excellent choice, regardless if your build or buy (there are more solutions nowadays).


Goal of the data stack

Subscription acquisition flow is not that simple: start a free trial or not, do they convert, do they refund, modelize the renewals, etc.


  • Pure analytics: relied on tooling (Amplitude)
  • Revenue modelization: had to do everything on their own, because the data Apple shares is a bit messy and hard to reconcile.


[💎@08:00] The big focus of the data stack at 8fit was how to predict revenue the earliest possible and in the most accurate way in order to know what to re)inject in acquisition.

Another important part was uniting all networks and costs to have all conversions in one place: every single step of the funnel from every single source and being able to segment by anything (OS, gender, activity, etc.). There are now new products in the market for that.


Challenges of acquisition for subscription apps

Challenge #1

Most subscriptions use "start a free trial" as an event for optimization (Facebook, UAC).

[💎@09:23] If you focus too much on people converting to free trial you might attract more people who don't convert after. For example younger audiences are a lot more aware about how to cancel a free trial so you might get good CPAs but a cohort that converts poorly.


In short: the algorithms from FB/UAC really deliver what you ask for.


Challenge #2

On the user side there is only one subscription to buy but on the developer side you have many (can be 100+): iOS and Android, 7 or 30 day trial, prices different by country, super premium subscription. This increases complexity as well.


For 8fit it was in the company mindset to constantly test pricing: introductory offers (free trials), removing the free trial, etc.


[💎@12:12] Constant testing on offers and pricing is the only way to survive for subscription apps. What you might have learned 3 years ago might be different today: audience you are approaching or overall market could be different, there might be subscription fatigue, etc.


How to differentiate yourself when doing performance marketing?

For 8fit, they came from the nutrition angle so that USP was doing good on conversion. That's therefore where a lot of the UA efforts were placed, as well as: "it's useless to go train everyday if you eat like shit".


8fit was against the kind of "quick fix" message and were into long-term healthy choice in life, education, etc.

Focusing on the long-term and specific positioning might not bring some of the quick wins but it helps building over time.


Calling out the free trial vs. positioning in advertising?

Thomas had a couple of cohorts showing that pushing free stuff was not the right choice in the long term. He banned the word "free" from the communication of the company besides inside the app: no "free download", no "free trial".


Replaced the attractiveness of the free trial with a push on personalization. So that when people hit the paywall they have a reason to subscribe: the product is built specifically for them.


Educating an audience

When pushing 2 messages at the same time, it would not work. Example:

  • Fitness + nutrition
  • Personalization + trial
  • Etc.

→ Had to deliver 1 message only. Video helped.


[💎@21:20] Something that worked is putting a longer funnel before the install. Example: sending people from Outbrain/Taboola to a landing page (article form) where 8fit had a lot more space to tell people what makes the app different.


Because of the drop off on the landing page the install would be more expensive but there was higher conversion, retention and renewal were compensating this effect. Long term gains were everywhere: users subscribed to push, were more receptive to promos, engaged on social, etc.


👉 Check [RESOURCE #16] Better Paid Content Marketing Copy (GEM MINING #16 - Better paid content marketing copy with Sandra Wu, Content Marketing Lead at Blinkist) for even more information on this (GrowthGems.co membership needed)


Did a playable geared at putting more time into the education before people hit the App Store.


[💎@23:45] On the App Store there is a bottleneck: everybody sees the same thing no matter which USP you advertise. Via web you can create a more coherent journey so people understand why they are going to see the App Store. It created a virtuous cycle (better install conversion, better place in the auction, ASO, etc.).


This was done in parallel with app campaigns (not all spend went to campaigns going through a landing page).


Playables are another version of this: people want to see more. Same with video on the store: it's about showing more before people download the app.


Retargeting viewers of inspirational videos with direct response

8fit created "inspirational" videos (like a big brand would).

[💎@26:15] Even though "inspirational" videos did not perform well on Facebook, what did work well was to retarget the video viewers with direct response ads and a more direct messaging.


They also did this with influencer videos, where they retargeted a specific influencer's audience with DR ads.


Apply learnings in other businesses

Sometimes it works, but sometimes it fails. So you have no certainty. It lessens the value of learning but hightens the value of consistently trying new things.


Technology Thomas is excited to test

Very excited about streamlining the creative process: how to produce more creatives but without putting too much resources before you know what works, move variants, etc. Here is the post (I think) by Eric Seufert that Thomas mentions.


A company called Network (used to have a very successful mobile game) developed a technology to deploy a large quantity of creatives and act on it fast (technology fast) to select winners a lot faster.


Doing this yourself would not work if you're spending 1 million / day, not really with small budgets. Smaller players could automate their testing much more.

[💎@31:00] Turns out Google has already developed internally predictive testing and can predict by 70% if a creative is going to outperform the current one or not before they spend money.


Today is about automation, maybe tomorrow is about prediction.


Figuring out which creatives work best

It can take many months to really understand the value of a specific creative.


[💎@32:45] When you find a creative that really performs on the top of the funnel (clicks, installs, etc.) it is extremely fast to see it. The algorithm makes most of the delivery go to it, you can scale your campaigns, etc. It is obvious. But it doesn't mean that it's the best for the long term because the cohort needs to mature.


Humans can balance quick data response with long-term maturing cohorts, brand consideration, what's happening in the space, etc.


For complex products you need to understand the full user journey, and the full journey take months in many case.


[💎@35:30] When using AEO and VO it's about which event you're feeding the algorithm/machine. Maybe you don't optimize for sign ups, or free trial, but for "free trial + this event completed within 24 hours".


The notes from this resource are only available to premium members.
↘ At this point, you know what to do ↙
Upgrade Your Plan

Where to start if you're building out a team for UA

Back then, people with experience were in gaming and it was hard to attract them. So hired people based on mindset (even if they were in the web) rather than experience and not only UA people.


Biggest takeaway for 8fit

They understood very only on that they needed a solid data stack and accessibility to data. They built they own data stack, scalable and custom to what they needed.

[💎@06:00] Investing in data is usually an excellent choice, regardless if your build or buy (there are more solutions nowadays).


Goal of the data stack

Subscription acquisition flow is not that simple: start a free trial or not, do they convert, do they refund, modelize the renewals, etc.


  • Pure analytics: relied on tooling (Amplitude)
  • Revenue modelization: had to do everything on their own, because the data Apple shares is a bit messy and hard to reconcile.


[💎@08:00] The big focus of the data stack at 8fit was how to predict revenue the earliest possible and in the most accurate way in order to know what to re)inject in acquisition.

Another important part was uniting all networks and costs to have all conversions in one place: every single step of the funnel from every single source and being able to segment by anything (OS, gender, activity, etc.). There are now new products in the market for that.


Challenges of acquisition for subscription apps

Challenge #1

Most subscriptions use "start a free trial" as an event for optimization (Facebook, UAC).

[💎@09:23] If you focus too much on people converting to free trial you might attract more people who don't convert after. For example younger audiences are a lot more aware about how to cancel a free trial so you might get good CPAs but a cohort that converts poorly.


In short: the algorithms from FB/UAC really deliver what you ask for.


Challenge #2

On the user side there is only one subscription to buy but on the developer side you have many (can be 100+): iOS and Android, 7 or 30 day trial, prices different by country, super premium subscription. This increases complexity as well.


For 8fit it was in the company mindset to constantly test pricing: introductory offers (free trials), removing the free trial, etc.


[💎@12:12] Constant testing on offers and pricing is the only way to survive for subscription apps. What you might have learned 3 years ago might be different today: audience you are approaching or overall market could be different, there might be subscription fatigue, etc.


How to differentiate yourself when doing performance marketing?

For 8fit, they came from the nutrition angle so that USP was doing good on conversion. That's therefore where a lot of the UA efforts were placed, as well as: "it's useless to go train everyday if you eat like shit".


8fit was against the kind of "quick fix" message and were into long-term healthy choice in life, education, etc.

Focusing on the long-term and specific positioning might not bring some of the quick wins but it helps building over time.


Calling out the free trial vs. positioning in advertising?

Thomas had a couple of cohorts showing that pushing free stuff was not the right choice in the long term. He banned the word "free" from the communication of the company besides inside the app: no "free download", no "free trial".


Replaced the attractiveness of the free trial with a push on personalization. So that when people hit the paywall they have a reason to subscribe: the product is built specifically for them.


Educating an audience

When pushing 2 messages at the same time, it would not work. Example:

  • Fitness + nutrition
  • Personalization + trial
  • Etc.

→ Had to deliver 1 message only. Video helped.


[💎@21:20] Something that worked is putting a longer funnel before the install. Example: sending people from Outbrain/Taboola to a landing page (article form) where 8fit had a lot more space to tell people what makes the app different.


Because of the drop off on the landing page the install would be more expensive but there was higher conversion, retention and renewal were compensating this effect. Long term gains were everywhere: users subscribed to push, were more receptive to promos, engaged on social, etc.


👉 Check [RESOURCE #16] Better Paid Content Marketing Copy (GEM MINING #16 - Better paid content marketing copy with Sandra Wu, Content Marketing Lead at Blinkist) for even more information on this (GrowthGems.co membership needed)


Did a playable geared at putting more time into the education before people hit the App Store.


[💎@23:45] On the App Store there is a bottleneck: everybody sees the same thing no matter which USP you advertise. Via web you can create a more coherent journey so people understand why they are going to see the App Store. It created a virtuous cycle (better install conversion, better place in the auction, ASO, etc.).


This was done in parallel with app campaigns (not all spend went to campaigns going through a landing page).


Playables are another version of this: people want to see more. Same with video on the store: it's about showing more before people download the app.


Retargeting viewers of inspirational videos with direct response

8fit created "inspirational" videos (like a big brand would).

[💎@26:15] Even though "inspirational" videos did not perform well on Facebook, what did work well was to retarget the video viewers with direct response ads and a more direct messaging.


They also did this with influencer videos, where they retargeted a specific influencer's audience with DR ads.


Apply learnings in other businesses

Sometimes it works, but sometimes it fails. So you have no certainty. It lessens the value of learning but hightens the value of consistently trying new things.


Technology Thomas is excited to test

Very excited about streamlining the creative process: how to produce more creatives but without putting too much resources before you know what works, move variants, etc. Here is the post (I think) by Eric Seufert that Thomas mentions.


A company called Network (used to have a very successful mobile game) developed a technology to deploy a large quantity of creatives and act on it fast (technology fast) to select winners a lot faster.


Doing this yourself would not work if you're spending 1 million / day, not really with small budgets. Smaller players could automate their testing much more.

[💎@31:00] Turns out Google has already developed internally predictive testing and can predict by 70% if a creative is going to outperform the current one or not before they spend money.


Today is about automation, maybe tomorrow is about prediction.


Figuring out which creatives work best

It can take many months to really understand the value of a specific creative.


[💎@32:45] When you find a creative that really performs on the top of the funnel (clicks, installs, etc.) it is extremely fast to see it. The algorithm makes most of the delivery go to it, you can scale your campaigns, etc. It is obvious. But it doesn't mean that it's the best for the long term because the cohort needs to mature.


Humans can balance quick data response with long-term maturing cohorts, brand consideration, what's happening in the space, etc.


For complex products you need to understand the full user journey, and the full journey take months in many case.


[💎@35:30] When using AEO and VO it's about which event you're feeding the algorithm/machine. Maybe you don't optimize for sign ups, or free trial, but for "free trial + this event completed within 24 hours".


The notes from this resource are only available to premium members.

The detailed notes taken for a resource are an easy way to see the gems in context to get a better understanding. They also include any relevant visuals from the source.
↘ At this point, you know what to do ↙
Request Access

Where to start if you're building out a team for UA

Back then, people with experience were in gaming and it was hard to attract them. So hired people based on mindset (even if they were in the web) rather than experience and not only UA people.


Biggest takeaway for 8fit

They understood very only on that they needed a solid data stack and accessibility to data. They built they own data stack, scalable and custom to what they needed.

[💎@06:00] Investing in data is usually an excellent choice, regardless if your build or buy (there are more solutions nowadays).


Goal of the data stack

Subscription acquisition flow is not that simple: start a free trial or not, do they convert, do they refund, modelize the renewals, etc.


  • Pure analytics: relied on tooling (Amplitude)
  • Revenue modelization: had to do everything on their own, because the data Apple shares is a bit messy and hard to reconcile.


[💎@08:00] The big focus of the data stack at 8fit was how to predict revenue the earliest possible and in the most accurate way in order to know what to re)inject in acquisition.

Another important part was uniting all networks and costs to have all conversions in one place: every single step of the funnel from every single source and being able to segment by anything (OS, gender, activity, etc.). There are now new products in the market for that.


Challenges of acquisition for subscription apps

Challenge #1

Most subscriptions use "start a free trial" as an event for optimization (Facebook, UAC).

[💎@09:23] If you focus too much on people converting to free trial you might attract more people who don't convert after. For example younger audiences are a lot more aware about how to cancel a free trial so you might get good CPAs but a cohort that converts poorly.


In short: the algorithms from FB/UAC really deliver what you ask for.


Challenge #2

On the user side there is only one subscription to buy but on the developer side you have many (can be 100+): iOS and Android, 7 or 30 day trial, prices different by country, super premium subscription. This increases complexity as well.


For 8fit it was in the company mindset to constantly test pricing: introductory offers (free trials), removing the free trial, etc.


[💎@12:12] Constant testing on offers and pricing is the only way to survive for subscription apps. What you might have learned 3 years ago might be different today: audience you are approaching or overall market could be different, there might be subscription fatigue, etc.


How to differentiate yourself when doing performance marketing?

For 8fit, they came from the nutrition angle so that USP was doing good on conversion. That's therefore where a lot of the UA efforts were placed, as well as: "it's useless to go train everyday if you eat like shit".


8fit was against the kind of "quick fix" message and were into long-term healthy choice in life, education, etc.

Focusing on the long-term and specific positioning might not bring some of the quick wins but it helps building over time.


Calling out the free trial vs. positioning in advertising?

Thomas had a couple of cohorts showing that pushing free stuff was not the right choice in the long term. He banned the word "free" from the communication of the company besides inside the app: no "free download", no "free trial".


Replaced the attractiveness of the free trial with a push on personalization. So that when people hit the paywall they have a reason to subscribe: the product is built specifically for them.


Educating an audience

When pushing 2 messages at the same time, it would not work. Example:

  • Fitness + nutrition
  • Personalization + trial
  • Etc.

→ Had to deliver 1 message only. Video helped.


[💎@21:20] Something that worked is putting a longer funnel before the install. Example: sending people from Outbrain/Taboola to a landing page (article form) where 8fit had a lot more space to tell people what makes the app different.


Because of the drop off on the landing page the install would be more expensive but there was higher conversion, retention and renewal were compensating this effect. Long term gains were everywhere: users subscribed to push, were more receptive to promos, engaged on social, etc.


👉 Check [RESOURCE #16] Better Paid Content Marketing Copy (GEM MINING #16 - Better paid content marketing copy with Sandra Wu, Content Marketing Lead at Blinkist) for even more information on this (GrowthGems.co membership needed)


Did a playable geared at putting more time into the education before people hit the App Store.


[💎@23:45] On the App Store there is a bottleneck: everybody sees the same thing no matter which USP you advertise. Via web you can create a more coherent journey so people understand why they are going to see the App Store. It created a virtuous cycle (better install conversion, better place in the auction, ASO, etc.).


This was done in parallel with app campaigns (not all spend went to campaigns going through a landing page).


Playables are another version of this: people want to see more. Same with video on the store: it's about showing more before people download the app.


Retargeting viewers of inspirational videos with direct response

8fit created "inspirational" videos (like a big brand would).

[💎@26:15] Even though "inspirational" videos did not perform well on Facebook, what did work well was to retarget the video viewers with direct response ads and a more direct messaging.


They also did this with influencer videos, where they retargeted a specific influencer's audience with DR ads.


Apply learnings in other businesses

Sometimes it works, but sometimes it fails. So you have no certainty. It lessens the value of learning but hightens the value of consistently trying new things.


Technology Thomas is excited to test

Very excited about streamlining the creative process: how to produce more creatives but without putting too much resources before you know what works, move variants, etc. Here is the post (I think) by Eric Seufert that Thomas mentions.


A company called Network (used to have a very successful mobile game) developed a technology to deploy a large quantity of creatives and act on it fast (technology fast) to select winners a lot faster.


Doing this yourself would not work if you're spending 1 million / day, not really with small budgets. Smaller players could automate their testing much more.

[💎@31:00] Turns out Google has already developed internally predictive testing and can predict by 70% if a creative is going to outperform the current one or not before they spend money.


Today is about automation, maybe tomorrow is about prediction.


Figuring out which creatives work best

It can take many months to really understand the value of a specific creative.


[💎@32:45] When you find a creative that really performs on the top of the funnel (clicks, installs, etc.) it is extremely fast to see it. The algorithm makes most of the delivery go to it, you can scale your campaigns, etc. It is obvious. But it doesn't mean that it's the best for the long term because the cohort needs to mature.


Humans can balance quick data response with long-term maturing cohorts, brand consideration, what's happening in the space, etc.


For complex products you need to understand the full user journey, and the full journey take months in many case.


[💎@35:30] When using AEO and VO it's about which event you're feeding the algorithm/machine. Maybe you don't optimize for sign ups, or free trial, but for "free trial + this event completed within 24 hours".