Andy Carvell (Partner at Phiture) shares insights from his experience at SoundCloud and talks about retention techniques and what to focus on at different stages of growth.
Lifecycle marketing starts with understanding the user journey and understanding users better than you currently do. This starts with asking questions to users, not by diving into analytics and looking at funnels.
Product is the biggest lever for engagement. But CRM is a great way to circumvent engineering backlog and iterate quickly. It also has built-in measurement and segmentation.
What really mattered at Elevate when A/B testing was the first step, the early experience. That’s where there was the most reach compared to lower-funnel tests.
At SoundCloud, Andy would run the boldest push notification experiments in Pakistan. He could then come back with data and get more leeway for tests in other places.
A challenge with early stage apps is that they have unrealistic expectations and think they’re ready for hyper growth in 99.9% of cases, they’re not ready). Instead, what an early stage app needs to work on is product-market fit. They need to iterate on the product and the marketing.
Focus on the product first. At Elevate, they started with monitoring Testflight cohorts (from a curated list) in order to understand engagement and conversion. They iterated on the product until they felt that the metrics were good (sign ups, activation), then they launched.
At one point, the impact of new features becomes marginal and can even be detrimental if things become too crowded.
In the growth phase, you need to upgrade your analytics to start tracking more things in order to get a deeper understanding of your engagement and the performance of your acquisition channels.
The growth phase is also where you want to experiment with virality. You can’t control virality but you maybe can start adding in some virality elements (share content, referral system, etc.). Build this in a rather cheap way first and if you see some traction, then you can optimize.
The growth stage is about exploring. You don’t know where the big levers for scale are yet, and it’s about discovering them. It might be virality, a market where you aren’t localized yet, etc.
When you’re at a huge scale (Headspace, VSCO, etc.), even incremental gains can be meaningful and especially on monetization. So at this stage you want to f*ocus more on monetization and subscription monetization*: increasing your average subscription lifetime by a month or increasing your conversion rate by .1% might be meaningful.
Once at scale, you should have an even bigger focus on retention: it’s a slow-moving metrics but brings compound growth and monetization gains. Also work on conversion optimization, start segmenting further your user base (since you can now have significant cohorts) and lean into AI and personalization.
When you’re early, you need to take big swings that bring big results. You don’t need sophisticated measurement to see the result or sophisticated A/B testing.
Lifecycle marketing starts with understanding the user journey and understanding users better than you currently do. This starts with asking questions to users, not by diving into analytics and looking at funnels.
Product is the biggest lever for engagement. But CRM is a great way to circumvent engineering backlog and iterate quickly. It also has built-in measurement and segmentation.
What really mattered at Elevate when A/B testing was the first step, the early experience. That’s where there was the most reach compared to lower-funnel tests.
At SoundCloud, Andy would run the boldest push notification experiments in Pakistan. He could then come back with data and get more leeway for tests in other places.
A challenge with early stage apps is that they have unrealistic expectations and think they’re ready for hyper growth in 99.9% of cases, they’re not ready). Instead, what an early stage app needs to work on is product-market fit. They need to iterate on the product and the marketing.
Focus on the product first. At Elevate, they started with monitoring Testflight cohorts (from a curated list) in order to understand engagement and conversion. They iterated on the product until they felt that the metrics were good (sign ups, activation), then they launched.
At one point, the impact of new features becomes marginal and can even be detrimental if things become too crowded.
In the growth phase, you need to upgrade your analytics to start tracking more things in order to get a deeper understanding of your engagement and the performance of your acquisition channels.
The growth phase is also where you want to experiment with virality. You can’t control virality but you maybe can start adding in some virality elements (share content, referral system, etc.). Build this in a rather cheap way first and if you see some traction, then you can optimize.
The growth stage is about exploring. You don’t know where the big levers for scale are yet, and it’s about discovering them. It might be virality, a market where you aren’t localized yet, etc.
When you’re at a huge scale (Headspace, VSCO, etc.), even incremental gains can be meaningful and especially on monetization. So at this stage you want to f*ocus more on monetization and subscription monetization*: increasing your average subscription lifetime by a month or increasing your conversion rate by .1% might be meaningful.
Once at scale, you should have an even bigger focus on retention: it’s a slow-moving metrics but brings compound growth and monetization gains. Also work on conversion optimization, start segmenting further your user base (since you can now have significant cohorts) and lean into AI and personalization.
When you’re early, you need to take big swings that bring big results. You don’t need sophisticated measurement to see the result or sophisticated A/B testing.
Lifecycle marketing starts with understanding the user journey and understanding users better than you currently do. This starts with asking questions to users, not by diving into analytics and looking at funnels.
Product is the biggest lever for engagement. But CRM is a great way to circumvent engineering backlog and iterate quickly. It also has built-in measurement and segmentation.
What really mattered at Elevate when A/B testing was the first step, the early experience. That’s where there was the most reach compared to lower-funnel tests.
At SoundCloud, Andy would run the boldest push notification experiments in Pakistan. He could then come back with data and get more leeway for tests in other places.
A challenge with early stage apps is that they have unrealistic expectations and think they’re ready for hyper growth in 99.9% of cases, they’re not ready). Instead, what an early stage app needs to work on is product-market fit. They need to iterate on the product and the marketing.
Focus on the product first. At Elevate, they started with monitoring Testflight cohorts (from a curated list) in order to understand engagement and conversion. They iterated on the product until they felt that the metrics were good (sign ups, activation), then they launched.
At one point, the impact of new features becomes marginal and can even be detrimental if things become too crowded.
In the growth phase, you need to upgrade your analytics to start tracking more things in order to get a deeper understanding of your engagement and the performance of your acquisition channels.
The growth phase is also where you want to experiment with virality. You can’t control virality but you maybe can start adding in some virality elements (share content, referral system, etc.). Build this in a rather cheap way first and if you see some traction, then you can optimize.
The growth stage is about exploring. You don’t know where the big levers for scale are yet, and it’s about discovering them. It might be virality, a market where you aren’t localized yet, etc.
When you’re at a huge scale (Headspace, VSCO, etc.), even incremental gains can be meaningful and especially on monetization. So at this stage you want to f*ocus more on monetization and subscription monetization*: increasing your average subscription lifetime by a month or increasing your conversion rate by .1% might be meaningful.
Once at scale, you should have an even bigger focus on retention: it’s a slow-moving metrics but brings compound growth and monetization gains. Also work on conversion optimization, start segmenting further your user base (since you can now have significant cohorts) and lean into AI and personalization.
When you’re early, you need to take big swings that bring big results. You don’t need sophisticated measurement to see the result or sophisticated A/B testing.
Notes for this resource are currently being transferred and will be available soon.
Andy was at SoundCloud for 4.5 years, where he built a team focused on retention and growth.
Everybody struggles with what happens after the subscription.
[💎@05:40] Lifecycle marketing starts with understanding the user journey and understanding users better than you currently do. This starts with asking questions to users, not by diving into analytics and looking at funnels.
When we have a lot of data available, we tend to dive directly into the analytics.
The App Store page is a great opportunity to sell the product, then the promise needs to be fulfilled within the first session.
[💎@07:45] Product is the biggest lever for engagement. But CRM is a great way to circumvent engineering backlog and iterate quickly. It also has built-in measurement and segmentation.
When talking about CRM in the mobile app context, Andy refers to platforms like Braze, Leanplum, Iterable, etc. They have rudimentary analytics and you can also pair this with your product analytics.
You can make messages look super native.
[💎@11:18] What really mattered at Elevate when A/B testing was the first step, the early experience. That’s where there was the most reach compared to lower-funnel tests.
The best channel depends on the audience of the app. SMS is still a good channel for some apps (e.g. GoodRX, with an older audience).
In-app messages only work when the user is already in the app.
We need to keep in mind that app developers are not the target audience for the vast majority of consumer apps. Not everybody hates push notifications.
[💎@21:05] At SoundCloud, Andy would run the boldest push notification experiments in Pakistan. He could then come back with data and get more leeway for tests in other places.
SoundCloud was sending around 5M push notifications a month.
The impact that you can drive with notifications is Reach x Relevance x Frequency. What they learned from their time at SoundCloud was that not all notifications are equal, and the really killer ones that are going to really supercharge your business have high reach, high relevance and high frequency.
Different stacks for different stages. You do not need to build everything from the mobile growth stack.
[💎@27:35] A challenge with early stage apps is that they have unrealistic expectations and think they’re ready for hyper growth in 99.9% of cases, they’re not ready). Instead, what an early stage app needs to work on is product-market fit. They need to iterate on the product and the marketing.
You want to get to retention curves that flatten out.
[💎@29:22] Focus on the product first. At Elevate, they started with monitoring Testflight cohorts (from a curated list) in order to understand engagement and conversion. They iterated on the product until they felt that the metrics were good (sign ups, activation), then they launched.
Don’t build a ton of internal tools at the beginning. You can use find hacky ways to do things: email surveys, zappier, etc.
[💎@32:00] At one point, the impact of new features becomes marginal and can even be detrimental if things become too crowded.
[💎@32:44] In the growth phase, you need to upgrade your analytics to start tracking more things in order to get a deeper understanding of your engagement and the performance of your acquisition channels.
[💎@33:22] The growth phase is also where you want to experiment with virality. You can’t control virality but you maybe can start adding in some virality elements (share content, referral system, etc.). Build this in a rather cheap way first and if you see some traction, then you can optimize.
Continue to iterate on the product, focus more on ASO and start thinking about international. A lot of companies focus on the market they know best but it’s good to think global, localize your app store pages, etc.
[💎@35:50] The growth stage is about exploring. You don’t know where the big levers for scale are yet, and it’s about discovering them. It might be virality, a market where you aren’t localized yet, etc.
You can also start building the foundations of CRM.
[💎@38:20] When you’re at a huge scale (Headspace, VSCO, etc.), even incremental gains can be meaningful and especially on monetization. So at this stage you want to focus more on monetization and subscription monetization: increasing your average subscription lifetime by a month or increasing your conversion rate by .1% might be meaningful.
It’s time to start upgrading your stack if you are still working with rudimentary tools.
[💎@39:40] Once at scale, you should have an even bigger focus on retention: it’s a slow-moving metrics but brings compound growth and monetization gains. Also work on conversion optimization, start segmenting further your user base (since you can now have significant cohorts) and lean into AI and personalization.
This is when it makes more sense to do A/B testing.
The next 10x is always going to be different.
[💎@41:42] When you’re early, you need to take big swings that bring big results. You don’t need sophisticated measurement to see the result or sophisticated A/B testing.
Phiture is looking for a Subscription Optimization Lead.
Andy was at SoundCloud for 4.5 years, where he built a team focused on retention and growth.
Everybody struggles with what happens after the subscription.
[💎@05:40] Lifecycle marketing starts with understanding the user journey and understanding users better than you currently do. This starts with asking questions to users, not by diving into analytics and looking at funnels.
When we have a lot of data available, we tend to dive directly into the analytics.
The App Store page is a great opportunity to sell the product, then the promise needs to be fulfilled within the first session.
[💎@07:45] Product is the biggest lever for engagement. But CRM is a great way to circumvent engineering backlog and iterate quickly. It also has built-in measurement and segmentation.
When talking about CRM in the mobile app context, Andy refers to platforms like Braze, Leanplum, Iterable, etc. They have rudimentary analytics and you can also pair this with your product analytics.
You can make messages look super native.
[💎@11:18] What really mattered at Elevate when A/B testing was the first step, the early experience. That’s where there was the most reach compared to lower-funnel tests.
The best channel depends on the audience of the app. SMS is still a good channel for some apps (e.g. GoodRX, with an older audience).
In-app messages only work when the user is already in the app.
We need to keep in mind that app developers are not the target audience for the vast majority of consumer apps. Not everybody hates push notifications.
[💎@21:05] At SoundCloud, Andy would run the boldest push notification experiments in Pakistan. He could then come back with data and get more leeway for tests in other places.
SoundCloud was sending around 5M push notifications a month.
The impact that you can drive with notifications is Reach x Relevance x Frequency. What they learned from their time at SoundCloud was that not all notifications are equal, and the really killer ones that are going to really supercharge your business have high reach, high relevance and high frequency.
Different stacks for different stages. You do not need to build everything from the mobile growth stack.
[💎@27:35] A challenge with early stage apps is that they have unrealistic expectations and think they’re ready for hyper growth in 99.9% of cases, they’re not ready). Instead, what an early stage app needs to work on is product-market fit. They need to iterate on the product and the marketing.
You want to get to retention curves that flatten out.
[💎@29:22] Focus on the product first. At Elevate, they started with monitoring Testflight cohorts (from a curated list) in order to understand engagement and conversion. They iterated on the product until they felt that the metrics were good (sign ups, activation), then they launched.
Don’t build a ton of internal tools at the beginning. You can use find hacky ways to do things: email surveys, zappier, etc.
[💎@32:00] At one point, the impact of new features becomes marginal and can even be detrimental if things become too crowded.
[💎@32:44] In the growth phase, you need to upgrade your analytics to start tracking more things in order to get a deeper understanding of your engagement and the performance of your acquisition channels.
[💎@33:22] The growth phase is also where you want to experiment with virality. You can’t control virality but you maybe can start adding in some virality elements (share content, referral system, etc.). Build this in a rather cheap way first and if you see some traction, then you can optimize.
Continue to iterate on the product, focus more on ASO and start thinking about international. A lot of companies focus on the market they know best but it’s good to think global, localize your app store pages, etc.
[💎@35:50] The growth stage is about exploring. You don’t know where the big levers for scale are yet, and it’s about discovering them. It might be virality, a market where you aren’t localized yet, etc.
You can also start building the foundations of CRM.
[💎@38:20] When you’re at a huge scale (Headspace, VSCO, etc.), even incremental gains can be meaningful and especially on monetization. So at this stage you want to focus more on monetization and subscription monetization: increasing your average subscription lifetime by a month or increasing your conversion rate by .1% might be meaningful.
It’s time to start upgrading your stack if you are still working with rudimentary tools.
[💎@39:40] Once at scale, you should have an even bigger focus on retention: it’s a slow-moving metrics but brings compound growth and monetization gains. Also work on conversion optimization, start segmenting further your user base (since you can now have significant cohorts) and lean into AI and personalization.
This is when it makes more sense to do A/B testing.
The next 10x is always going to be different.
[💎@41:42] When you’re early, you need to take big swings that bring big results. You don’t need sophisticated measurement to see the result or sophisticated A/B testing.
Phiture is looking for a Subscription Optimization Lead.
Andy was at SoundCloud for 4.5 years, where he built a team focused on retention and growth.
Everybody struggles with what happens after the subscription.
[💎@05:40] Lifecycle marketing starts with understanding the user journey and understanding users better than you currently do. This starts with asking questions to users, not by diving into analytics and looking at funnels.
When we have a lot of data available, we tend to dive directly into the analytics.
The App Store page is a great opportunity to sell the product, then the promise needs to be fulfilled within the first session.
[💎@07:45] Product is the biggest lever for engagement. But CRM is a great way to circumvent engineering backlog and iterate quickly. It also has built-in measurement and segmentation.
When talking about CRM in the mobile app context, Andy refers to platforms like Braze, Leanplum, Iterable, etc. They have rudimentary analytics and you can also pair this with your product analytics.
You can make messages look super native.
[💎@11:18] What really mattered at Elevate when A/B testing was the first step, the early experience. That’s where there was the most reach compared to lower-funnel tests.
The best channel depends on the audience of the app. SMS is still a good channel for some apps (e.g. GoodRX, with an older audience).
In-app messages only work when the user is already in the app.
We need to keep in mind that app developers are not the target audience for the vast majority of consumer apps. Not everybody hates push notifications.
[💎@21:05] At SoundCloud, Andy would run the boldest push notification experiments in Pakistan. He could then come back with data and get more leeway for tests in other places.
SoundCloud was sending around 5M push notifications a month.
The impact that you can drive with notifications is Reach x Relevance x Frequency. What they learned from their time at SoundCloud was that not all notifications are equal, and the really killer ones that are going to really supercharge your business have high reach, high relevance and high frequency.
Different stacks for different stages. You do not need to build everything from the mobile growth stack.
[💎@27:35] A challenge with early stage apps is that they have unrealistic expectations and think they’re ready for hyper growth in 99.9% of cases, they’re not ready). Instead, what an early stage app needs to work on is product-market fit. They need to iterate on the product and the marketing.
You want to get to retention curves that flatten out.
[💎@29:22] Focus on the product first. At Elevate, they started with monitoring Testflight cohorts (from a curated list) in order to understand engagement and conversion. They iterated on the product until they felt that the metrics were good (sign ups, activation), then they launched.
Don’t build a ton of internal tools at the beginning. You can use find hacky ways to do things: email surveys, zappier, etc.
[💎@32:00] At one point, the impact of new features becomes marginal and can even be detrimental if things become too crowded.
[💎@32:44] In the growth phase, you need to upgrade your analytics to start tracking more things in order to get a deeper understanding of your engagement and the performance of your acquisition channels.
[💎@33:22] The growth phase is also where you want to experiment with virality. You can’t control virality but you maybe can start adding in some virality elements (share content, referral system, etc.). Build this in a rather cheap way first and if you see some traction, then you can optimize.
Continue to iterate on the product, focus more on ASO and start thinking about international. A lot of companies focus on the market they know best but it’s good to think global, localize your app store pages, etc.
[💎@35:50] The growth stage is about exploring. You don’t know where the big levers for scale are yet, and it’s about discovering them. It might be virality, a market where you aren’t localized yet, etc.
You can also start building the foundations of CRM.
[💎@38:20] When you’re at a huge scale (Headspace, VSCO, etc.), even incremental gains can be meaningful and especially on monetization. So at this stage you want to focus more on monetization and subscription monetization: increasing your average subscription lifetime by a month or increasing your conversion rate by .1% might be meaningful.
It’s time to start upgrading your stack if you are still working with rudimentary tools.
[💎@39:40] Once at scale, you should have an even bigger focus on retention: it’s a slow-moving metrics but brings compound growth and monetization gains. Also work on conversion optimization, start segmenting further your user base (since you can now have significant cohorts) and lean into AI and personalization.
This is when it makes more sense to do A/B testing.
The next 10x is always going to be different.
[💎@41:42] When you’re early, you need to take big swings that bring big results. You don’t need sophisticated measurement to see the result or sophisticated A/B testing.
Phiture is looking for a Subscription Optimization Lead.
