Eric Seufert (Mobile Dev Memo, Analyst and Strategy Consultant at Heracles) talks with Thomas Petit (Growth Consultant) about all things mobile subscriptions: product strategy, how to price test and the future of the subscription model.
Even with mobile web clicks not necessarily cheaper and more drop-offs on the web funnel, ads leading to a web onboarding flow can still be worth it: it can allow you to broaden the audience beyond the reach of app install campaigns (some networks don't target LAT people, some people don't click on app install ads and end up seeing less of them).
A long web onboarding flow is actually more challenging because once people download an app you can get a great completion rate even with longer in-app onboarding.
With a web onboarding you only pay 3% to Stripe instead of 30% to the app store but the added friction might bring your conversion rate down 27%, making things even.
The real benefit of a web onboarding flow is that people become your direct customers. It is a massive benefit to have that direct relationship, as it brings flexibility when it comes to cancellations, refunds, renewals, etc.
Having a direct relationship with the customer allows you to do things like partial refunds: instead of cancelling a yearly subscription directly for the customer, you can offer them to refund half of the subscription while keeping their access on. It's a win-win.
With the IDFA deprecation coming up, a lot of people are trying to move to web. But it's a different mindset and very few people have the skills to understand the app store ecosystem AND the web ecosystem at the same time.
You are losing a lot of users at that App Store page step. You can't control it, and you think you could do better with landing pages. But still the bounce rate can be higher on the web!
A lot of people try to increase just one metric and think it could trickle down all the way. But early drop-offs (like on the App Store page) can filter out low-intent people that would never convert anyway.
The growth hacking mindset when it comes to funnels is not the right one. You can endlessly A/B test an onboarding flow yet still have your conversion rate go down, or fail to increase revenue.
You can not just focus on A/B testing specific parts of the funnel because these local optimizations often disregard the impact on down-funnel metrics. Another critical aspect is the traffic composition.
Even if you consider both deeper-funnel events and traffic composition, be careful about over optimizing for revenue. Short term gains can end up being long term losses and prevent you from really scaling. Example: great payback period but high churn.
Put secondary metrics ("tradeoff metrics") to every test you're doing. It's more complex but leads you to think bigger picture. Example: retention when optimizing for revenue, email sales when adding SSO, etc.
Eric has seen some games where subscription players (often high LTV ones) end up up paying more on in-app purchases after subscribing. Players get a subscription to be more competitive, but as more players subscribe as well they still need to buy IAPs.
A lot of games are not doing subscriptions because they have already found a way to get their players to spend as much as possible.
The beauty of the subscription model is the predictability/security but a major flaw is that you have a high floor and a low ceiling. - High floor: high barrier to entry because some people would pay for a daily/weekly pass but not a subscription so you're leaving money on the table. - Low ceiling: nothing more to "extract" from a user that has already paid.
You can even sell non digital goods to your subscribed users. This is what the Sweat app is doing with a shop. The community wants branded goods. Another example with 8fit: super expensive subscription where they could deliver fruits & vegetables through Instacart.
Think about how you can fight against the "low-ceiling" of your subscriptions. What can you upsell (another product, another subscription, a bundle with another app through a partnership, etc.)? Subscriptions don't have to be a binary thing.
What's the probability that the price you set for your subscription is the max they're willing to spend? If you're already willing to buy the subscription, you'd probably be willing to pay more.
If you're an app like Calm or Tinder, a +2% gain is huge and small price optimizations end up being worth doing (especially because you have the data to do it).
If you find 5-10% increments in monetization, it might make a huge difference in the volume you can unlock through UA (passing a bid threshold can unlock acquisition volume). Example: an app that had already iterated on its paywall finds +15-17% increase in revenue that unlocked massive volume.
Think about the state of mind in which the user is when he reaches the paywall. Don't think about cosmetic changes (bullet points, buttons, etc.). Example:show 30% discount if they skip the paywall, then if they skip again then email with 50% discount, etc. But you can go even further: daily pass when they hit the paywall the second time, free month if they invite somebody when they hit the paywall for the third time, etc.
The best experiments are the ones that work on the onboarding screens before the paywall. Anecdote: an app Thomas works with had tested having a longer loading screen ("preparing your plan"), which increased intent because it hinted at personalization (even though there was none).
If you make the skip button more obvious on your paywall (instead of making it discreet) you churn people less, you get less dissatisfied people and you get an opportunity to monetize them another way.
Instead of trying to optimize the subscription bundle at the highest price possible right out of the gate:
- Prove out that initial payers are ready to pay for a subscription
- Collect more data
- Upsell other things they could find value in and purchase (digital goods, different kinds of content, etc.)
Always raise your price: the highest price always wins. Only a small percentage of people are willing to pay for in-app subscription, and those people are not really price sensitive. Example: speechify at $150, tests Thomas did with first $50 vs. $100, etc.
Price elasticity is typically very low, but don't push the fact that conversion doesn't change that much when you increase prices. It might be best to leave some money on the table (more users who pay less) if it means that you get more data, your community grows, etc.
There are beneficial effects of having a bigger userbase: word of mouth, viral effect, etc. but customer churn also becomes less meaningful.
In crowded spaces, people are not going to compare the apps' content that much (short term). But the power of brand is what comes in when you reach a degree of maturity, with often 1 or 2 players. Example: Peloton, Headspace, Calm. True in language learning too.
The problem is that people over experiment but don't test radical enough things. The less data you have and the more you haven't been down the road of experimentation, the more radical you should be.
Even with mobile web clicks not necessarily cheaper and more drop-offs on the web funnel, ads leading to a web onboarding flow can still be worth it: it can allow you to broaden the audience beyond the reach of app install campaigns (some networks don't target LAT people, some people don't click on app install ads and end up seeing less of them).
A long web onboarding flow is actually more challenging because once people download an app you can get a great completion rate even with longer in-app onboarding.
With a web onboarding you only pay 3% to Stripe instead of 30% to the app store but the added friction might bring your conversion rate down 27%, making things even.
The real benefit of a web onboarding flow is that people become your direct customers. It is a massive benefit to have that direct relationship, as it brings flexibility when it comes to cancellations, refunds, renewals, etc.
Having a direct relationship with the customer allows you to do things like partial refunds: instead of cancelling a yearly subscription directly for the customer, you can offer them to refund half of the subscription while keeping their access on. It's a win-win.
With the IDFA deprecation coming up, a lot of people are trying to move to web. But it's a different mindset and very few people have the skills to understand the app store ecosystem AND the web ecosystem at the same time.
You are losing a lot of users at that App Store page step. You can't control it, and you think you could do better with landing pages. But still the bounce rate can be higher on the web!
A lot of people try to increase just one metric and think it could trickle down all the way. But early drop-offs (like on the App Store page) can filter out low-intent people that would never convert anyway.
The growth hacking mindset when it comes to funnels is not the right one. You can endlessly A/B test an onboarding flow yet still have your conversion rate go down, or fail to increase revenue.
You can not just focus on A/B testing specific parts of the funnel because these local optimizations often disregard the impact on down-funnel metrics. Another critical aspect is the traffic composition.
Even if you consider both deeper-funnel events and traffic composition, be careful about over optimizing for revenue. Short term gains can end up being long term losses and prevent you from really scaling. Example: great payback period but high churn.
Put secondary metrics ("tradeoff metrics") to every test you're doing. It's more complex but leads you to think bigger picture. Example: retention when optimizing for revenue, email sales when adding SSO, etc.
Eric has seen some games where subscription players (often high LTV ones) end up up paying more on in-app purchases after subscribing. Players get a subscription to be more competitive, but as more players subscribe as well they still need to buy IAPs.
A lot of games are not doing subscriptions because they have already found a way to get their players to spend as much as possible.
The beauty of the subscription model is the predictability/security but a major flaw is that you have a high floor and a low ceiling. - High floor: high barrier to entry because some people would pay for a daily/weekly pass but not a subscription so you're leaving money on the table. - Low ceiling: nothing more to "extract" from a user that has already paid.
You can even sell non digital goods to your subscribed users. This is what the Sweat app is doing with a shop. The community wants branded goods. Another example with 8fit: super expensive subscription where they could deliver fruits & vegetables through Instacart.
Think about how you can fight against the "low-ceiling" of your subscriptions. What can you upsell (another product, another subscription, a bundle with another app through a partnership, etc.)? Subscriptions don't have to be a binary thing.
What's the probability that the price you set for your subscription is the max they're willing to spend? If you're already willing to buy the subscription, you'd probably be willing to pay more.
If you're an app like Calm or Tinder, a +2% gain is huge and small price optimizations end up being worth doing (especially because you have the data to do it).
If you find 5-10% increments in monetization, it might make a huge difference in the volume you can unlock through UA (passing a bid threshold can unlock acquisition volume). Example: an app that had already iterated on its paywall finds +15-17% increase in revenue that unlocked massive volume.
Think about the state of mind in which the user is when he reaches the paywall. Don't think about cosmetic changes (bullet points, buttons, etc.). Example:show 30% discount if they skip the paywall, then if they skip again then email with 50% discount, etc. But you can go even further: daily pass when they hit the paywall the second time, free month if they invite somebody when they hit the paywall for the third time, etc.
The best experiments are the ones that work on the onboarding screens before the paywall. Anecdote: an app Thomas works with had tested having a longer loading screen ("preparing your plan"), which increased intent because it hinted at personalization (even though there was none).
If you make the skip button more obvious on your paywall (instead of making it discreet) you churn people less, you get less dissatisfied people and you get an opportunity to monetize them another way.
Instead of trying to optimize the subscription bundle at the highest price possible right out of the gate:
- Prove out that initial payers are ready to pay for a subscription
- Collect more data
- Upsell other things they could find value in and purchase (digital goods, different kinds of content, etc.)
Always raise your price: the highest price always wins. Only a small percentage of people are willing to pay for in-app subscription, and those people are not really price sensitive. Example: speechify at $150, tests Thomas did with first $50 vs. $100, etc.
Price elasticity is typically very low, but don't push the fact that conversion doesn't change that much when you increase prices. It might be best to leave some money on the table (more users who pay less) if it means that you get more data, your community grows, etc.
There are beneficial effects of having a bigger userbase: word of mouth, viral effect, etc. but customer churn also becomes less meaningful.
In crowded spaces, people are not going to compare the apps' content that much (short term). But the power of brand is what comes in when you reach a degree of maturity, with often 1 or 2 players. Example: Peloton, Headspace, Calm. True in language learning too.
The problem is that people over experiment but don't test radical enough things. The less data you have and the more you haven't been down the road of experimentation, the more radical you should be.
Even with mobile web clicks not necessarily cheaper and more drop-offs on the web funnel, ads leading to a web onboarding flow can still be worth it: it can allow you to broaden the audience beyond the reach of app install campaigns (some networks don't target LAT people, some people don't click on app install ads and end up seeing less of them).
A long web onboarding flow is actually more challenging because once people download an app you can get a great completion rate even with longer in-app onboarding.
With a web onboarding you only pay 3% to Stripe instead of 30% to the app store but the added friction might bring your conversion rate down 27%, making things even.
The real benefit of a web onboarding flow is that people become your direct customers. It is a massive benefit to have that direct relationship, as it brings flexibility when it comes to cancellations, refunds, renewals, etc.
Having a direct relationship with the customer allows you to do things like partial refunds: instead of cancelling a yearly subscription directly for the customer, you can offer them to refund half of the subscription while keeping their access on. It's a win-win.
With the IDFA deprecation coming up, a lot of people are trying to move to web. But it's a different mindset and very few people have the skills to understand the app store ecosystem AND the web ecosystem at the same time.
You are losing a lot of users at that App Store page step. You can't control it, and you think you could do better with landing pages. But still the bounce rate can be higher on the web!
A lot of people try to increase just one metric and think it could trickle down all the way. But early drop-offs (like on the App Store page) can filter out low-intent people that would never convert anyway.
The growth hacking mindset when it comes to funnels is not the right one. You can endlessly A/B test an onboarding flow yet still have your conversion rate go down, or fail to increase revenue.
You can not just focus on A/B testing specific parts of the funnel because these local optimizations often disregard the impact on down-funnel metrics. Another critical aspect is the traffic composition.
Even if you consider both deeper-funnel events and traffic composition, be careful about over optimizing for revenue. Short term gains can end up being long term losses and prevent you from really scaling. Example: great payback period but high churn.
Put secondary metrics ("tradeoff metrics") to every test you're doing. It's more complex but leads you to think bigger picture. Example: retention when optimizing for revenue, email sales when adding SSO, etc.
Eric has seen some games where subscription players (often high LTV ones) end up up paying more on in-app purchases after subscribing. Players get a subscription to be more competitive, but as more players subscribe as well they still need to buy IAPs.
A lot of games are not doing subscriptions because they have already found a way to get their players to spend as much as possible.
The beauty of the subscription model is the predictability/security but a major flaw is that you have a high floor and a low ceiling. - High floor: high barrier to entry because some people would pay for a daily/weekly pass but not a subscription so you're leaving money on the table. - Low ceiling: nothing more to "extract" from a user that has already paid.
You can even sell non digital goods to your subscribed users. This is what the Sweat app is doing with a shop. The community wants branded goods. Another example with 8fit: super expensive subscription where they could deliver fruits & vegetables through Instacart.
Think about how you can fight against the "low-ceiling" of your subscriptions. What can you upsell (another product, another subscription, a bundle with another app through a partnership, etc.)? Subscriptions don't have to be a binary thing.
What's the probability that the price you set for your subscription is the max they're willing to spend? If you're already willing to buy the subscription, you'd probably be willing to pay more.
If you're an app like Calm or Tinder, a +2% gain is huge and small price optimizations end up being worth doing (especially because you have the data to do it).
If you find 5-10% increments in monetization, it might make a huge difference in the volume you can unlock through UA (passing a bid threshold can unlock acquisition volume). Example: an app that had already iterated on its paywall finds +15-17% increase in revenue that unlocked massive volume.
Think about the state of mind in which the user is when he reaches the paywall. Don't think about cosmetic changes (bullet points, buttons, etc.). Example:show 30% discount if they skip the paywall, then if they skip again then email with 50% discount, etc. But you can go even further: daily pass when they hit the paywall the second time, free month if they invite somebody when they hit the paywall for the third time, etc.
The best experiments are the ones that work on the onboarding screens before the paywall. Anecdote: an app Thomas works with had tested having a longer loading screen ("preparing your plan"), which increased intent because it hinted at personalization (even though there was none).
If you make the skip button more obvious on your paywall (instead of making it discreet) you churn people less, you get less dissatisfied people and you get an opportunity to monetize them another way.
Instead of trying to optimize the subscription bundle at the highest price possible right out of the gate:
- Prove out that initial payers are ready to pay for a subscription
- Collect more data
- Upsell other things they could find value in and purchase (digital goods, different kinds of content, etc.)
Always raise your price: the highest price always wins. Only a small percentage of people are willing to pay for in-app subscription, and those people are not really price sensitive. Example: speechify at $150, tests Thomas did with first $50 vs. $100, etc.
Price elasticity is typically very low, but don't push the fact that conversion doesn't change that much when you increase prices. It might be best to leave some money on the table (more users who pay less) if it means that you get more data, your community grows, etc.
There are beneficial effects of having a bigger userbase: word of mouth, viral effect, etc. but customer churn also becomes less meaningful.
In crowded spaces, people are not going to compare the apps' content that much (short term). But the power of brand is what comes in when you reach a degree of maturity, with often 1 or 2 players. Example: Peloton, Headspace, Calm. True in language learning too.
The problem is that people over experiment but don't test radical enough things. The less data you have and the more you haven't been down the road of experimentation, the more radical you should be.
Notes for this resource are currently being transferred and will be available soon.
You probably know who Thomas is, but just in case:
More and more companies running ads to a web destination, where users can both register and subscribe. Then they download the app. This allows apps to avoid the 30% app store fees.
Eric mentioned that the web onboarding flow could present 2 advantages:
1. Avoiding the 30% platform fee
2. People are used to fill out multi-part forms on the web
When first trying this in 2016/2017, Thomas' motivations for a web onboarding flow were:
All 3 assumptions were wrong, but it was still interesting to do it.
[💎 @08:40] Even with mobile web clicks not necessarily cheaper and more drop-offs on the web funnel, ads leading to a web onboarding flow can still be worth it: it can allow you to broaden the audience beyond the reach of app install campaigns (some networks don't target LAT people, some people don't click on app install ads and end up seeing less of them).
In Thomas' experience:
The user doesn't care about the platform: the user cares about the app's benefit
[💎 @10:09] The real benefit of a web onboarding flow is that people become your direct customers. It is a massive benefit to have that direct relationship, as it brings flexibility. The app stores bring a lot of convenience, but the customer ends up being in relation with Apple when it comes to cancellations, refunds, renewals, etc.
[💎 @10:52] Having a direct relationship with the customer allows you to do partial refunds: instead of cancelling a yearly subscription directly for the customer, you can offer them to refund half of the subscription while keeping their access on. It's a win-win.
Now there are introductory offers, promo codes, more flexibility on free trial, etc. But a referral program work on the App Store is still very tricky.
30% to Apple would be ok if you had direct access to the customer, not just on Apple's terms.
People still overestimate the advantages of web flows. There is not that much mobile web traffic. If you go outside of mobile app distribution channels to promote an app, it's very difficult.
The app store process can be a really helpful filter. Once users have downloaded, it shows that there is a lot of intent. On the web you need a lot of social proof to get someone to pull out their credit card vs. the App Store where users know they could get their money back from Apple (i.e less risks).
[💎 @17:05] With the IDFA deprecation coming up, a lot of people are trying to move to web. But it's a different mindset and very few people have the skills to understand the app store ecosystem AND the web ecosystem at the same time.
Challenges with web flows:
[💎 @20:22] The growth hacking mindset when it comes to funnels is not the right one. You can endlessly A/B test an onboarding flow yet still have your conversion rate go down, or get the same revenue.
[💎 @21:10] You can not just focus on A/B testing specific parts of the funnel because these local optimizations often disregard the impact on down-funnel metrics. Another critical aspect is the traffic composition.
The fact that the App Store pages are standardized puts in some constraints that prevents things from getting "out of control".
Even in-app, it's harder to A/B test as many variants on each step of the funnel which forces to adopt a more high-level view.
The constraints (App Store pages and limited A/B testing) forces you to think creatively on mobile.
If you A/B test everything too much, everything ends up being a porn website.
Whether it's mobile or web, smart growth people are at the intersection of product, data and marketing.
People coming from the marketing side segment the composition of traffic heavily: channel, OS, etc. But they look at relatively early events like signup and purchase.
People coming from product look very deep into the funnel and form hypothesis, but often disregard the traffic composition which makes it hard to understand why the product is reacting (e.g. when that traffic composition changes and brings challenges).
[💎 @27:12] Even if you consider both deeper-funnel events and traffic composition, be careful about over optimizing for revenue. Short term gains can end up being long term losses and prevent you from really scaling. Example: great payback period but high churn.
[💎 @28:12] Put secondary metrics ("tradeoff metrics") to every test you're doing. It's more complex but leads you to think bigger picture. Example: retention when optimizing for revenue, email sales when adding SSO, etc.
The vast majority of the mobile landscape, especially gaming, don't implement subscriptions in a meaningful way.
[💎 @29:30] Eric has seen some games where subscription players (often high LTV ones) end up up paying more on in-app purchases after subscribing. Players get a subscription to be more competitive, but as more players subscribe as well they still need to buy IAPs.
Thomas is bullish on subscriptions moving to a hybrid model.
[💎 @30:47] A lot of games are not doing subscriptions because they have already found a way to get their players to spend as much as possible.
[💎 @31:27] The beauty of the subscription model is the predictability/security but a major flaw is that you have a high floor and a low ceiling.
Plus, in several countries like Germany and China the subscription model is frowned upon.
Not everything is meant to be a subscription
Once you have a subscriber that is hooked to your product, you should upsell him. Examples: "premium" subscriptions, family plans (cf. Spotify), extra IAPs, non digital goods, etc.
[💎 @33:33] You can even sell non digital goods to your subscribed users. This is what the Sweat app is doing with a shop. The community wants branded goods. Another example with 8fit: super expensive subscription where they could deliver fruits & vegetables through Instacart.
[💎 @35:34] Think about how you can fight against the "low-ceiling" of your subscriptions. What can you upsell (another product, another subscription, a bundle with another app through a partnership, etc.)? Subscriptions don't have to be a binary thing.
[💎 @37:00] What's the probability that the price you set for your subscription is the max they're willing to spend? If you're already willing to buy the subscription, you'd probably be willing to pay more.
Is it worth it?
A lot of people are testing things that are too small, especially if they are startups.
[💎 @38:09] If you're an app like Calm or Tinder, a +2% gain is huge and small price optimizations end up being worth doing (especially because you have the data to do it).
[💎 @39:23] The problem is that people over experiment but don't test radical enough things. The less data you have and the more you haven't been down the road of experimentation, the more radical you should be.
At one point, wins from price optimizations become so marginal that you should either focus elsewhere or flip things up and change everything.
You should not make the parallel between mobile subscription businesses, SaaS businesses and content subscription businesses:
A very high amount of renewals are coming from people that have just forgotten about the subscription.
3-4 years ago, genius UA people could make a real difference for the business. This is harder to do now, with creatives being the main lever.
[💎 @40:52] If you find 5-10% increments in monetization, it might make a huge difference in the volume you can unlock through UA (passing a bid threshold can unlock acquisition volume). Example: an app that had already iterated on its paywall finds +15-17% that unlocked massive volume.
[💎 @41:55] Think about the state of mind in which the user is when he reaches the paywall. Don't think about cosmetic changes (bullet points, buttons, etc.). Example: show 30% discount if they skip the paywall, then if they skip again then email with 50% discount, etc. But you can go even further: daily pass when they hit the paywall the second time, free month if they invite somebody when they hit the paywall for the third time, etc.
If people only hit the paywall once then go out, you're too aggressive with your paywall
[💎 @42:57] The best experiments are the ones that work on the onboarding screens before the paywall. Anecdote: an app Thomas works with had tested having a longer loading screen ("preparing your plan"), which increased intent because it hinted at personalization (even though there was none).
Think about the whole flow: how they get to the paywall, what they are going to see after, etc.
[💎 @44:50] If you make the skip button more obvious on your paywall (instead of making it discreet) you churn people less, you get less dissatisfied people and you get an opportunity to monetize them another way.
Don't try hiding the skip buttons
For most companies, micro pricing optimizations are not the best way to leverage users' intent and extend the low ceiling.
Optimizing the subscription price based on the user (person A sees price X, person B sees price Y) is technically hard to do AND you can run into problems with your community if they find out.
[💎 @47:35] Instead of trying to optimize the subscription bundle at the highest price possible right out of the gate:
It can be worth lowering your subscription price if you can then sell something else through in-app purchases. Example: subscribers to Disney + (cheap subscription) had to pay a high price to watch the Mulan movie.
[💎 @52:06] Always raise your price: the highest price always wins. Only a small percentage of people are willing to pay for in-app subscription, and those people are not really price sensitive. Example: speechify at $150, tests Thomas did with first $50 vs. $100, etc.
[💎 @54:02] Price elasticity is typically very low, but don't push the fact that conversion doesn't change that much when you increase prices. It might be best to leave some money on the table (more users who pay less) if it means that you get more data, your community grows, etc.
[💎 @54:40] There are beneficial effects of having a bigger userbase: word of mouth, viral effect, etc. but customer churn also becomes less meaningful.
Setting the bar lower to onboard more people gives you the opportunity to later on find who's more valuable and upselling them.
Think about how you can bring value to your users from a product perspective.
Thomas mentioned a sleep app that went free, but with in-app coaching at hundreds of dollars.
Eric: apps that keep lowering endlessly their subscription price to scale and monetize through IAP would eventually become freemium apps.
Sleep and meditation are becoming crowded space. Is there a race to the bottom?
The price matters to nobody, not before and not now. The differentiation is on branding, the user experience, etc.
[💎 @01:01:50] In crowded spaces, people are not going to compare the apps' content that much (short term). But the power of brand is what comes in when you reach a degree of maturity, with often 1 or 2 players. Example: Peloton, Headspace, Calm. True in language learning too.
When companies are VC-backed, even though they could be profitable as #3/#4/#5 if they keep trying to win a category they might explode doing it.
In spaces that become crowded, you need to build a brand. But you need to build you brand it before you start plateauing.
That said there is always space to innovate and be #3 or #4 with good revenue. Example: Fitness AI.
Not everyone realizes that if you're funded, VCs expect you to become a category leader.
Because you're getting paid upfront when you have a subscription app, you can get one off the ground by bootstrapping.
But the plateau and the "explosion" is becoming the most common scenario.
A lot of second-time founders now take some pre-seed money to accelerate in the beginning, reach great profitability and repay back the pre-seed investors.
In smaller gaming companies, they pay back investors in dividends. Also seeing more companies paying employees with royalty bonuses.
Race to the bottom in video streaming with Disney +? Now flat churn for Netflix, etc.
Spotify has some good defensibility.
For video, there are a lot of players. And people don't necessarily have high loyalty, and will follow the content. That's why Amazon was smart there, because you're tied to the platform. Netflix has been paying to create their own content and is trying to build loyalty, same with Disney.
On Quibi: they approached it with a movie release playbook, and that was probably a mistake. But it might have also come down to the fact that they didn't have one "hit" show. They also didn't own the actual content.
If you're going after this kind of market, you can't be a small player and go after it by iterating.
You probably know who Thomas is, but just in case:
More and more companies running ads to a web destination, where users can both register and subscribe. Then they download the app. This allows apps to avoid the 30% app store fees.
Eric mentioned that the web onboarding flow could present 2 advantages:
1. Avoiding the 30% platform fee
2. People are used to fill out multi-part forms on the web
When first trying this in 2016/2017, Thomas' motivations for a web onboarding flow were:
All 3 assumptions were wrong, but it was still interesting to do it.
[💎 @08:40] Even with mobile web clicks not necessarily cheaper and more drop-offs on the web funnel, ads leading to a web onboarding flow can still be worth it: it can allow you to broaden the audience beyond the reach of app install campaigns (some networks don't target LAT people, some people don't click on app install ads and end up seeing less of them).
In Thomas' experience:
The user doesn't care about the platform: the user cares about the app's benefit
[💎 @10:09] The real benefit of a web onboarding flow is that people become your direct customers. It is a massive benefit to have that direct relationship, as it brings flexibility. The app stores bring a lot of convenience, but the customer ends up being in relation with Apple when it comes to cancellations, refunds, renewals, etc.
[💎 @10:52] Having a direct relationship with the customer allows you to do partial refunds: instead of cancelling a yearly subscription directly for the customer, you can offer them to refund half of the subscription while keeping their access on. It's a win-win.
Now there are introductory offers, promo codes, more flexibility on free trial, etc. But a referral program work on the App Store is still very tricky.
30% to Apple would be ok if you had direct access to the customer, not just on Apple's terms.
People still overestimate the advantages of web flows. There is not that much mobile web traffic. If you go outside of mobile app distribution channels to promote an app, it's very difficult.
The app store process can be a really helpful filter. Once users have downloaded, it shows that there is a lot of intent. On the web you need a lot of social proof to get someone to pull out their credit card vs. the App Store where users know they could get their money back from Apple (i.e less risks).
[💎 @17:05] With the IDFA deprecation coming up, a lot of people are trying to move to web. But it's a different mindset and very few people have the skills to understand the app store ecosystem AND the web ecosystem at the same time.
Challenges with web flows:
[💎 @20:22] The growth hacking mindset when it comes to funnels is not the right one. You can endlessly A/B test an onboarding flow yet still have your conversion rate go down, or get the same revenue.
[💎 @21:10] You can not just focus on A/B testing specific parts of the funnel because these local optimizations often disregard the impact on down-funnel metrics. Another critical aspect is the traffic composition.
The fact that the App Store pages are standardized puts in some constraints that prevents things from getting "out of control".
Even in-app, it's harder to A/B test as many variants on each step of the funnel which forces to adopt a more high-level view.
The constraints (App Store pages and limited A/B testing) forces you to think creatively on mobile.
If you A/B test everything too much, everything ends up being a porn website.
Whether it's mobile or web, smart growth people are at the intersection of product, data and marketing.
People coming from the marketing side segment the composition of traffic heavily: channel, OS, etc. But they look at relatively early events like signup and purchase.
People coming from product look very deep into the funnel and form hypothesis, but often disregard the traffic composition which makes it hard to understand why the product is reacting (e.g. when that traffic composition changes and brings challenges).
[💎 @27:12] Even if you consider both deeper-funnel events and traffic composition, be careful about over optimizing for revenue. Short term gains can end up being long term losses and prevent you from really scaling. Example: great payback period but high churn.
[💎 @28:12] Put secondary metrics ("tradeoff metrics") to every test you're doing. It's more complex but leads you to think bigger picture. Example: retention when optimizing for revenue, email sales when adding SSO, etc.
The vast majority of the mobile landscape, especially gaming, don't implement subscriptions in a meaningful way.
[💎 @29:30] Eric has seen some games where subscription players (often high LTV ones) end up up paying more on in-app purchases after subscribing. Players get a subscription to be more competitive, but as more players subscribe as well they still need to buy IAPs.
Thomas is bullish on subscriptions moving to a hybrid model.
[💎 @30:47] A lot of games are not doing subscriptions because they have already found a way to get their players to spend as much as possible.
[💎 @31:27] The beauty of the subscription model is the predictability/security but a major flaw is that you have a high floor and a low ceiling.
Plus, in several countries like Germany and China the subscription model is frowned upon.
Not everything is meant to be a subscription
Once you have a subscriber that is hooked to your product, you should upsell him. Examples: "premium" subscriptions, family plans (cf. Spotify), extra IAPs, non digital goods, etc.
[💎 @33:33] You can even sell non digital goods to your subscribed users. This is what the Sweat app is doing with a shop. The community wants branded goods. Another example with 8fit: super expensive subscription where they could deliver fruits & vegetables through Instacart.
[💎 @35:34] Think about how you can fight against the "low-ceiling" of your subscriptions. What can you upsell (another product, another subscription, a bundle with another app through a partnership, etc.)? Subscriptions don't have to be a binary thing.
[💎 @37:00] What's the probability that the price you set for your subscription is the max they're willing to spend? If you're already willing to buy the subscription, you'd probably be willing to pay more.
Is it worth it?
A lot of people are testing things that are too small, especially if they are startups.
[💎 @38:09] If you're an app like Calm or Tinder, a +2% gain is huge and small price optimizations end up being worth doing (especially because you have the data to do it).
[💎 @39:23] The problem is that people over experiment but don't test radical enough things. The less data you have and the more you haven't been down the road of experimentation, the more radical you should be.
At one point, wins from price optimizations become so marginal that you should either focus elsewhere or flip things up and change everything.
You should not make the parallel between mobile subscription businesses, SaaS businesses and content subscription businesses:
A very high amount of renewals are coming from people that have just forgotten about the subscription.
3-4 years ago, genius UA people could make a real difference for the business. This is harder to do now, with creatives being the main lever.
[💎 @40:52] If you find 5-10% increments in monetization, it might make a huge difference in the volume you can unlock through UA (passing a bid threshold can unlock acquisition volume). Example: an app that had already iterated on its paywall finds +15-17% that unlocked massive volume.
[💎 @41:55] Think about the state of mind in which the user is when he reaches the paywall. Don't think about cosmetic changes (bullet points, buttons, etc.). Example: show 30% discount if they skip the paywall, then if they skip again then email with 50% discount, etc. But you can go even further: daily pass when they hit the paywall the second time, free month if they invite somebody when they hit the paywall for the third time, etc.
If people only hit the paywall once then go out, you're too aggressive with your paywall
[💎 @42:57] The best experiments are the ones that work on the onboarding screens before the paywall. Anecdote: an app Thomas works with had tested having a longer loading screen ("preparing your plan"), which increased intent because it hinted at personalization (even though there was none).
Think about the whole flow: how they get to the paywall, what they are going to see after, etc.
[💎 @44:50] If you make the skip button more obvious on your paywall (instead of making it discreet) you churn people less, you get less dissatisfied people and you get an opportunity to monetize them another way.
Don't try hiding the skip buttons
For most companies, micro pricing optimizations are not the best way to leverage users' intent and extend the low ceiling.
Optimizing the subscription price based on the user (person A sees price X, person B sees price Y) is technically hard to do AND you can run into problems with your community if they find out.
[💎 @47:35] Instead of trying to optimize the subscription bundle at the highest price possible right out of the gate:
It can be worth lowering your subscription price if you can then sell something else through in-app purchases. Example: subscribers to Disney + (cheap subscription) had to pay a high price to watch the Mulan movie.
[💎 @52:06] Always raise your price: the highest price always wins. Only a small percentage of people are willing to pay for in-app subscription, and those people are not really price sensitive. Example: speechify at $150, tests Thomas did with first $50 vs. $100, etc.
[💎 @54:02] Price elasticity is typically very low, but don't push the fact that conversion doesn't change that much when you increase prices. It might be best to leave some money on the table (more users who pay less) if it means that you get more data, your community grows, etc.
[💎 @54:40] There are beneficial effects of having a bigger userbase: word of mouth, viral effect, etc. but customer churn also becomes less meaningful.
Setting the bar lower to onboard more people gives you the opportunity to later on find who's more valuable and upselling them.
Think about how you can bring value to your users from a product perspective.
Thomas mentioned a sleep app that went free, but with in-app coaching at hundreds of dollars.
Eric: apps that keep lowering endlessly their subscription price to scale and monetize through IAP would eventually become freemium apps.
Sleep and meditation are becoming crowded space. Is there a race to the bottom?
The price matters to nobody, not before and not now. The differentiation is on branding, the user experience, etc.
[💎 @01:01:50] In crowded spaces, people are not going to compare the apps' content that much (short term). But the power of brand is what comes in when you reach a degree of maturity, with often 1 or 2 players. Example: Peloton, Headspace, Calm. True in language learning too.
When companies are VC-backed, even though they could be profitable as #3/#4/#5 if they keep trying to win a category they might explode doing it.
In spaces that become crowded, you need to build a brand. But you need to build you brand it before you start plateauing.
That said there is always space to innovate and be #3 or #4 with good revenue. Example: Fitness AI.
Not everyone realizes that if you're funded, VCs expect you to become a category leader.
Because you're getting paid upfront when you have a subscription app, you can get one off the ground by bootstrapping.
But the plateau and the "explosion" is becoming the most common scenario.
A lot of second-time founders now take some pre-seed money to accelerate in the beginning, reach great profitability and repay back the pre-seed investors.
In smaller gaming companies, they pay back investors in dividends. Also seeing more companies paying employees with royalty bonuses.
Race to the bottom in video streaming with Disney +? Now flat churn for Netflix, etc.
Spotify has some good defensibility.
For video, there are a lot of players. And people don't necessarily have high loyalty, and will follow the content. That's why Amazon was smart there, because you're tied to the platform. Netflix has been paying to create their own content and is trying to build loyalty, same with Disney.
On Quibi: they approached it with a movie release playbook, and that was probably a mistake. But it might have also come down to the fact that they didn't have one "hit" show. They also didn't own the actual content.
If you're going after this kind of market, you can't be a small player and go after it by iterating.
You probably know who Thomas is, but just in case:
More and more companies running ads to a web destination, where users can both register and subscribe. Then they download the app. This allows apps to avoid the 30% app store fees.
Eric mentioned that the web onboarding flow could present 2 advantages:
1. Avoiding the 30% platform fee
2. People are used to fill out multi-part forms on the web
When first trying this in 2016/2017, Thomas' motivations for a web onboarding flow were:
All 3 assumptions were wrong, but it was still interesting to do it.
[💎 @08:40] Even with mobile web clicks not necessarily cheaper and more drop-offs on the web funnel, ads leading to a web onboarding flow can still be worth it: it can allow you to broaden the audience beyond the reach of app install campaigns (some networks don't target LAT people, some people don't click on app install ads and end up seeing less of them).
In Thomas' experience:
The user doesn't care about the platform: the user cares about the app's benefit
[💎 @10:09] The real benefit of a web onboarding flow is that people become your direct customers. It is a massive benefit to have that direct relationship, as it brings flexibility. The app stores bring a lot of convenience, but the customer ends up being in relation with Apple when it comes to cancellations, refunds, renewals, etc.
[💎 @10:52] Having a direct relationship with the customer allows you to do partial refunds: instead of cancelling a yearly subscription directly for the customer, you can offer them to refund half of the subscription while keeping their access on. It's a win-win.
Now there are introductory offers, promo codes, more flexibility on free trial, etc. But a referral program work on the App Store is still very tricky.
30% to Apple would be ok if you had direct access to the customer, not just on Apple's terms.
People still overestimate the advantages of web flows. There is not that much mobile web traffic. If you go outside of mobile app distribution channels to promote an app, it's very difficult.
The app store process can be a really helpful filter. Once users have downloaded, it shows that there is a lot of intent. On the web you need a lot of social proof to get someone to pull out their credit card vs. the App Store where users know they could get their money back from Apple (i.e less risks).
[💎 @17:05] With the IDFA deprecation coming up, a lot of people are trying to move to web. But it's a different mindset and very few people have the skills to understand the app store ecosystem AND the web ecosystem at the same time.
Challenges with web flows:
[💎 @20:22] The growth hacking mindset when it comes to funnels is not the right one. You can endlessly A/B test an onboarding flow yet still have your conversion rate go down, or get the same revenue.
[💎 @21:10] You can not just focus on A/B testing specific parts of the funnel because these local optimizations often disregard the impact on down-funnel metrics. Another critical aspect is the traffic composition.
The fact that the App Store pages are standardized puts in some constraints that prevents things from getting "out of control".
Even in-app, it's harder to A/B test as many variants on each step of the funnel which forces to adopt a more high-level view.
The constraints (App Store pages and limited A/B testing) forces you to think creatively on mobile.
If you A/B test everything too much, everything ends up being a porn website.
Whether it's mobile or web, smart growth people are at the intersection of product, data and marketing.
People coming from the marketing side segment the composition of traffic heavily: channel, OS, etc. But they look at relatively early events like signup and purchase.
People coming from product look very deep into the funnel and form hypothesis, but often disregard the traffic composition which makes it hard to understand why the product is reacting (e.g. when that traffic composition changes and brings challenges).
[💎 @27:12] Even if you consider both deeper-funnel events and traffic composition, be careful about over optimizing for revenue. Short term gains can end up being long term losses and prevent you from really scaling. Example: great payback period but high churn.
[💎 @28:12] Put secondary metrics ("tradeoff metrics") to every test you're doing. It's more complex but leads you to think bigger picture. Example: retention when optimizing for revenue, email sales when adding SSO, etc.
The vast majority of the mobile landscape, especially gaming, don't implement subscriptions in a meaningful way.
[💎 @29:30] Eric has seen some games where subscription players (often high LTV ones) end up up paying more on in-app purchases after subscribing. Players get a subscription to be more competitive, but as more players subscribe as well they still need to buy IAPs.
Thomas is bullish on subscriptions moving to a hybrid model.
[💎 @30:47] A lot of games are not doing subscriptions because they have already found a way to get their players to spend as much as possible.
[💎 @31:27] The beauty of the subscription model is the predictability/security but a major flaw is that you have a high floor and a low ceiling.
Plus, in several countries like Germany and China the subscription model is frowned upon.
Not everything is meant to be a subscription
Once you have a subscriber that is hooked to your product, you should upsell him. Examples: "premium" subscriptions, family plans (cf. Spotify), extra IAPs, non digital goods, etc.
[💎 @33:33] You can even sell non digital goods to your subscribed users. This is what the Sweat app is doing with a shop. The community wants branded goods. Another example with 8fit: super expensive subscription where they could deliver fruits & vegetables through Instacart.
[💎 @35:34] Think about how you can fight against the "low-ceiling" of your subscriptions. What can you upsell (another product, another subscription, a bundle with another app through a partnership, etc.)? Subscriptions don't have to be a binary thing.
[💎 @37:00] What's the probability that the price you set for your subscription is the max they're willing to spend? If you're already willing to buy the subscription, you'd probably be willing to pay more.
Is it worth it?
A lot of people are testing things that are too small, especially if they are startups.
[💎 @38:09] If you're an app like Calm or Tinder, a +2% gain is huge and small price optimizations end up being worth doing (especially because you have the data to do it).
[💎 @39:23] The problem is that people over experiment but don't test radical enough things. The less data you have and the more you haven't been down the road of experimentation, the more radical you should be.
At one point, wins from price optimizations become so marginal that you should either focus elsewhere or flip things up and change everything.
You should not make the parallel between mobile subscription businesses, SaaS businesses and content subscription businesses:
A very high amount of renewals are coming from people that have just forgotten about the subscription.
3-4 years ago, genius UA people could make a real difference for the business. This is harder to do now, with creatives being the main lever.
[💎 @40:52] If you find 5-10% increments in monetization, it might make a huge difference in the volume you can unlock through UA (passing a bid threshold can unlock acquisition volume). Example: an app that had already iterated on its paywall finds +15-17% that unlocked massive volume.
[💎 @41:55] Think about the state of mind in which the user is when he reaches the paywall. Don't think about cosmetic changes (bullet points, buttons, etc.). Example: show 30% discount if they skip the paywall, then if they skip again then email with 50% discount, etc. But you can go even further: daily pass when they hit the paywall the second time, free month if they invite somebody when they hit the paywall for the third time, etc.
If people only hit the paywall once then go out, you're too aggressive with your paywall
[💎 @42:57] The best experiments are the ones that work on the onboarding screens before the paywall. Anecdote: an app Thomas works with had tested having a longer loading screen ("preparing your plan"), which increased intent because it hinted at personalization (even though there was none).
Think about the whole flow: how they get to the paywall, what they are going to see after, etc.
[💎 @44:50] If you make the skip button more obvious on your paywall (instead of making it discreet) you churn people less, you get less dissatisfied people and you get an opportunity to monetize them another way.
Don't try hiding the skip buttons
For most companies, micro pricing optimizations are not the best way to leverage users' intent and extend the low ceiling.
Optimizing the subscription price based on the user (person A sees price X, person B sees price Y) is technically hard to do AND you can run into problems with your community if they find out.
[💎 @47:35] Instead of trying to optimize the subscription bundle at the highest price possible right out of the gate:
It can be worth lowering your subscription price if you can then sell something else through in-app purchases. Example: subscribers to Disney + (cheap subscription) had to pay a high price to watch the Mulan movie.
[💎 @52:06] Always raise your price: the highest price always wins. Only a small percentage of people are willing to pay for in-app subscription, and those people are not really price sensitive. Example: speechify at $150, tests Thomas did with first $50 vs. $100, etc.
[💎 @54:02] Price elasticity is typically very low, but don't push the fact that conversion doesn't change that much when you increase prices. It might be best to leave some money on the table (more users who pay less) if it means that you get more data, your community grows, etc.
[💎 @54:40] There are beneficial effects of having a bigger userbase: word of mouth, viral effect, etc. but customer churn also becomes less meaningful.
Setting the bar lower to onboard more people gives you the opportunity to later on find who's more valuable and upselling them.
Think about how you can bring value to your users from a product perspective.
Thomas mentioned a sleep app that went free, but with in-app coaching at hundreds of dollars.
Eric: apps that keep lowering endlessly their subscription price to scale and monetize through IAP would eventually become freemium apps.
Sleep and meditation are becoming crowded space. Is there a race to the bottom?
The price matters to nobody, not before and not now. The differentiation is on branding, the user experience, etc.
[💎 @01:01:50] In crowded spaces, people are not going to compare the apps' content that much (short term). But the power of brand is what comes in when you reach a degree of maturity, with often 1 or 2 players. Example: Peloton, Headspace, Calm. True in language learning too.
When companies are VC-backed, even though they could be profitable as #3/#4/#5 if they keep trying to win a category they might explode doing it.
In spaces that become crowded, you need to build a brand. But you need to build you brand it before you start plateauing.
That said there is always space to innovate and be #3 or #4 with good revenue. Example: Fitness AI.
Not everyone realizes that if you're funded, VCs expect you to become a category leader.
Because you're getting paid upfront when you have a subscription app, you can get one off the ground by bootstrapping.
But the plateau and the "explosion" is becoming the most common scenario.
A lot of second-time founders now take some pre-seed money to accelerate in the beginning, reach great profitability and repay back the pre-seed investors.
In smaller gaming companies, they pay back investors in dividends. Also seeing more companies paying employees with royalty bonuses.
Race to the bottom in video streaming with Disney +? Now flat churn for Netflix, etc.
Spotify has some good defensibility.
For video, there are a lot of players. And people don't necessarily have high loyalty, and will follow the content. That's why Amazon was smart there, because you're tied to the platform. Netflix has been paying to create their own content and is trying to build loyalty, same with Disney.
On Quibi: they approached it with a movie release playbook, and that was probably a mistake. But it might have also come down to the fact that they didn't have one "hit" show. They also didn't own the actual content.
If you're going after this kind of market, you can't be a small player and go after it by iterating.
