Scott Dodson (CGO of Lingvist - language learning app) talks about getting better return on acquisition spend.
Growth comes from loops so you want to:
- Identify & amplify your core loop(s). For Lingvist it is word of mouth and paid UA - Identify, build and support a system of compounding loops
If you are not already, figure out your ROI and its relation with your ROAS so you can make early adjustments.
For their creative process, Lingvist switched their OKR focus to diversity & output (10 concepts tested) as the lead metric and ROAS improvement (ROAS from X to Y) as a lag metric.
Lingvist created weekly "ad-cept" (concept) meeting with a process opened up beyond the creative team. Preparation was mandatory and ideally with quickly executable tests.
They unified performance results in a daily data bundle to analyze creative performance depending on audience and creative type in order to have a fast performance feedback loop
You want to get users that get into the door (open your app) and get them to experience the value of the product through a "aha" moment so they become a regular user. For Lingvist it is key because users need to try the app and get some things wrong so the app reminds them of words right on time.
A flat retention curve is critical to growth. If every new cohort that you have trends to 0 at some point in time then it doesn't matter how many people you acquire: your growth plateaus.
You want to look at a lot of different behaviors from people when they start using the app to find what makes them stick around. For example Lingvist looked at which language, features - voice recognition, specific course, number of cards done, etc.) kept users around.
Do an analysis to find the behavior pattern of the first month that is predictive of 12th week retention. Then look at what behavior during the first week is predictive of the first month habit forming behavior so you can optimize for it.
Once you know which behavior you want people to achieve when they become a user, you can focus efforts on it through the onboarding system for example.
Growth comes from loops so you want to:
- Identify & amplify your core loop(s). For Lingvist it is word of mouth and paid UA - Identify, build and support a system of compounding loops
If you are not already, figure out your ROI and its relation with your ROAS so you can make early adjustments.
For their creative process, Lingvist switched their OKR focus to diversity & output (10 concepts tested) as the lead metric and ROAS improvement (ROAS from X to Y) as a lag metric.
Lingvist created weekly "ad-cept" (concept) meeting with a process opened up beyond the creative team. Preparation was mandatory and ideally with quickly executable tests.
They unified performance results in a daily data bundle to analyze creative performance depending on audience and creative type in order to have a fast performance feedback loop
You want to get users that get into the door (open your app) and get them to experience the value of the product through a "aha" moment so they become a regular user. For Lingvist it is key because users need to try the app and get some things wrong so the app reminds them of words right on time.
A flat retention curve is critical to growth. If every new cohort that you have trends to 0 at some point in time then it doesn't matter how many people you acquire: your growth plateaus.
You want to look at a lot of different behaviors from people when they start using the app to find what makes them stick around. For example Lingvist looked at which language, features - voice recognition, specific course, number of cards done, etc.) kept users around.
Do an analysis to find the behavior pattern of the first month that is predictive of 12th week retention. Then look at what behavior during the first week is predictive of the first month habit forming behavior so you can optimize for it.
Once you know which behavior you want people to achieve when they become a user, you can focus efforts on it through the onboarding system for example.
Growth comes from loops so you want to:
- Identify & amplify your core loop(s). For Lingvist it is word of mouth and paid UA - Identify, build and support a system of compounding loops
If you are not already, figure out your ROI and its relation with your ROAS so you can make early adjustments.
For their creative process, Lingvist switched their OKR focus to diversity & output (10 concepts tested) as the lead metric and ROAS improvement (ROAS from X to Y) as a lag metric.
Lingvist created weekly "ad-cept" (concept) meeting with a process opened up beyond the creative team. Preparation was mandatory and ideally with quickly executable tests.
They unified performance results in a daily data bundle to analyze creative performance depending on audience and creative type in order to have a fast performance feedback loop
You want to get users that get into the door (open your app) and get them to experience the value of the product through a "aha" moment so they become a regular user. For Lingvist it is key because users need to try the app and get some things wrong so the app reminds them of words right on time.
A flat retention curve is critical to growth. If every new cohort that you have trends to 0 at some point in time then it doesn't matter how many people you acquire: your growth plateaus.
You want to look at a lot of different behaviors from people when they start using the app to find what makes them stick around. For example Lingvist looked at which language, features - voice recognition, specific course, number of cards done, etc.) kept users around.
Do an analysis to find the behavior pattern of the first month that is predictive of 12th week retention. Then look at what behavior during the first week is predictive of the first month habit forming behavior so you can optimize for it.
Once you know which behavior you want people to achieve when they become a user, you can focus efforts on it through the onboarding system for example.
Notes for this resource are currently being transferred and will be available soon.
Lingvist ROAS on UAC over time
Programs like Reforge change your approach from a funnel model to a loop model.
An example for Lingvist is the viral loop of word of mouth
Another example of viral loop they tried is the incentivized referral program they created, but it did not perform as expected so they stopped it
And also the Paid User Acquisition loop: do you earn more money off of new subscribers to buy enough ads to get more users?
Company generated content/SEO: taking money to invest in content to get searches and get new users
PR & Earned media loop
All these loops can be interconnected and tell the story of how Lingvist grows
[💎@11:30] Growth comes from loops so you want to:
When the paid UA loop spins the other loops spin too: every user that comes into an ad is potentially going to tell friends, etc.
[💎@12:40] If you are not already, figure out your ROI and its relation with your ROAS so you can make early adjustments.
How Lingvist improved their UA, mostly thanks to improvements in the creative process
Looked at different kind of motivations, from not caring to much higher motivation. Also figured out where competitors stand (e.g. Duolingo at the "Sudoku" motivation, italki on the "Athlete" motivation) and determined they are attacking:
First generation creatives
Upgrade in-house improvements (including giving a preview of what the app looks like)
Then change their in-house process and moved to creating a bunch of concepts: become a machine and see what sticks.
Examples:
Some results
A best performer with C being a close second. B good TOFU metrics but not after. C high CPI but great retention.
[💎@22:22] You want to get users that get into the door (open your app) and get them to experience the value of the product through a "aha" moment so they become a regular user. For Lingvist it is key because users need to try the app and get some things wrong so the app reminds them of words right on time.
[💎@23:50] A flat retention curve is critical to growth. If every new cohort that you have trends to 0 at some point in time then it doesn't matter how many people you acquire: your growth plateaus.
When some portion of your userbase sticks around you can build a stack of userbase cohort after cohort.
The engagement cycle varies by company/industry (daily for whatsapp, yearly for tax service, etc.). For Lingvist it is weekly and flat engagement is defined at 12 weeks.
Massive drop-off and it trends towards 0 but very, very slowly.
[💎@25:43] You want to look at a lot of different behaviors from people when they start using the app to find what makes them stick around. For example Lingvist looked at which language, features - voice recognition, specific course, number of cards done, etc.)
Doing 50 cards per session is what gets Lingvist users to be retained the most.
[💎@26:10] Do an analysis to find the behavior pattern of the first month that is predictive of 12th week retention. For Lingvist if somebody does two 50 cards set per week during the first 28 days, they are very likely to stick around through the 12th week.
Then look at what behavior during the first week is predictive of the first month habit forming behavior.
If a user does 4 sets of 50 cards in the first week, they are very likely to do the habit forming behavior and do at least one more set or more by the end of the month.
[💎@27:15] Once you know which behavior you want people to achieve when they become a user, you can focus efforts on it through the onboarding system for example.
Lingvist created a new onboarding system that focused on the 50 cards set (from 100) with some positive reinforcement.
They rolled out this new onboarding gradually (hence the curve)
Lingvist ROAS on UAC over time
Programs like Reforge change your approach from a funnel model to a loop model.
An example for Lingvist is the viral loop of word of mouth
Another example of viral loop they tried is the incentivized referral program they created, but it did not perform as expected so they stopped it
And also the Paid User Acquisition loop: do you earn more money off of new subscribers to buy enough ads to get more users?
Company generated content/SEO: taking money to invest in content to get searches and get new users
PR & Earned media loop
All these loops can be interconnected and tell the story of how Lingvist grows
[💎@11:30] Growth comes from loops so you want to:
When the paid UA loop spins the other loops spin too: every user that comes into an ad is potentially going to tell friends, etc.
[💎@12:40] If you are not already, figure out your ROI and its relation with your ROAS so you can make early adjustments.
How Lingvist improved their UA, mostly thanks to improvements in the creative process
Looked at different kind of motivations, from not caring to much higher motivation. Also figured out where competitors stand (e.g. Duolingo at the "Sudoku" motivation, italki on the "Athlete" motivation) and determined they are attacking:
First generation creatives
Upgrade in-house improvements (including giving a preview of what the app looks like)
Then change their in-house process and moved to creating a bunch of concepts: become a machine and see what sticks.
Examples:
Some results
A best performer with C being a close second. B good TOFU metrics but not after. C high CPI but great retention.
[💎@22:22] You want to get users that get into the door (open your app) and get them to experience the value of the product through a "aha" moment so they become a regular user. For Lingvist it is key because users need to try the app and get some things wrong so the app reminds them of words right on time.
[💎@23:50] A flat retention curve is critical to growth. If every new cohort that you have trends to 0 at some point in time then it doesn't matter how many people you acquire: your growth plateaus.
When some portion of your userbase sticks around you can build a stack of userbase cohort after cohort.
The engagement cycle varies by company/industry (daily for whatsapp, yearly for tax service, etc.). For Lingvist it is weekly and flat engagement is defined at 12 weeks.
Massive drop-off and it trends towards 0 but very, very slowly.
[💎@25:43] You want to look at a lot of different behaviors from people when they start using the app to find what makes them stick around. For example Lingvist looked at which language, features - voice recognition, specific course, number of cards done, etc.)
Doing 50 cards per session is what gets Lingvist users to be retained the most.
[💎@26:10] Do an analysis to find the behavior pattern of the first month that is predictive of 12th week retention. For Lingvist if somebody does two 50 cards set per week during the first 28 days, they are very likely to stick around through the 12th week.
Then look at what behavior during the first week is predictive of the first month habit forming behavior.
If a user does 4 sets of 50 cards in the first week, they are very likely to do the habit forming behavior and do at least one more set or more by the end of the month.
[💎@27:15] Once you know which behavior you want people to achieve when they become a user, you can focus efforts on it through the onboarding system for example.
Lingvist created a new onboarding system that focused on the 50 cards set (from 100) with some positive reinforcement.
They rolled out this new onboarding gradually (hence the curve)
Lingvist ROAS on UAC over time
Programs like Reforge change your approach from a funnel model to a loop model.
An example for Lingvist is the viral loop of word of mouth
Another example of viral loop they tried is the incentivized referral program they created, but it did not perform as expected so they stopped it
And also the Paid User Acquisition loop: do you earn more money off of new subscribers to buy enough ads to get more users?
Company generated content/SEO: taking money to invest in content to get searches and get new users
PR & Earned media loop
All these loops can be interconnected and tell the story of how Lingvist grows
[💎@11:30] Growth comes from loops so you want to:
When the paid UA loop spins the other loops spin too: every user that comes into an ad is potentially going to tell friends, etc.
[💎@12:40] If you are not already, figure out your ROI and its relation with your ROAS so you can make early adjustments.
How Lingvist improved their UA, mostly thanks to improvements in the creative process
Looked at different kind of motivations, from not caring to much higher motivation. Also figured out where competitors stand (e.g. Duolingo at the "Sudoku" motivation, italki on the "Athlete" motivation) and determined they are attacking:
First generation creatives
Upgrade in-house improvements (including giving a preview of what the app looks like)
Then change their in-house process and moved to creating a bunch of concepts: become a machine and see what sticks.
Examples:
Some results
A best performer with C being a close second. B good TOFU metrics but not after. C high CPI but great retention.
[💎@22:22] You want to get users that get into the door (open your app) and get them to experience the value of the product through a "aha" moment so they become a regular user. For Lingvist it is key because users need to try the app and get some things wrong so the app reminds them of words right on time.
[💎@23:50] A flat retention curve is critical to growth. If every new cohort that you have trends to 0 at some point in time then it doesn't matter how many people you acquire: your growth plateaus.
When some portion of your userbase sticks around you can build a stack of userbase cohort after cohort.
The engagement cycle varies by company/industry (daily for whatsapp, yearly for tax service, etc.). For Lingvist it is weekly and flat engagement is defined at 12 weeks.
Massive drop-off and it trends towards 0 but very, very slowly.
[💎@25:43] You want to look at a lot of different behaviors from people when they start using the app to find what makes them stick around. For example Lingvist looked at which language, features - voice recognition, specific course, number of cards done, etc.)
Doing 50 cards per session is what gets Lingvist users to be retained the most.
[💎@26:10] Do an analysis to find the behavior pattern of the first month that is predictive of 12th week retention. For Lingvist if somebody does two 50 cards set per week during the first 28 days, they are very likely to stick around through the 12th week.
Then look at what behavior during the first week is predictive of the first month habit forming behavior.
If a user does 4 sets of 50 cards in the first week, they are very likely to do the habit forming behavior and do at least one more set or more by the end of the month.
[💎@27:15] Once you know which behavior you want people to achieve when they become a user, you can focus efforts on it through the onboarding system for example.
Lingvist created a new onboarding system that focused on the 50 cards set (from 100) with some positive reinforcement.
They rolled out this new onboarding gradually (hence the curve)
