The Science of Sticky Apps study guide · Lesson 7 of 11
Retention Architecture
Onboarding & the Aha Moment · Push notifications · Investment & cohort retention
Unofficial study guide for Hooked: How to Build Habit-Forming Products by Nir Eyal, Ryan Hoover. FlashFeed is not affiliated with the author or publisher.
This is the lesson's study guide from the FlashFeed app, plus its key facts and a self-test.
What you'll be able to do
- Define the Aha Moment as feeling core value and match Facebook, Twitter, and Duolingo to their onboarding thresholds.
- Explain how teams locate the Aha Moment empirically from cohort data and cite the day-3 drop-off.
- State the three rules of effective push notifications (timing, personalization, curiosity/FOMO) and what not to do.
- Explain the Investment (Hook) stage as raising switching cost via the endowment effect, and identify a Duolingo streak as accumulated investment.
- Read cohort retention benchmarks (Day 1/7/30) and explain why a small retention gain can mean a large profit gain.
Onboarding & the Aha Moment
Retention isn't decided over months — it's decided in hours. The entire onboarding discipline exists to collapse the distance between "downloaded the app" and "understood why it matters."
Retention is won in the first session
Roughly 80% of users stop using an app after the third day following install. That single statistic reframes onboarding as a survival problem: the app has two or three sessions to prove its value, and then most people are gone for good. The solution isn't a better tutorial — it's faster delivery of the Aha Moment, the instant the app's core value becomes emotionally real. Finishing a walkthrough is not the Aha Moment; feeling the value is.
Product teams locate the Aha Moment empirically: they analyze which early actions correlate with long-term retention in cohort data. When a specific behavior — reaching it within session one — predicts a user still being active 30 days later, that behavior is the Aha Moment. The engineering and UX then bend around getting every new user to that moment as fast as possible.
The canonical examples each involve a concrete threshold. Facebook landed on "7 friends in 10 days" — hit that social density and the feed becomes meaningful; miss it and the product is an empty room. Twitter's equivalent was "follow 30 accounts" — enough signal to fill a timeline worth returning to. Duolingo's is immediate and sensory: completing the first lesson and receiving the animated XP reward — the gamified dopamine hit that makes the habit feel attainable right now.
All three share a structure: a specific, measurable threshold that converts a passive visitor into someone with a felt reason to come back. Onboarding design is the art of removing every obstacle between install and that threshold.
Push notifications
Getting a user to the Aha Moment buys time; keeping them returning requires a re-engagement channel. Push notifications are that channel — but only when used with restraint and precision.
Push: the re-engagement channel
Push notifications are the primary channel for reactivating users who have left the habit loop. When a user misses a day — or a week — a well-timed notification is often the only thing that pulls them back before they forget the app exists. Done badly, push becomes noise that triggers a permanent opt-out; done well, it mimics the gentle nudge of a real commitment.
Three rules separate effective push from spam. First, timing: notifications land better when they arrive at the moment a user has historically opened the app, not on a fixed broadcast schedule. Sending at the same fixed interval every day regardless of context is precisely what not to do — it is the one approach that is not among the three effective rules. Second, personalization: using the user's name and behavioral history — their streak length, their last completed action, their stated goal — outperforms generic copy significantly, because it signals the app is paying attention. Third, curiosity and the FOMO gap: copy that withholds just enough information ("Your streak is at risk" rather than "Open the app") creates mild anxiety that motivates a tap without being manipulative.
The underlying mechanism is the curiosity gap — the mild discomfort of an unresolved question. Combined with FOMO (fear of missing out on a streak, a reward, or a social event), even a short notification can feel urgent. The risk is habituation: if every push is framed as urgent, none are. The craft is calibrating frequency and intensity so the channel retains its pull.
Investment & cohort retention
The Aha Moment and a well-timed push notification win the first weeks. Whether a user stays for months depends on what they have built inside the product, and whether retention is working is measured one cohort at a time.
Investment and stored value
Investment is the fourth stage of the Hook Model: after the reward, the user puts something back into the product. Nir Eyal's Hooked calls the result stored value, which accrues as content, data, followers, reputation and skill the more the product is used. On his own site Eyal lists inviting friends, stating preferences, building virtual assets and learning new features as investments, and says they can make the next trigger more engaging, the action easier and the reward more exciting. He cites LinkedIn: an early product manager there, Josh Elman, found that users who entered even a little information were much more likely to return.
Investment holds users in two ways. It makes the product more valuable with each use, and it raises the switching cost of leaving, because the endowment effect leads people to overvalue what they already own. Everything a user has built stays behind if they go.
The streak as investment
Duolingo's streak count is the clearest example in this pack. Each day of practice adds to a number the user built and can lose. A 100-day streak is accumulated sunk cost: past investment that has no bearing on future value but feels devastatingly real, so the threat of a reset triggers loss aversion. The counter does not measure fluency; it records the user's own run of daily practice.
Cohort analysis
Whether any of this works shows up in cohort analysis. A cohort is a group of users who share a starting point, such as the same install or sign-up date. Cohort retention tracks how many of that group return over time, and N-day retention is the share of the cohort active on the Nth day after they started. Common checkpoints are Day 1, Day 7 and Day 30, with longer views such as Day 90. Grouping users by early behavior instead of by date is how teams find the Aha Moment: the actions that correlate with higher long-term retention.
What typical retention looks like
Benchmarks show how steep the curve is. The mobile measurement company Adjust, in retention benchmarks published in 2025, put the median across all app categories at 26% on Day 1, 13% on Day 7 and 7% on Day 30. iOS apps ran slightly higher (27%, 14% and 8%) and Android slightly lower (24% on Day 1, 6% on Day 30). Retention also varies widely by category, from 29% on Day 1 for games and social apps to 13% for food and drink, so a single median is a reference point, not a target.
Why retention pays
LTV (Lifetime Value) is the total value a user generates over their lifetime, and it grows the longer a user stays: a 2001 Bain & Company note by Frederick Reichheld observes that customers generate increasing profits each year they stay. A widely cited figure attributed to Reichheld, repeated in Harvard Business Review in 2014, holds that raising customer retention rates by 5% increases profits by 25% to 95%, and that acquiring a new customer costs five to 25 times more than keeping one. The figure describes company profits, not app LTV specifically, and the range varies by business. The direction is the lesson: small gains in retention compound, which is why retention is treated as more capital-efficient than paid acquisition.
Key terms
- Aha Moment
- The instant an app's core value becomes emotionally real — feeling the value, not finishing a tutorial. Facebook's was '7 friends in 10 days'; Twitter's was following 30 accounts. Onboarding should deliver every user here fast.
- Push notification
- The primary channel for reactivating users who have left the habit loop. The most effective ones are contextually timed, personalized to the user's state, and carry a curiosity or FOMO gap.
- Investment (Hook)
- The fourth Hook stage: users put something into the product — data, content, social connections, reputation. Investment makes the product more valuable on the next use and raises switching costs via the endowment effect: everything a user builds inside the app raises the cost of leaving.
- Cohort analysis
- Tracking the percentage of users from a given install cohort who return on Day 1, 7, 30, and 90. It is also how teams identify the Aha Moment — the early actions that correlate with long-term retention.
- LTV (Lifetime Value)
- The total value a user generates over their lifetime. Paid acquisition loops are sustainable only when it greatly exceeds CAC.
Key concepts
- This social network located its Aha Moment as a specific social-density threshold new users had to hit early on.
- Reach that density and the feed becomes meaningful; miss it and the product feels like an empty room.
- Its famous benchmark: reaching 7 friends within the first 10 days of signup.
- This microblogging platform defined its Aha Moment as reaching enough signal to make the timeline worth returning to.
- The threshold was about following enough sources to fill a feed.
- Its specific onboarding target was following 30 accounts.
Duolingo
- This language-learning app's Aha Moment is immediate and sensory rather than social.
- It hinges on finishing your very first exercise and getting an instant gamified payoff.
- New users feel the value the moment an animated XP celebration fires after lesson one — a dopamine hit that makes the habit feel attainable.
Key facts
Retention is won in the first session
Roughly 80% of users stop using an app after the third day following install. The entire onboarding architecture exists to do one thing: deliver the user to the moment the app's value becomes real, before day three takes them.
When Value Feels Real
An app's Aha Moment is the instant its core value becomes emotionally real to the user — not the moment they finish a tutorial.
Twitter's Onboarding Number
Twitter's Aha-Moment onboarding threshold was following 30 accounts — enough signal to fill a timeline worth returning to.
Duolingo's Instant Payoff
Duolingo's Aha Moment is immediate and sensory: finishing the first lesson and getting the animated XP reward, a dopamine hit rather than a slow-building realization.
Hunting The Aha Moment
Product teams find an app's Aha Moment by analyzing cohort data for which early actions correlate with long-term retention, then re-engineering onboarding to steer every new user through that sequence.
Push: the re-engagement channel
Push notifications are the main way to reactivate users who left the habit loop. Personalized notifications — using the user's name and behavioral history — significantly outperform generic ones at driving people back.
Timed To Your Habits
The timing rule for push says notifications should land at moments a user has historically opened the app — not at a fixed interval sent regardless of context.
Copy That Pays Attention
Personalized push copy — referencing a user's name, streak, or behavioral history — outperforms generic copy because it signals the app is actually paying attention to them.
The Curiosity Gap Trick
Push copy that withholds just enough — like 'Your streak is at risk' — creates a curiosity gap and mild FOMO that motivates a tap, unlike copy that resolves everything upfront.
By the numbers
- Share of users who stop using an app after day 3 following install
- 80 %
- Median Day 1 retention across all apps (Adjust benchmark, 2025)
- 26 %
- Median Day 7 retention across all apps (Adjust benchmark, 2025)
- 13 %
- Median Day 30 retention across all apps (Adjust benchmark, 2025)
- 7 %
- Upper end of the profit increase attributed to a 5% rise in customer retention (Bain)
- 95 %
Matched pairs
Each item on the left goes with the one on the right.
- 7 friends in 10 days
- follow 30 accounts
- Duolingo
- first lesson + animated XP
- Timing
- send at high-probability moments — post-commute, lunch, evening
- Personalization
- reference the user's specific state — '2 days from your longest streak'
- Curiosity / FOMO gap
- 'Your friend just overtook your rank' — competitive anxiety plus social proof
- Duolingo
- streak count
Step by step
Engineer onboarding to deliver every new user to the Aha Moment
- Pull cohort data on users who stuck around long-term versus those who churned.
- Find the early actions that correlate with long-term retention.
- Define the Aha Moment as that exact action sequence (e.g. 7 friends in 10 days).
- Re-engineer onboarding so every new user is driven to that action as fast as possible.
- Confirm the value is felt emotionally, not just that a tutorial was completed.
Teams find the Aha Moment by analysis, not intuition — then build the funnel backward from it.
Self-test
Answer each question from memory, then check your work against the answer key.
1. Facebook famously identified a single onboarding threshold as its Aha Moment. What was it?
- AFollow 30 accounts
- B5 friends in the first session
- C10 friends in 30 days
- D7 friends in 10 days
2. What best defines an app's Aha Moment?
- AThe day-3 point when most users churn
- BThe instant the app's core value becomes emotionally real
- CThe screen where a user grants push permission
- DThe moment a user finishes the onboarding walkthrough
3. What was Twitter's Aha-Moment onboarding threshold?
- AFollow 30 accounts
- B7 friends in 10 days
- CAdd 5 topics of interest
- DComplete the first lesson for animated XP
4. Which describes Duolingo's Aha Moment?
- AMaintaining a 7-day streak
- BReaching 7 friends in 10 days
- CCompleting the first lesson and receiving the animated XP reward
- DFollowing 30 accounts
5. Which of these is NOT an example of an app's Aha Moment?
- ADuolingo's first lesson plus animated XP
- BFinishing the onboarding walkthrough
- CFacebook's 7 friends in 10 days
- DTwitter's follow 30 accounts
6. How do product teams identify an app's Aha Moment?
- ACopy whatever the market-leading competitor does
- BPick the final screen of the onboarding tutorial
- CAnalyze which early actions correlate with long-term retention in cohort data
- DAsk users in a survey what they enjoyed most
7. Roughly what share of users stop using an app after the third day following install?
- A95%
- B65%
- C50%
- D80%
8. A specific early behavior qualifies as the Aha Moment when it does what?
- APredicts a user still being active about 30 days later
- BMatches a market-leading competitor's feature
- CIs rated most enjoyable in a user survey
- DAppears on the final onboarding screen
9. The ~80% day-3 drop-off reframes onboarding as which kind of problem?
- AA survival problem — prove value within two or three sessions
- BA tutorial-length problem — the walkthrough is too short
- CA paid-acquisition problem — not enough installs
- DA pricing problem — the app costs too much up front
10. Which is NOT a valid way to locate an app's Aha Moment?
- AExamining the early actions of users who stayed active
- BChoosing the last screen of the onboarding tutorial
- CFinding behaviors that predict 30-day retention
- DAnalyzing which early actions correlate with long-term retention
11. Which is NOT one of the three rules for effective push notifications?
- ATime it to moments of high probability of action
- BOpen a curiosity or FOMO gap
- CSend at the same fixed interval every day regardless of context
- DPersonalize it to the user's specific state
12. What is a push notification's primary role?
- AA paid user-acquisition channel
- BA required step in the onboarding walkthrough
- CThe Investment stage of the Hook Model
- DThe primary channel for reactivating users who left the habit loop
13. The timing rule says push notifications should arrive when?
- AAt the same fixed interval every day
- BAt moments the user has historically opened the app
- COnly at midnight to avoid daytime noise
- DAs often as possible to stay top of mind
14. Why does personalized push copy (name and behavioral history) outperform generic copy?
- AIt is cheaper to send at scale
- BIt bypasses app-store notification rules
- CIt removes any need for good timing
- DIt signals the app is paying attention to the user
15. Which push copy best uses the curiosity / FOMO gap?
- A'Your streak is at risk'
- B'Open the app'
- C'Thanks for installing our app'
- D'A new update is available'
16. In the Hook Model's Investment stage, what do users put into the product?
- AA one-time app download
- BMoney for a subscription only
- CData, content, social connections, and reputation
- DA single push-notification opt-in
17. A user has spent years building hundreds of professional connections and a detailed profile inside an app. Which Hook Model stage does this represent?
- AAction
- BVariable Reward
- CTrigger
- DInvestment
18. Which Duolingo feature is the clearest example of accumulated investment?
- AThe Super Duolingo upgrade prompt
- BThe hearts (lives) system
- CStreak count
- DPush notifications
19. How does the Investment stage keep users coming back over time?
- AEach investment makes the product more valuable next use and raises switching costs
- BIt happens before the user reaches the Aha Moment
- CIt sends the highest volume of notifications
- DIt lowers the cost of paid acquisition
20. Investment raises switching costs primarily through which psychological effect?
- AVariable-ratio reinforcement
- BThe endowment effect
- CThe goal gradient effect
- DThe curiosity gap
21. In Adjust's 2025 benchmark, median Day 1 retention across all apps is roughly what?
- A7%
- B50%
- C13%
- D26%
22. In Adjust's 2025 benchmark, median Day 7 retention across all apps is roughly what?
- A7%
- B35%
- C13%
- D26%
23. In Adjust's 2025 benchmark, median Day 30 retention across all apps is roughly what?
- A7%
- B2%
- C13%
- D26%
24. Which correctly ranks Adjust's 2025 median retention figures from highest to lowest?
- ADay 1 26% > Day 7 7% > Day 30 13%
- BDay 1 26% > Day 7 13% > Day 30 7%
- CDay 1 7% > Day 7 26% > Day 30 13%
- DDay 1 13% > Day 7 7% > Day 30 26%
25. A figure attributed to Bain & Company says a 5% rise in customer retention can raise profits by 25–95%. What does it suggest?
- AIt only influences Day 1 numbers
- BIt is far more capital-efficient than paid acquisition
- CIt matters less than lowering CAC
- DIt has no effect on lifetime value
Go deeper
- ReadHooked by Nir Eyal, Ryan Hoover · on Amazon ↗
- StudyThe Science of Sticky Apps in the FlashFeed app
These are affiliate links: if you buy a book through one, FlashFeed may earn a small commission at no extra cost to you. The commission helps pay for FlashFeed's servers.
As an Amazon Associate we earn from qualifying purchases.
Answer key
- 1. D 7 friends in 10 daysFacebook's Aha Moment was '7 friends in 10 days' — reach that social density and the feed becomes meaningful; miss it and the app is an empty room. 'Follow 30 accounts' was Twitter's threshold, and Duolingo's was the first lesson's animated XP reward.
- 2. B The instant the app's core value becomes emotionally realThe Aha Moment is feeling the value, not finishing a tutorial. Completing a walkthrough is explicitly NOT the Aha Moment.
- 3. A Follow 30 accountsTwitter's threshold was following 30 accounts — enough signal to fill a timeline worth returning to. '7 friends in 10 days' was Facebook's.
- 4. C Completing the first lesson and receiving the animated XP rewardDuolingo's is immediate and sensory: the first lesson plus the gamified XP dopamine hit. The account and friend thresholds belong to Twitter and Facebook.
- 5. B Finishing the onboarding walkthroughFinishing a walkthrough is explicitly not the Aha Moment — feeling the value is. The other three are the canonical measurable thresholds.
- 6. C Analyze which early actions correlate with long-term retention in cohort dataTeams analyze which early actions correlate with long-term retention in cohort data, then re-engineer onboarding to deliver every new user to that action sequence.
- 7. D 80%About 80% of users stop after day 3, which reframes onboarding as a survival problem — two or three sessions to prove value.
- 8. A Predicts a user still being active about 30 days laterWhen a behavior reached in session one predicts 30-day retention in cohort data, that behavior is the Aha Moment. Surveys and tutorials do not define it.
- 9. A A survival problem — prove value within two or three sessionsThe day-3 stat reframes onboarding as a survival problem: the app has two or three sessions to prove value before most users are gone. It is not about price, tutorial length, or acquisition.
- 10. B Choosing the last screen of the onboarding tutorialThe Aha Moment is located empirically from cohort data, not by picking a tutorial screen — finishing a tutorial is not the Aha Moment.
- 11. C Send at the same fixed interval every day regardless of contextEffective push follows timing, personalization, and a curiosity/FOMO gap. Arbitrary fixed intervals are exactly what these rules replace.
- 12. D The primary channel for reactivating users who left the habit loopPush is the main re-engagement channel for pulling lapsed users back. Investment is a separate Hook phase, not the push definition.
- 13. B At moments the user has historically opened the appNotifications land better at the user's historical open times. A fixed broadcast interval regardless of context is exactly what NOT to do.
- 14. D It signals the app is paying attention to the userReferencing the user's streak, last action, or goal signals attention, which significantly beats generic copy. It does not replace the timing rule.
- 15. A 'Your streak is at risk'Withholding just enough ('Your streak is at risk') creates a curiosity gap and mild FOMO that motivates a tap. 'Open the app' resolves nothing.
- 16. C Data, content, social connections, and reputationInvestment is the fourth Hook stage where users deposit data, content, connections, and reputation, making the product more valuable next use.
- 17. D InvestmentAccumulated connections and profile detail are deposits the user makes into the product -- exactly what the Hook Model describes as the Investment stage, which raises switching costs via the endowment effect.
- 18. C Streak countDuolingo's streak count is accumulated investment: a 100-day streak is sunk cost that feels devastatingly real. Push notifications are a re-engagement channel; the hearts system and an upgrade prompt store nothing the user has built up.
- 19. A Each investment makes the product more valuable next use and raises switching costsStored investment makes the product more valuable with each use and raises switching costs, so leaving means losing what the user built.
- 20. B The endowment effectThe endowment effect makes users value what they have put in, raising switching costs. The curiosity gap belongs to push notifications, and the goal gradient effect to speeding up near a goal.
- 21. D 26%Adjust's 2025 benchmark puts median Day 1 retention across all apps at 26%. 13% is its Day 7 figure and 7% its Day 30 figure.
- 22. C 13%Adjust's 2025 benchmark puts median Day 7 retention at 13%, between the 26% Day 1 figure and the 7% Day 30 figure.
- 23. A 7%Adjust's 2025 benchmark puts median Day 30 retention at 7%, the lowest of the three checkpoints (Day 1 26%, Day 7 13%).
- 24. B Day 1 26% > Day 7 13% > Day 30 7%Retention decays over time: Adjust's 2025 medians are Day 1 26%, Day 7 13%, Day 30 7%. The other orderings scramble those values.
- 25. B It is far more capital-efficient than paid acquisitionKeeping customers costs far less than winning new ones (HBR, 2014: acquiring a new customer costs five to 25 times more than keeping one), so a small retention gain can lift profits a lot. That makes retention more capital-efficient than paid acquisition.
Sources
- Adjust — What makes a good mobile app retention rate (2025) — www.adjust.com/blog/what-makes-a-good-retention-rate/
- Harvard Business Review — The Value of Keeping the Right Customers (2014) — hbr.org/2014/10/the-value-of-keeping-the-right-customers
- Bain & Company — Prescription for cutting costs, Fred Reichheld (2001) — media.bain.com/Images/BB_Prescription_cutting_costs.pdf
- Nir Eyal — User Investment: Make Your Users Do the Work — www.nirandfar.com/makeyourusersdothework/
- Nir Eyal — The Hooked Model: How to Manufacture Desire in 4 Steps — www.nirandfar.com/how-to-manufacture-desire/
- Dan Silvestre — Hooked by Nir Eyal: Summary and Notes — dansilvestre.com/summaries/hooked-book-summary/
- Amplitude — What Is Cohort Retention Analysis — amplitude.com/explore/analytics/cohort-retention-analysis