Funnels & Retention
Every product has a funnel, whether you have mapped it or not. Users discover your product, sign up, try it, decide whether to keep using it, pay for it, and tell others about it. At every step, people drop off. Understanding where they drop off — and why — is the difference between growing sustainably and pouring users into a leaky bucket.
The AARRR Framework
Dave McClure's pirate metrics framework, named for its acronym, breaks the user lifecycle into five stages:
Acquisition — How do users find you?
Channels: organic search, paid ads, referrals, content
Metric: Visitors, signups by channel, cost per acquisition
Activation — Do users have a good first experience?
The "aha moment" when they get value for the first time
Metric: % of signups who complete a core action
Retention — Do users come back?
The make-or-break stage for long-term viability
Metric: % of users active in week/month N who return in N+1
Revenue — How do you make money?
Conversion to paid, expansion, average revenue per user
Metric: Trial-to-paid conversion, ARPU, LTV
Referral — Do users tell others?
Organic growth through word of mouth or built-in sharing
Metric: Referral rate, viral coefficient, NPS
The framework is sequential but not linear. Users cycle through retention and revenue repeatedly. And referral feeds back into acquisition, creating a loop.
Where Most Products Leak: Activation
Acquisition gets most of the attention because it is the most visible. Marketing budgets, ad campaigns, SEO strategies — all focused on getting people through the door. But the biggest leak in most products is activation: users sign up and never experience the product's value.
Typical activation problem:
1,000 users sign up this month
600 complete the onboarding flow
300 perform a core action
150 come back in week 2
That's an activation rate of 30% (core action) and only
15% weekly retention. You're losing 70% of users before
they even understand what the product does.
Finding the Aha Moment
The aha moment is the point where a user first experiences real value. It is different for every product:
Product Aha moment
---------------------------------------------------------------------------
Dropbox First file synced across devices
Slack First message exchange with a teammate
Zoom First video call completed without issues
Notion First page shared with a collaborator
GitHub First pull request merged
To find yours, compare users who retained with users who churned. What actions did retained users take in their first session or first week that churned users did not?
Analysis approach:
1. Define "retained" (e.g., active in month 3)
2. Pull all actions from retained users' first 7 days
3. Pull all actions from churned users' first 7 days
4. Compare: which actions have the highest correlation with retention?
Example findings:
- Users who created a project in day 1: 68% month-3 retention
- Users who invited a teammate in week 1: 74% month-3 retention
- Users who only browsed the dashboard: 12% month-3 retention
Aha moment: Creating a project AND inviting a teammate
Once you identify the aha moment, your activation strategy is to get every new user to that moment as fast as possible. Remove friction before it. Defer complexity until after it.
Building a Conversion Funnel
A conversion funnel tracks users through a sequence of steps, measuring the drop-off at each stage.
Step 1: Define the Steps
Map the journey from first touch to desired outcome:
Example SaaS funnel:
Step 1: Lands on homepage
Step 2: Clicks "Sign up"
Step 3: Completes registration form
Step 4: Verifies email
Step 5: Completes onboarding wizard
Step 6: Creates first project
Step 7: Invites a teammate
Step 8: Returns in week 2
Step 2: Measure Each Step
Step Users Conversion Drop-off
-----------------------------------------------------------
Homepage visit 10,000 100% -
Click "Sign up" 2,000 20% 80%
Complete registration 1,400 70% 30%
Verify email 1,100 79% 21%
Complete onboarding 800 73% 27%
Create first project 500 63% 38%
Invite teammate 200 40% 60%
Return in week 2 120 60% 40%
Step 3: Find the Biggest Leak
In this example, two steps stand out: homepage to signup (80% drop-off) and project creation to teammate invitation (60% drop-off). These are your highest-leverage opportunities.
Do not try to optimize every step simultaneously. Fix the biggest leak first. A 10% improvement in the worst step usually has more impact than a 2% improvement across all steps.
Step 4: Diagnose & Fix
For each leak, diagnose the cause:
Homepage to signup (80% drop-off):
Possible causes:
- Value proposition unclear on homepage
- Signup requires credit card (friction)
- Wrong audience visiting (acquisition targeting issue)
Investigation: Session recordings, heatmaps, user interviews
Fix: Clearer headline, remove credit card requirement for trial,
add social proof, refine ad targeting
Project creation to teammate invitation (60% drop-off):
Possible causes:
- Users don't realize it's a collaboration tool
- Invitation flow is buried in settings
- Users want to try it solo first before involving others
Investigation: Usability testing, in-app survey, funnel analysis
Fix: Prompt invitation during project creation, show collaborative
features prominently, add "invite team" CTA on dashboard
Retention: The Metric That Matters Most
Retention is the foundation of sustainable growth. You can acquire millions of users, but if they do not stay, you are filling a bathtub with the drain open.
Retention Curves
A retention curve plots the percentage of users who remain active over time, typically by cohort (users who signed up in the same week or month).
Healthy retention curve (flattens):
Week 0: 100%
Week 1: 65%
Week 2: 50%
Week 4: 40%
Week 8: 35%
Week 12: 33%
Week 24: 31%
The curve flattens around 30-35%. This means roughly a third
of users find lasting value. This product has product-market fit.
Unhealthy retention curve (goes to zero):
Week 0: 100%
Week 1: 40%
Week 2: 20%
Week 4: 10%
Week 8: 4%
Week 12: 1%
Week 24: 0%
The curve never flattens. Every user eventually leaves.
This product does not have product-market fit.
If your retention curve goes to zero, no amount of acquisition will save you. Fix the product first.
What "Good" Retention Looks Like
Benchmarks vary dramatically by product category:
Product type Good monthly retention
---------------------------------------------------------------------------
Social/messaging 60-80%
SaaS (B2B) 85-95%
E-commerce 25-35%
Consumer subscription 60-70%
Gaming (casual) 10-20%
Gaming (core) 30-50%
Fitness apps 20-40%
Compare yourself to your category, not to a universal benchmark. A 40% monthly retention rate is excellent for e-commerce and terrible for B2B SaaS.
Types of Retention
User retention: Is the user still active?
(logged in, performed any action)
Feature retention: Is the user still using a specific feature?
(useful for measuring feature success)
Revenue retention: Is the user still paying?
Net revenue retention includes expansion revenue.
>100% NRR means existing customers are growing.
Dollar retention: How much revenue from a cohort is retained?
Accounts for downgrades and expansions.
Cohort Analysis
Cohort analysis groups users by when they started and tracks their behavior over time. It is the most important analytical technique for understanding retention.
Why Cohorts Matter
Aggregate metrics hide trends. If your overall retention is 40%, that could mean:
Scenario A (healthy):
January cohort: 40% retained
February cohort: 40% retained
March cohort: 40% retained
Consistent. No trend.
Scenario B (improving):
January cohort: 30% retained
February cohort: 40% retained
March cohort: 50% retained
Getting better. Product improvements are working.
Scenario C (deteriorating):
January cohort: 50% retained
February cohort: 40% retained
March cohort: 30% retained
Getting worse. Something is breaking.
The aggregate is 40% in all three scenarios, but the stories are completely different. Cohort analysis reveals the truth.
Building a Cohort Table
Week 0 Week 1 Week 2 Week 4 Week 8 Week 12
Jan cohort 1,000 600 420 350 310 290
Feb cohort 1,200 780 560 430 380 -
Mar cohort 1,500 1,050 780 610 - -
Apr cohort 1,800 1,350 1,020 - - -
As percentages:
Week 0 Week 1 Week 2 Week 4 Week 8 Week 12
Jan cohort 100% 60% 42% 35% 31% 29%
Feb cohort 100% 65% 47% 36% 32% -
Mar cohort 100% 70% 52% 41% - -
Apr cohort 100% 75% 57% - - -
This table tells a clear story: each successive cohort retains better than the last. Whatever product changes were made between January and April are working. Read the table diagonally (same calendar week) to see how all cohorts behaved during a specific period, and vertically (same cohort age) to see how retention at each stage changes over time.
Segmenting Cohorts
Do not just segment by time. Segment by behavior, channel, plan, and persona:
Useful cohort segments:
- By acquisition channel (organic vs paid vs referral)
- By onboarding completion (completed vs abandoned)
- By first-week behavior (invited teammates vs solo)
- By plan type (free vs trial vs paid)
- By company size (SMB vs mid-market vs enterprise)
You will often find that aggregate retention hides massive differences between segments. Paid acquisition users might retain at 20% while organic users retain at 50%. This changes your strategy entirely.
Common Pitfalls
- Measuring vanity funnels — tracking steps that do not matter (e.g., "viewed pricing page") while ignoring steps that do (e.g., "completed core action"). Every funnel step should map to a meaningful user behavior.
- Not segmenting cohorts — aggregate retention hides the truth. A blended 40% retention might be 80% for one segment and 10% for another. You need to know which.
- Optimizing acquisition when retention is broken — pouring water into a bucket with a hole in the bottom. Fix the hole first.
- Confusing correlation with causation in aha moment analysis — the fact that retained users invited teammates does not necessarily mean inviting teammates causes retention. It might just indicate a certain type of user. Run experiments to validate.
- Ignoring the activation step — most products focus on acquisition and retention but skip activation. This is where the biggest drop-off happens and where improvements have the most leverage.
- Measuring retention too broadly — "logged in" is a weak definition of active. Define retention around core value delivery: did the user actually do the thing your product exists for?
- Not tracking cohort trends over time — a single retention snapshot tells you nothing about trajectory. Track cohorts over time to see whether things are getting better or worse.
Key Takeaways
- The AARRR funnel (Acquisition, Activation, Retention, Revenue, Referral) is the lifecycle framework for understanding where users drop off and why.
- Activation is the most underinvested stage in most products. Find the aha moment and remove every obstacle between signup and that moment.
- Retention curves reveal product-market fit. If the curve flattens, you have it. If it goes to zero, no amount of acquisition will compensate.
- Cohort analysis is essential. Aggregate metrics hide trends, segment differences, and the true impact of product changes.
- Fix the biggest leak in the funnel first. A 10% improvement at the worst stage beats a 2% improvement everywhere.
- Retention is the foundation. Without it, acquisition spend is wasted and growth is unsustainable.