North Star & the AARRR funnel — measure what matters
🎯 Goal: Choose ONE North Star metric that reflects the real value customers get, and read growth health through the AARRR funnel to see where customers drop and fix the right spot.
The North Star Metric is a single number reflecting the real value customers receive — not a "vanity metric" like downloads or likes. When the whole team aims at one North Star, decisions stay consistent instead of each department optimizing its own thing. Beneath it sits the AARRR funnel (Dave McClure’s "Pirate Metrics"): Acquisition → Activation → Retention → Referral → Revenue. You’ll leave with a funnel table that shows where customers drop so you fix the real leak instead of guessing.
Lesson content
🎯
A North Star ties to the MOMENT the customer gets value. Airbnb picks nights booked, Spotify picks time spent listening, Grab picks completed rides, early Facebook picked friends connected within 10 days. Common thread: if that number rises, customers are genuinely better off and the business is more durable. By contrast "app downloads" is vanity — download then delete still counts. A good North Star test: (1) reflects customer value; (2) predicts future revenue; (3) the team can move it through daily actions.
🧭
The AARRR funnel — 5 stages, top to bottom: • Acquisition: people discover you — ask "which channel, at what cost?". • Activation: they hit a first "wow" — "what % reach value on first use?". • Retention: they come back — "after 7/30 days, what % is still active?". • Referral: they invite others — "how many new users does each bring?". • Revenue: they pay — "revenue per customer, paying rate". Read the funnel to find the worst leak, then fix it first — clearing one bottleneck often lifts the whole funnel.
🧰
TOOL — AARRR funnel table (each stage gets one question + one measurable metric):
Stage
Core question
Metric
Acquisition
Where do customers come from?
Visits, cost per customer (CAC)
Activation
Do they hit a "wow" first time?
% completing first step, real-use rate
Retention
Do they come back?
% still active after 7/30 days
Referral
Do they invite others?
Viral coefficient K, invites sent
Revenue
Do they pay?
Revenue per customer, paying rate
Tip: don’t just watch the final number. The stage with the biggest drop-off is where to fix first.
🌏
Case study — MoMo & Coolmate (step by step): 1) MoMo uses a North Star tied to successful transactions per month — each transaction is one moment of real value (convenient payment). 2) High frequency builds a daily app-opening habit → strong Retention → cross-selling many services. 3) Coolmate could use repeat customers as its North Star (estimated 50%+ return) — reflecting real satisfaction, not one-off buyers who vanish. 4) For an online shop, the funnel usually "leaks" worst at Retention: it pulls people in but fails to make them buy again — and patching Retention is far cheaper than endlessly buying new customers.
⚠️
4 traps with metrics: • Vanity metrics: bragging downloads/likes/views — loud but says nothing about value received. • Measuring too much: 30 metrics means none drives decisions; have ONE North Star. • A North Star detached from customer value: if maximizing it hurts customers (e.g., forcing ads), it’s the wrong metric. • Watching totals, ignoring the funnel: revenue rises but you don’t know which stage did it, and can’t tell where to patch when it falls.
✅
Checklist to lock a North Star & read the funnel: ① Does this number rise when customers are better off? ② Does it predict long-term revenue? ③ Can the team move it through daily work? ④ Have I attached a metric to each AARRR stage? ⑤ Which stage drops most — will I fix it first?
Practice exercise
🔬 APPLIED EXERCISE: (1) Write one North Star for your idea — it must reflect real value, NOT downloads/likes. (2) Fill the AARRR funnel: one metric per stage. (3) Guess which stage "leaks" worst and why. (4) Note one thing you’ll do next week to lift that stage.
Worked template: FILLED EXAMPLE — a vocabulary app: North Star = "words a learner actually remembers after 7 days" (not downloads). Funnel: Acquisition = installs; Activation = % finishing lesson one; Retention = % still learning after 7 days; Referral = friends invited; Revenue = % buying Pro. Guessed leak: Retention — many install but quit after 2 days → next week add a 3-min daily reminder to build the habit.
Quick quiz
1. What should the North Star Metric reflect?
→ The real value customers receive
The North Star must reflect real value customers receive, not a vanity metric like downloads/likes.
2. Why is "app downloads" usually a vanity metric?
→ Because it counts even if users delete right after, showing nothing about value received
Downloads are a vanity metric: they count even after deletion, showing nothing about value received.
3. In the AARRR funnel, which question does "Retention" answer?
→ Do customers come back to use it again?
In AARRR, Retention answers "do customers come back to use it again?" — core to durable growth.
4. A shop pulls in many customers but few buy a second time. By AARRR, which stage should it fix first?
→ Retention — patch the keep-them stage first, usually cheaper than buying new customers
Pulling customers in but few buying twice is a Retention leak; patching it is usually cheaper than buying new customers.
The funnel order is Acquisition → Activation → Retention → Referral → Revenue; the right order reveals the leak.
6. What signals you picked a GOOD North Star?
→ When it rises, customers are better off and it predicts future revenue
A good North Star is a number that, when it rises, means customers are better off and it predicts future revenue.
Advanced
A deeper framework
At an advanced level, don’t just watch the total — compute the conversion rate between stages of the AARRR funnel. Final revenue = starting users × (rate through each stage). Because these rates multiply, one steep drop chokes the whole funnel even if other stages are great.
Prioritization trick: find the stage with the biggest drop-off and lift it, because improving the bottleneck beats pouring more users into an already-fine top of funnel. This is why we "fix Retention before raising Acquisition".
Reading an AARRR funnel with numbers
Acquisition: 10,000 arrivals
CAC ~15,000₫ each
Activation: 40% reach first value
4,000 "wow" users
Retention after 30 days: 25%
~1,000 still active
Revenue: 8% pay
~320 paying customers
The worst drop is Activation→Retention (4,000 down to 1,000). Lifting this segment beats pouring more money into Acquisition.
Common trap: A "growth illusion" from vanity metrics: bragging 1 million downloads while 30-day Retention is only 2% — meaning almost nobody got value. A big top-line can hide a leaky funnel; always inspect retention rates, not just the number at the top.
Advanced questions
1. Your team brags "the app hit 1 million downloads" but only 2% are still active after 30 days. How to read the real health, and what should the North Star be?
→ Worrying: downloads are a vanity metric; the North Star should tie to real value like % still active/returning
1M downloads but 2% retention means almost nobody got value. Downloads are a vanity metric; the North Star must reflect real value (active/returning users), else the whole team optimizes the wrong thing.
2. Funnel: 10,000 in → 40% activation → 25% retention → 8% pay. Which stage most deserves lifting?
→ The Activation→Retention segment, the steepest drop, whose lift amplifies the whole funnel
Rates multiply, so the steepest drop chokes the funnel. From 4,000 "wow" down to ~1,000 is the bottleneck — lifting retention here beats adding more users at the top.
3. Why usually "fix Retention before raising Acquisition"?
→ Because if Retention leaks, every acquired customer falls out — buying more pours into an empty bucket
Leaky Retention turns Acquisition into pouring water into an empty bucket. Patching retention first is cheaper and makes every later Acquisition dollar more profitable.
🎯 Real-life mission
REAL-LIFE MISSION: Pick one product you use daily (app, shop, service). Guess its North Star (a metric reflecting real value, NOT downloads), then map its AARRR funnel: for each stage note one metric they could measure and guess which stage "leaks" worst. Explain why, and name one thing they should do to patch that stage.