📈Metrics 101 for PMs
AARRR, funnels, leading vs lagging — the basic vocabulary of measuring a product. Master this in week 1.
PMs are evaluated on the metrics they move. If you can't speak fluently about acquisition, activation, retention, revenue, and referral — or distinguish leading from lagging indicators — you'll lose credibility with engineering, sales, and execs alike.
Every product has a small set of metrics that actually matter — the rest are noise. PM job: identify the North Star, build the funnel that shows where the journey breaks, instrument it well, and watch the leading indicators that predict the lagging ones.
The metric vocabulary
Acquisition (A). How many people land in the product. Sources, channels, CAC.
Activation (A). How many of those land users hit the "aha" moment — the first experience of the product's value. Time-to-first-value, activation rate.
Retention (R). How many keep coming back. D1, D7, D30 retention. Cohort curves.
Revenue (R). How many start paying, how much they pay, expansion. MRR, ARPU, LTV.
Referral (R). How many bring others. Viral coefficient (K-factor), referral rate.
Dave McClure's "AARRR" pirate metrics is the classic SaaS / consumer framework. Useful as scaffolding even if you don't apply it religiously.
Leading vs lagging
Lagging indicators are the outcomes (revenue, retention, NPS). Easy to see, hard to influence quickly. You learn from them but you can't steer with them.
Leading indicators are the behaviors that predict the outcomes. (For Slack: weekly active days per user. For Dropbox: files synced in week 1. For Shopify: stores reaching $1K in sales.) You can influence these now and the lagging follow.
Great PMs identify the 2-3 leading indicators that predict their North Star and instrument the team around those. Tracking only lagging is steering by the rearview mirror.
The funnel
Every product has a journey from "saw an ad" to "evangelist." Map it as a funnel. Measure the conversion rate at each step. The biggest drop-off is usually the highest-leverage thing to fix.
Funnel for SaaS sign-up:
- 1000 land on the homepage
- 200 click 'sign up' (20% conversion)
- 120 complete sign-up (60% completion)
- 50 complete onboarding (42%)
- 25 active at day 7 (50%)
- 10 active at day 30 (40%)
- 4 paying at day 90 (40%)
The lever isn't always at the top — often it's at activation or D7 retention. Look for the worst conversion rate.
North Star Metric
One metric that, if it grows, the business is succeeding. For Airbnb: nights booked. For Spotify: time spent listening. For Slack: messages sent in active workspaces. The North Star is the one number everyone in the company can recite.
It should be (a) measurable, (b) tied to customer value, (c) a leading indicator of revenue.
How to instrument
Before you ship anything: define the metrics. Define the events. Define the dashboard. Engineering implements during the build. The shipped feature should be observable from day 1.
The pattern that fails: ship first, instrument later. By the time you realize you're missing an event, weeks of usage data are unrecoverable.
The metric tree
Build a "metric tree" — North Star at the top, input metrics below, drivers below those. Every team in the company should be able to point to the leaf they're moving and how it rolls up.
Key frameworks
Acquisition, Activation, Retention, Revenue, Referral. The default funnel framework.
One metric the whole org rallies around. Captures customer value and is a leading indicator of revenue.
Leading = behaviors you can influence now. Lagging = outcomes. Steer with leading.
Real-world examples
Airbnb famously chose 'Nights Booked' as their North Star — not revenue, not bookings, not visits. It captures both guest value (they found a place to stay) and host value (they got a booking). Every team's KRs ladder up to nights booked.
Slack discovered early that workspaces that sent 2,000 messages had ~90% retention, vs. much lower below that threshold. The 2000-message threshold became their activation metric — and the entire onboarding was designed to drive workspaces past it.
Go deeper — recommended reading
Interview questions (2)
Q1Walk me through how you'd choose a North Star Metric for a new product.metricsmid▼
Three criteria for a good North Star:
- Captures customer value. Not vanity. If this metric grows, customers are getting more value from your product.
- Leads revenue. It should correlate with and precede revenue growth.
- Singular and memorable. The whole company can recite it.
Process:
- Map the customer journey. What's the moment of value delivery? (For Airbnb, that's the night spent at a host's place. For Slack, the message sent.)
- Quantify it. "Nights booked per quarter," "weekly active workspaces sending >50 messages."
- Pressure-test: does this metric have an obvious failure mode? (E.g., 'time on page' is a bad North Star because confusion increases it.)
- Validate the leading-to-lagging relationship with 6+ months of historical data.
The non-obvious move: pick the metric that both sides of your marketplace or both buyer and user benefit from. That alignment prevents the metric from being gamed at one side's expense.
Q2Your engagement metric is up 15% but revenue is flat. What's happening?metricssenior▼
Five hypotheses to investigate, in order of likelihood:
- The new engagement is from non-paying users. Free users are using more; paid users are flat. Slice by segment.
- The engagement is shallow. Sessions went up but session quality (depth, completion) went down. Check leading indicators of value.
- The engagement is in a non-monetized feature. Users discovered a feature that doesn't drive revenue. Cool for retention long-term, but not your revenue lever.
- Pricing/packaging is broken. Users get value but the offer doesn't capture it. Look at conversion from free to paid.
- The engagement metric is the wrong North Star. If 15% lifts don't move revenue, the metric isn't truly leading.
The diagnostic: slice the engagement metric by paying status, by feature, by cohort. The pattern usually emerges within a day of careful analysis.