Lean Analytics: the one metric that matters right now
Croll and Yoskovitz's data layer for the lean movement — what makes a metric good, the five stages every startup moves through, and how your business model picks your numbers for you.
The theory in one paragraph
The lean loop says build, measure, learn — Lean Analytics is the book about the middle word. Croll and Yoskovitz's core discipline is the One Metric That Matters: at any moment, one number best captures whether your current stage is working, and everything else is context. Which number isn't a matter of taste — it falls out of two coordinates: your business model (marketplace, SaaS, e-commerce, media, mobile app, user-generated content) and your stage. Their five stages run empathy (do people have this problem?), stickiness (do they keep using the solution?), virality (do they bring others?), revenue (will they pay sustainably?), and scale (does it grow efficiently?). Skipping a stage — chasing virality before stickiness, revenue before empathy — is how startups optimize their way into a wall.
How it works
The mechanics — as Alistair Croll defined them, not the folklore version.
What makes a metric good
Four tests: comparative (against last month, a cohort, a competitor), understandable (people remember and discuss it), a ratio or rate (raw totals only go up; rates expose change), and behavior-changing (you know in advance what you'd do differently at different values). 'Total registered users' fails all four; 'week-4 retention by signup cohort' passes them — vanity versus actionable in one example.
The five stages gate each other
Empathy metrics are qualitative — interview counts, problem validation. Stickiness is engagement and retention. Virality is invitation and referral rates. Revenue is unit economics. Scale is efficiency and channel economics. The gating is the insight: viral growth on a leaky product fills a bathtub with the drain open, and revenue optimization before stickiness monetizes people who are already leaving.
Your business model chooses your numbers
A marketplace lives on liquidity and take rate; SaaS on churn, MRR, and CAC payback; e-commerce on repeat purchase rate and basket size; media on engaged time and RPM. The book's model-by-stage grid is why two equally smart teams should track completely different dashboards — and why copying a famous company's KPIs is copying their business model by accident.
The OMTM is a focus device, not a philosophy of ignorance
Teams hear 'one metric' and object that businesses are complicated. The authors' point is about attention, not dashboards: you still collect everything, but the whole team optimizes one number per stage — and crucially, you draw the line in the sand first. Deciding 'we move on when week-4 retention crosses 30%' before the data arrives is what separates learning from rationalizing; the OMTM without a pre-committed threshold is just a favorite number.
The person behind it
Alistair Croll & Benjamin Yoskovitz
Analyst-entrepreneur and product executive · co-authors in Eric Ries's Lean Series
Croll co-founded web-performance company Coradiant and chairs tech conferences; Yoskovitz built and sold startups before running product at scale. Lean Analytics was their answer to what founders kept asking after reading The Lean Startup: measure what, exactly? The book's dozens of benchmark numbers — real conversion, churn, and engagement baselines per business model — made it the reference the movement was missing.
Lineage — The measurement layer of the Lean Startup movement →
How to apply it this week
Each step maps to a field in the North Star Metrics tool — finishing the read means finishing the work.
Classify your business model honestly
Marketplace, SaaS, e-commerce, media, mobile app, or user-generated content — pick the one that describes where money will actually come from, not the one from your pitch deck. Hybrid models pick the engine that must work first.
Locate your stage in the five
The uncomfortable test: what's genuinely unproven? If you can't show people keep using the product, you're at stickiness no matter how good the revenue model looks on paper. Most teams place themselves one stage ahead of their evidence.
North Star Metrics · supporting metricsPick the OMTM the grid points to
Model × stage → metric: SaaS at stickiness = cohort retention; marketplace at revenue = take-rate on repeat transactions; e-commerce at scale = CAC by channel. Run it through the four tests — comparative, understandable, ratio, behavior-changing.
North Star Metrics · your north starDraw the line in the sand
Write the number that means 'this stage is done' — using the book's benchmarks or your market's — and what you'll do if you're still under it in eight weeks. Pre-commitment is the anti-rationalization technology.
North Star Metrics · targetMake the OMTM ambient, review the threshold monthly
One number on the wall, in standup, in the investor update. When it crosses the line, change stages and change the metric — clinging to a conquered metric is how sticky products forget to monetize.
Feeds your Readiness Score · Grow
Build it, don't just read it
The steps above are the North Star Metrics tool's structure. Open it and work through them with your own startup — your readiness score starts building from the first field.
Free account · no card required
See it in the wild
Teardowns from our benchmarks library where this framework is doing real work.
Benchmark teardown
Duolingo
A company that visibly lives at stickiness: streaks and retention cohorts as the OMTM, monetization layered on only after the habit held
Read the teardown
Benchmark teardown
Calo
Subscription food where the honest metrics are churn and repeat rate — pre-committed demand makes the model's numbers legible
Read the teardown
Benchmark teardown
Breadfast
Unit economics as the line in the sand — a q-commerce survivor because the revenue-stage metrics were honest through Egypt's macro storms
Read the teardown
Frequently asked questions
What is the One Metric That Matters (OMTM)?
It's Lean Analytics' central discipline: at any given stage of your startup, one metric best captures whether the current bet is working, and the whole team should focus on moving it. You still track other numbers as context, but the OMTM gets the attention, the target ('line in the sand'), and the decisions.
What are the five stages of Lean Analytics?
Empathy (validating people have the problem), stickiness (proving they keep using your solution), virality (getting users to bring others), revenue (making the economics work), and scale (growing efficiently into new markets and channels). Each stage gates the next — optimizing virality before stickiness spreads a product people abandon.
What makes a good metric versus a vanity metric?
A good metric is comparative, understandable, expressed as a ratio or rate, and changes your behavior — you know in advance what you'd do at different values. Vanity metrics (total signups, page views, cumulative downloads) only ever go up and justify no decision; their actionable twins are rates: retention by cohort, conversion, repeat-purchase percentage.
How do you choose a north star metric for your business model?
Let the model choose: marketplaces track liquidity (transactions, fill rate), SaaS tracks retention and MRR-related ratios, e-commerce tracks repeat purchase rate and basket economics, media tracks engaged attention, UGC tracks content-creation participation. Then adjust for stage — the same SaaS company's OMTM moves from cohort retention to CAC payback as it matures.
What is the difference between Lean Analytics and Lean Startup?
The Lean Startup defines the loop — build, measure, learn — and the philosophy of validated learning; Lean Analytics is the field guide for the measure step, published in Ries's own Lean Series. It supplies what the parent framework assumes: which metric to watch per business model and stage, what benchmark counts as good, and when to declare a stage done.
Related frameworks
North Star Metric
Sean Ellis · 2017
AARRR Pirate Metrics
Dave McClure · 2007
The Lean Startup
Eric Ries · 2011
Sources
Independent educational summary written by StartupKit from public sources. Lean Analytics is the work of Alistair Croll & Benjamin Yoskovitz; this page is not affiliated with or endorsed by the author.