Summary
Lean Analytics is Alistair Croll and Benjamin Yoskovitz's attempt to give early-stage companies a practical framework for using data to move faster. It sits in the tradition of Lean Startup thinking — the idea that the job of an early company is not to execute a plan but to learn what plan to execute, as quickly and cheaply as possible. The book's contribution is to make that process more specific: what exactly should you be measuring, when, and why?
The book's core framework is the One Metric That Matters (OMTM) — the single number that best captures the current state of your business at any given stage. The key word is "current": the right metric changes as a company moves through stages. An early-stage marketplace should obsess over activation rates. A growth-stage SaaS company should obsess over churn. A company trying to monetize should focus on revenue per user. Using the wrong metric — vanity metrics like raw signup numbers that feel good but don't predict survival — is how companies run in circles.
Croll and Yoskovitz map six business models (e-commerce, SaaS, mobile, media, user-generated content, two-sided marketplace) against five stages of startup development (Empathy, Stickiness, Virality, Revenue, Scale). For each combination, they suggest which metrics matter most. The resulting matrix is the book's most referenced artifact: a cheat sheet for figuring out what to measure given your model and where you are. It is prescriptive and opinionated, which makes it useful even when you disagree with specific suggestions.
The book's weakness is its age. Published in 2013, it predates the maturation of product analytics tools, the dominance of mobile-first products, and the shift in growth tactics that followed. Some specific benchmarks and examples are dated. But the underlying framework — focus on one metric, know your stage, distinguish leading from lagging indicators, design experiments rather than just tracking dashboards — remains sound. Lean Analytics is best read alongside current practitioners who have updated the specific benchmarks.
Key takeaways
- 1.
The One Metric That Matters: at any given stage, there is one number that best captures whether the business is working. Measuring many things without prioritizing one leads to analysis paralysis.
- 2.
Vanity metrics (page views, total signups, downloads) feel good but don't predict survival. Actionable metrics (activation rate, retention, revenue per user) reveal whether the business model actually works.
- 3.
The right metric changes at every stage. Early on, measure whether people care (engagement, activation). Later, measure whether the engine is scaling (virality, unit economics).
- 4.
A good metric is comparative: not '10,000 signups' but '10,000 signups, up 15% week-over-week.' Change over time is what generates learning.
- 5.
Every business model has a defining metric. SaaS companies live and die on churn. Marketplaces on liquidity. E-commerce on repeat purchase rate. Knowing your model tells you what to obsess over.
- 6.
The Empathy stage precedes all quantitative analytics. Before measuring, you need to know you are solving a real problem for a real person. Qualitative research first, then metrics.
- 7.
Dashboards full of green metrics and a declining business are common. The question is not whether your metrics look good but whether they predict the health of the model.
- 8.
Experimentation requires a hypothesis. 'We changed the button color' is an observation. 'We changed the button color because we hypothesize it will increase activation by reducing cognitive friction' is an experiment.
Discussion questions
Use these on your own, with a book club, or as chat starters in Superbook.
- 1.
What is the One Metric That Matters for the business or project you're currently focused on? If you can't name one number, what does that tell you?
- 2.
Think of a metric you track that you suspect is a vanity metric. What would you replace it with that is more directly tied to value delivered?
- 3.
Croll and Yoskovitz say the right metric changes at each stage. At what stage is your current project, and does your measurement focus match that stage?
- 4.
The book distinguishes between leading indicators (metrics that predict future performance) and lagging indicators (metrics that confirm past performance). Which kind does your team primarily track?
- 5.
Have you ever been in a situation where the dashboard was green but the business was declining? What was the gap between the metrics you tracked and what was actually happening?
- 6.
They argue for qualitative research before quantitative metrics. How often do you actually talk to users before designing experiments or interpreting data?
- 7.
Their business model matrix is prescriptive. Does your business fit cleanly into one of their six models, or does it straddle multiple ones? How does that complicate the framework?
- 8.
What is the difference between a dashboard and an insight? When was the last time your data actually changed a decision?
- 9.
The book was published in 2013. Which parts feel dated to you, and which principles seem timeless regardless of which tools or platforms are current?
- 10.
Lean Analytics assumes you can run controlled experiments. What structural barriers — organizational, technical, or market-related — make experimentation hard in your context?
- 11.
They distinguish between 'good enough' benchmarks and 'world class' benchmarks for each metric. Knowing the benchmark changes what counts as progress. What benchmarks do you use in your field, and where did they come from?
- 12.
If you could only track three metrics for the next ninety days in your current work, what would they be and why?
Themes
Frequently asked questions
-
What is the One Metric That Matters?
Croll and Yoskovitz's term for the single number that best represents whether your business is healthy at its current stage. The concept is not that you track only one metric, but that you prioritize one above all others and let it drive decisions. The right OMTM changes as you move through stages.
-
Is Lean Analytics still relevant in 2026?
The framework is sound; some benchmarks and examples are dated. The core ideas — OMTM, stages of startup development, distinguishing vanity from actionable metrics, designing experiments — remain useful. Read it alongside current writing on product analytics to update the specific numbers.
-
Who is this book for?
Founders, product managers, and growth teams at early-stage companies who want a more disciplined approach to measurement. It is most useful for people who already have some data but aren't sure what to focus on or how to use it to make decisions.
-
What are vanity metrics?
Metrics that look impressive but don't actually predict whether your business will succeed — raw signup counts, total downloads, page views without engagement context. They feel good and are easy to show stakeholders, but they don't tell you whether people are getting value or whether you should change anything.
-
How does Lean Analytics relate to Lean Startup?
Lean Analytics is an extension and operationalization of Eric Ries's Lean Startup framework. Where Lean Startup describes the Build-Measure-Learn loop in principle, Lean Analytics is primarily about the Measure and Learn steps — specifically what to measure and how to draw valid conclusions from it.
