Lean Analytics by Croll and Yoskovitz — Cliff Notes Summary

Direct Answer

Lean Analytics: Use Data to Build a Better Startup Faster by Alistair Croll and Benjamin Yoskovitz (O'Reilly, 2013) is the operating manual that turned Eric Ries's Lean Startup theory into a working metrics discipline. Croll (Year One Labs partner, co-founder of the Strata conference) and Yoskovitz (Highline Beta partner, former GoInstant VP Product) argue that most startups track the wrong metrics, and the right metric depends on two variables — your business model and your startup stage.

Their signature contribution is the One Metric That Matters (OMTM): at every stage, pick the single number that matters most, rally the team around it, and ignore everything else until it's solved. The book also delivers the Bullshit-Metric Detector (Vanity vs. Actionable), the Five-Stage progression (Empathy → Stickiness → Virality → Revenue → Scale), and the Six Business Model Frameworks (e-commerce, SaaS, free mobile app, media, user-generated content, two-sided marketplace) — each with stage-appropriate metric stacks and Lines in the Sand benchmarks for "good enough." It is the missing instrumentation layer between Ries's Lean Startup (2011) and Ash Maurya's Running Lean (2010), and it remains the canonical reference for modern growth and PLG teams using Amplitude, Mixpanel, Pendo, and Heap.

1. Part One — Stop Lying To Yourself (Chapters 1-4)

1.1 Chapter 1 — We're All Liars

Croll and Yoskovitz open with the founder's hardest truth: every startup founder has a story they tell themselves about why their startup is working. The data usually disagrees. The chapter frames the entire book as a tool for catching yourself in the lie before the market does it for you.

The authors borrow from Eric Ries's "Innovation Accounting" but immediately go operational: which numbers, in which order, at which stage.

The opening case study is Airbnb's 2009 pivot — the founders had a marketplace with listings but no bookings. The OMTM was not "users," it was "bookings per active listing," and that single number forced the famous professional-photography intervention that re-priced and re-listed hosts in New York.

The right metric forced the right action.

1.2 Chapter 2 — How To Keep Score

The authors introduce the four properties of a good metric:

1. Comparative — it tells you whether you're better or worse than yesterday or another segment.
2. Understandable — everyone on the team can recite it from memory.
3. A ratio or rate — absolute numbers lie; ratios (conversion %, churn %, ARPU) survive growth.
4. Behavior-changing — looking at it actually causes someone to do something different tomorrow.

Metrics that fail any of these four tests are noise.

1.3 Chapter 3 — Deciding What To Do With Your Life

This is the chapter that introduces the One Metric That Matters (OMTM) — the signature Croll/Yoskovitz contribution. Verbatim: "The One Metric That Matters is your rallying point — pick it + ignore everything else." The argument is brutally simple: a team chasing five metrics chases none of them.

A team chasing one metric ships. The OMTM is chosen, not discovered — leadership picks it deliberately for the current stage and re-picks when the stage changes.

1.4 Chapter 4 — Data-Driven Versus Data-Informed

The authors push back on naive empiricism. Data-driven companies let the number make the decision; data-informed companies let the number sharpen a judgment call humans still own. The book recommends data-informed — partly because early-stage data is noisy, partly because some decisions (mission, ethics, hiring) are not metric problems.

2. Part Two — The Bullshit-Metric Detector (Chapters 5-6)

2.1 Chapter 5 — Vanity Versus Actionable Metrics

The chapter every startup founder needs printed on their wall. Croll and Yoskovitz: "Vanity metrics always go up — actionable metrics drive decisions." Examples of vanity: total registered users, total page views, total downloads, social-media followers. They only move one direction (up), they hide problems (churned users still count as registered), and they don't inform a next action.

Examples of actionable: activation rate (signups who reach the "aha" moment), D1/D7/D30 retention, conversion rate by funnel step, cohort revenue retention. These ratios surface problems early and tell you exactly which lever to pull.

2.2 Chapter 6 — The Discipline of "One Metric"

The authors walk through how to operationalize OMTM: weekly all-hands review, dashboard-front-and-center, every team OKR rolling up to it, hiring questions tied to it. The chapter's named case study is Localmind (Yoskovitz's prior company) — their OMTM was "questions answered within five minutes," which determined supply-side liquidity in a Q&A marketplace.

Everything from push-notification timing to onboarding copy to recruiting moderators was justified by that one number.

3. Part Three — The Five Stages of a Startup (Chapters 7-12)

The book's spine. Every startup, regardless of business model, moves through five sequential stages — and the OMTM changes at each one.

3.1 Chapter 7 — Stage One: Empathy

Before product, before metrics, you need empathy — qualitative evidence that a real customer has a real problem worth paying to solve. Metrics here are interviews completed, problem-validation rate, and solution-fit signal strength. This stage is inherently qualitative; trying to A/B-test your way through Empathy is the most common founder mistake.

3.2 Chapter 8 — Stage Two: Stickiness

Once you have a product, the OMTM is engagement and retention, not growth. "Don't pour water into a leaky bucket." Stickiness metrics: DAU/MAU ratio, D30 retention curve, session frequency. The line in the sand: a healthy consumer app holds 30%+ at D30; a healthy SaaS holds 90%+ monthly logo retention.

Until you clear those bars, virality and revenue spend is wasted.

3.3 Chapter 9 — Stage Three: Virality

Only after stickiness do you turn on virality — the rate at which existing users bring new users. The OMTM is the viral coefficient (K-factor) — invites sent per user multiplied by acceptance rate. K > 1 means exponential growth from existing users alone; K between 0.5 and 1 is "viral assist" to paid acquisition; K < 0.5 means viral is a rounding error and you need a different growth motion.

Hotmail's P.S. Signature (K ≈ 0.4 in the late 1990s, foundationally documented by Adam Penenberg's *Viral Loop*) and Dropbox's referral program (K-factor near 0.5 with 35% activation lift) are the canonical examples.

3.4 Chapter 10 — Stage Four: Revenue

The OMTM shifts to revenue per user and CAC vs. LTV ratio. The line in the sand: LTV ≥ 3× CAC, with CAC paid back within 12 months for SaaS, 6 months for transactional commerce.

Below those benchmarks, every dollar of growth spend destroys value. The chapter introduces negative churn — the threshold where expansion revenue from existing customers exceeds churn losses, the holy grail of SaaS economics now codified as Net Revenue Retention (NRR) > 100%.

3.5 Chapter 11 — Stage Five: Scale

Now the OMTM becomes channel/segment economics — which acquisition channel, which customer segment, which geography returns capital efficiently at volume. The discipline shifts from "find product-market fit" to "find channel-market fit." Companies that nail Stage 5 (HubSpot's inbound playbook, Salesforce's field-sales motion, Shopify's partner ecosystem) become category leaders.

Companies that try to scale without finishing Stickiness or Revenue die spectacularly — the authors cite Color Labs (which raised $41M before validating stickiness) as the cautionary tale.

3.6 Chapter 12 — The Stage Gate

The authors are explicit: you do not move to the next stage until the current stage clears its Line in the Sand. Pouring virality fuel on a leaky-bucket product wastes money. Pouring revenue effort on a product no one returns to wastes more. The stage gate is the disciplined version of the OMTM.

4. Part Four — The Six Business Model Frameworks (Chapters 13-18)

The book's most-photocopied chapters. Croll and Yoskovitz argue "the right metric depends on your business model + your stage." They lay out six canonical models — each with its own metric stack.

4.1 Chapter 13 — E-Commerce

Stage OMTMs: Empathy → completed purchases; Stickiness → repeat purchase rate (loyalty model) or conversion rate (transactional model); Virality → referral-driven order share; Revenue → Customer Lifetime Value (CLV) and contribution margin per order; Scale → channel-by-channel CAC.

The chapter splits e-commerce into three sub-models — Loyalty (Amazon), Acquisition (Wayfair), and Hybrid — each with different metric weights. The line in the sand for healthy e-commerce: 2-3% baseline conversion rate, 30%+ repeat purchase rate within 90 days.

4.2 Chapter 14 — SaaS

Stage OMTMs: Empathy → signup rate from cold landing pages; Stickiness → activation rate (signups who reach the aha moment) and monthly retention; Virality → seat expansion within accounts; Revenue → MRR, ARPU, gross logo churn < 2% monthly; Scale → Net Revenue Retention (NRR) > 110%, CAC payback < 12 months, Rule of 40.

The SaaS chapter is the most cited and the metric stack here is now the default operating dashboard for every venture-funded B2B SaaS company.

4.3 Chapter 15 — Free Mobile App

Stage OMTMs: Empathy → installs from organic search; Stickiness → DAU/MAU ratio (>20% is healthy, >50% is exceptional, see Facebook's 65% peak), D1/D7/D30 retention; Virality → invite acceptance rate; Revenue → ARPU and ARPPU (average revenue per paying user — typically <5% of users pay in F2P games); Scale → LTV per install vs.

Paid-install CPI. The chapter draws heavily on Zynga's 2010-2012 data and Supercell's Clash of Clans (sustained $5M/day revenue at peak driven by 2% paying-user share).

4.4 Chapter 16 — Media

Stage OMTMs: Empathy → session depth; Stickiness → time on site and return visit rate; Virality → social shares per article; Revenue → CPM (cost per thousand ad impressions), fill rate, viewability; Scale → revenue per session across channels. The line in the sand: a healthy ad-supported media business needs $3-10 CPM with >50% fill rate and >70% viewability.

Below those, programmatic economics collapse.

4.5 Chapter 17 — User-Generated Content

Stage OMTMs: Empathy → contributor acquisition; Stickiness → engagement funnel (visit → read → comment → contribute → moderate); Virality → contributor-driven invite rate; Revenue → ad revenue per contributor; Scale → community health score. The chapter introduces the 90-9-1 rule (90% lurk, 9% contribute occasionally, 1% contribute heavily) and uses Wikipedia, Reddit, and Stack Overflow as the named case studies.

4.6 Chapter 18 — Two-Sided Marketplaces

Stage OMTMs: Empathy → supply-side and demand-side problem validation; Stickiness → liquidity (probability a buyer's search finds a seller match within N minutes); Virality → cross-side referral; Revenue → take rate and GMV growth; Scale → demand-match rate by category and geography.

The chapter draws on Airbnb, eBay, Etsy, and Uber — and the framing of "supply liquidity as the OMTM" is the direct intellectual ancestor of Bill Gurley's liquidity-first marketplace investing thesis and Lenny Rachitsky's modern marketplace playbook.

5. Part Five — Lines in the Sand and Putting It Into Practice (Chapters 19-30)

The back half of the book is operating manual — how to actually run a metrics-driven startup week by week.

5.1 Chapters 19-22 — Lines in the Sand (Industry Benchmarks)

Croll and Yoskovitz publish numerical benchmarks for "good enough" at each stage of each business model. Examples: SaaS monthly churn < 2%, e-commerce conversion > 2%, mobile D30 retention > 15%, marketplace take rate 5-20%, media CPM $3-10. These benchmarks were the first systematic public dataset of their kind and are still cited verbatim in Reforge programs and a16z's SaaS metrics primers.

5.2 Chapters 23-26 — Selling Inside the Enterprise

The authors detour into how to apply Lean Analytics inside large companies — intrapreneur territory. Metrics here are framed as proof-of-concept thresholds that unlock the next budget cycle. The case studies include IBM's internal startup model and Intuit's "Design for Delight" methodology.

5.3 Chapters 27-30 — Conclusion and Getting Started

The final chapters compress the entire framework into a 100-day startup operating plan: Day 1-30 establish OMTM and Lines in the Sand; Day 31-60 instrument and report weekly; Day 61-90 run experiments against the OMTM; Day 91-100 stage-gate review. The recommendation: a one-page weekly metrics review with the OMTM at the top, the next-most-important three metrics underneath, and a written "what changed and why" commentary.

6. Frameworks at a Glance

The Lean Analytics frameworks that travel directly into modern growth and PLG operations:

• The One Metric That Matters (OMTM) — pick one number, rally the team, ignore the rest. Now the spine of Reforge's growth curriculum and the North Star Metric discipline taught by Sean Ellis and Brian Balfour.
• The Bullshit-Metric Detector (Vanity vs. Actionable) — kills bad dashboards. Embedded in Amplitude and Mixpanel activation analytics by default.
• The Five Stages (Empathy → Stickiness → Virality → Revenue → Scale) — the canonical startup progression. Y Combinator and First Round Capital stage-gate guidance maps onto this directly.
• The Six Business Model Frameworks — the metric stack for e-commerce, SaaS, free mobile, media, UGC, and marketplaces. Still the reference for product-team onboarding.
• Lines in the Sand — numerical benchmarks for "good enough" at each stage. Cited verbatim in a16z SaaS metrics primers and OpenView PLG benchmarks.
• The Stage Gate — do not advance until the current stage's OMTM clears the line. The discipline that prevents premature scaling — the failure mode Steve Blank named as the #1 startup killer.

7. What Holds Up, What Has Aged

What still holds (2025-2027):

• OMTM remains the single most-adopted metrics discipline in modern growth — every Series A through pre-IPO company runs a North Star Metric process that is functionally OMTM under a different name.
• The Vanity vs. Actionable distinction is now standard in every product-analytics tool (Amplitude, Mixpanel, Pendo, Heap default dashboards specifically downplay vanity counts).
• The Six Business Model Frameworks are still the cleanest taxonomy in the industry. Every PLG primer and SaaS benchmark report uses some variant of this split.
• The Stickiness-before-Virality sequencing has been re-validated repeatedly by Sean Ellis's Hacking Growth research and Reforge's retention-first curriculum.
• Lines in the Sand benchmarks still anchor industry reports from a16z, OpenView, Bessemer, and SaaS Capital, though the specific numbers have shifted (modern SaaS NRR targets are 120%+ for elite, not 110%+).

What has aged:

• The book underweights product-led growth (PLG) as a distinct motion — the term barely existed in 2013. Modern PLG metrics (PQL-to-paid conversion, expansion via seat-based upsell, free-to-paid time) need to be slotted onto the framework but aren't named there.
• The 2013 mobile app chapter is dated — F2P gaming economics have evolved (subscriptions overtaking IAP, Apple's ATT crushing attribution, sub-2% paying-user economics now require category-level scale).
• The media chapter underestimated the collapse of programmatic CPMs and the rise of subscription bundling (Substack, The Athletic, Patreon) and creator-economy economics.
• The instrumentation prescriptions are pre-AI. Modern teams use AI-driven analytics tools (Amplitude AI, Mixpanel's Spark, Hex's AI notebooks) to automate the metric tracking and surface anomalies without a human writing SQL. ChatGPT and Claude now also help PMs reason about which OMTM to choose at a given stage — a meta-layer the original book could not anticipate.
• The lineage from Lean Analytics extends directly into modern PLG canon: Sean Ellis's *Hacking Growth* (2017), Brian Balfour's Reforge content, Elena Verna's PLG essays, and Kyle Poyar's Growth Unhinged newsletter — all of which build on the OMTM/stage/business-model triangle.

FAQ

Is Lean Analytics worth reading after Lean Startup and Running Lean? Yes — the three compose. Ries gives the philosophy, Maurya gives the canvas, and Croll/Yoskovitz give the instrumentation. Lean Analytics is the only one of the three that gives you specific metrics to track at specific stages.

Does the OMTM mean we should only track one metric? No — track many, focus on one. The OMTM is the team's rallying point and the metric you optimize first. Supporting metrics still get reviewed weekly; they just don't get the team's morning energy.

How does this apply to a B2B sales org? Directly. Pick the OMTM at each pipeline stage — top of funnel = conversations created; middle = qualified opportunities generated; bottom = closed-won revenue; expansion = NRR. The vanity-vs-actionable test kills most sales dashboards (lead count, MQLs, demos booked are mostly vanity; SQL-to-opportunity conversion, opportunity-to-close, and NRR are actionable).

Are the Lines in the Sand benchmarks still accurate? Directionally yes, numerically partially. SaaS churn targets are tighter (<1% monthly for elite vs the book's <2%), NRR targets are higher (120%+ vs 110%+), e-commerce conversion is roughly the same. Use the book's framework, refresh the numbers from current OpenView and a16z benchmark reports.

Should I read this if I'm a growth PM or a sales leader, not a founder? Yes — arguably more useful for operators than founders. The OMTM and stage-gate discipline applies to any function (growth, sales, customer success) that needs to focus a team around the one number that moves the business.

Bottom Line

Read this book if you've ever stared at a fifty-metric dashboard and felt no closer to knowing what to do. Lean Analytics gives you the discipline to pick one number, defend it, and build a team rhythm around it — and the Six Business Model Frameworks plus Lines in the Sand give you the specific number to pick for your specific business.

It is the single best operating reference for any founder, growth PM, or revenue leader who needs to move from "we have lots of data" to "we know what to do tomorrow." Thirteen years after publication, OMTM is still the most-stolen idea in growth — which is the highest possible compliment a metrics book can earn.

Sources

• Croll, Alistair & Yoskovitz, Benjamin — *Lean Analytics: Use Data to Build a Better Startup Faster* (O'Reilly, 2013)
• Ries, Eric — *The Lean Startup* (Crown Business, 2011)
• Maurya, Ash — *Running Lean* (O'Reilly, 2010)
• Ellis, Sean & Brown, Morgan — *Hacking Growth* (Crown Business, 2017)
• Balfour, Brian — Reforge Growth Series Curriculum (2017-2026)
• A16z — Sixteen Startup Metrics & SaaS Benchmarks Reports (Andreessen Horowitz, ongoing)
• OpenView Partners — Annual SaaS Benchmarks Report and PLG Index (2020-2026)
• Rachitsky, Lenny — Lenny's Newsletter Marketplace and PLG Essays
• Gurley, Bill — Above the Crowd — All Markets Are Not Created Equal (2012)
• Penenberg, Adam — *Viral Loop* (Hyperion, 2009)
• Blank, Steve — *The Four Steps to the Epiphany* (K&S Ranch, 2005)
• Amplitude & Mixpanel — Product Analytics Benchmark Reports (2024-2026)