Lean Analytics: Use Data to Build a Better Startup Faster
Source: https://www.oreilly.com/library/view/lean-analytics/9781449335687/ ↗
Croll and Yoskovitz wrote the operational companion to Ries's theoretical argument in The Lean Startup.
Where Ries argues that validated learning is the right unit of progress, Croll and Yoskovitz tell you which numbers to actually look at at each stage of a startup and what they mean.
The "One Metric That Matters" framing is their most portable idea — at any given moment there is usually a single number whose movement contains most of what you need to know, and picking it forces the hard conversation about what you are actually trying to do.
Read alongside Doerr for the goal-setting layer and Varian for the pricing mechanics. More useful than it is elegant.
Central argument
Croll and Yoskovitz argue that most startups fail not from lack of data but from measuring the wrong things at the wrong time. Their central thesis is that each stage of a startup's development has a single metric that matters most — the One Metric That Matters (OMTM) — and that identifying it forces clarity about what the business is actually optimizing for. They map specific metrics to specific business models and growth stages, turning validated learning from an abstract principle into an operational checklist.
Critique
The stage-gate model underlying OMTM assumes a relatively linear progression through startup phases — empathy, stickiness, virality, revenue, scale — which maps poorly onto the messier reality of established digital products that must simultaneously optimize across multiple dimensions. A CPO running a mature product with competing stakeholder demands cannot easily subordinate everything to one number without creating blind spots; the framework's strength as a forcing function becomes a liability when the real problem is managing trade-offs between retention, monetization, and growth at the same time. The book also predates the widespread normalization of experimentation infrastructure, so its metric recommendations occasionally feel prescriptive in ways that assume data access conditions that no longer represent the hard part of the problem.
Why it matters for product
For a product leader, the OMTM concept is most valuable as a diagnostic tool for team alignment rather than a literal measurement rule — if a product team cannot agree on a single metric to move this quarter, that disagreement usually surfaces a deeper unresolved strategic question about what the product is for. The book's model-specific metric breakdowns (SaaS vs. marketplace vs. media vs. e-commerce) give CPOs a concrete vocabulary for challenging product teams who are hiding behind vanity metrics or composite scores that cannot be acted on. Read alongside Doerr's OKR framework, it helps distinguish the goal-setting layer from the measurement layer — a confusion that generates significant organizational dysfunction in product organizations.