A Walkthrough of the Lean Startup Skill: From Idea to Validated Learning
Seven commands that run the complete Build-Measure-Learn loop — from MVP design through experiment planning to pivot-or-persevere decisions.
Eric Ries's The Lean Startup introduced a framework for building products through validated learning rather than build-and-hope. The Build-Measure-Learn loop is elegant in theory and demanding in practice: it requires disciplined hypothesis formulation, honest measurement, and willingness to act on what you learn even when the data is inconvenient.
The Lean Startup BookSkill has seven commands that operationalize the framework — from designing your first MVP through running a full BML cycle to making the pivot-or-persevere decision. Here's what each does.
The Seven Commands
/mvp-designer — Define What You're Actually Testing
What it does: Walks you through Ries's MVP design process. You describe your idea, identify your core assumptions, determine which assumption is riskiest, and design the minimum viable product that tests it. The command helps you distinguish between a real MVP (a learning instrument) and a "just ship something" approach.
What you get: An MVP spec with a clear hypothesis, the specific assumption being tested, the MVP format (concierge, wizard of oz, landing page, or another type), and the metric that will confirm or disconfirm the hypothesis.
When to use it: Before building anything. The MVP designer ensures you're building to learn, not just building.
/experiment-plan — Set Up Validated Learning
What it does: Converts your MVP hypothesis into a structured experiment. Defines the specific metric you'll measure, the baseline, the success threshold, the time period, and what you'll do with the results. Applies Ries's innovation accounting framework.
What you get: An experiment design document: the hypothesis, the metric, the measurement method, the success/failure thresholds, and the learning question the experiment answers.
When to use it: After MVP design, before running the experiment. The plan forces clarity about what "it worked" and "it didn't work" actually mean before you're looking at the data.
/metrics-audit — Sort Your Metrics
What it does: Reviews your current metrics — for your startup, product, or marketing — and classifies each as actionable (tells you what to do differently) or vanity (makes you feel good but doesn't guide decisions). Helps you redesign your metrics dashboard around actionable metrics.
What you get: A metrics dashboard design: your vanity metrics (to stop tracking or deprioritize), your actionable metrics (to center your decisions on), and the cohort analysis or A/B test framework for each.
When to use it: Any time you're tracking a lot of metrics but not sure which ones to act on. Also useful after launch, when early data starts accumulating.
/customer-discovery — Plan Your Hypothesis Tests
What it does: Helps you design customer discovery interviews and behavioral tests. Not "would you use this?" surveys (which produce unreliable data) but structured conversations and experiments that reveal actual behavior.
What you get: An interview guide and hypothesis map — the specific questions to ask, the behaviors to observe, and the hypotheses each conversation is testing.
When to use it: Before building your MVP, to validate that the problem is real, and alongside early product releases to understand why people are or aren't engaging.
/pivot-or-persevere — Make the High-Stakes Call
What it does: Runs a structured analysis of your current results against your hypothesis. Takes you through Ries's framework for evaluating whether the data supports persevering (continuing on the current path with confidence) or pivoting (making a significant change to one or more components of the business model).
What you get: A pivot analysis with a recommendation: persevere, specific type of pivot, or more data needed. Includes which type of pivot (zoom-in, zoom-out, customer segment, customer need, platform, business architecture, value capture, engine of growth, channel, or technology) fits the data pattern.
When to use it: When you have enough data to evaluate your core hypothesis — typically after one or more completed Build-Measure-Learn cycles. This is the most consequential command in the skill.
/innovation-accounting — Set Up Learning Milestones
What it does: Helps you establish innovation accounting for your startup: the learning milestones that replace traditional financial projections in early-stage validation. Instead of "we'll hit $X revenue by month 6," you set "we'll achieve X% conversion on the landing page, then Y% activation, then Z% retention" as the learning sequence.
What you get: An innovation accounting framework: your learning milestones in sequence, the metrics for each, and the investment level you'll make at each stage based on the results.
When to use it: When setting up a new product or startup, or when transitioning from early experimentation to more structured execution.
/build-measure-learn — Run a Full Cycle
What it does: Runs a complete Build-Measure-Learn cycle on a single feature, hypothesis, or initiative. Integrates the MVP design, experiment planning, measurement, and learning phases into a single workflow.
What you get: A completed BML cycle report: what you built, what you measured, what you learned, and what the next hypothesis is.
When to use it: When you want to run a disciplined end-to-end cycle rather than individual components.
Recommended Sequence for a New Venture
/customer-discovery— validate the problem before building/mvp-designer— design what you'll build and what it tests/experiment-plan— set up the measurement framework/metrics-audit— ensure you're tracking the right things/build-measure-learn— run the full cycle/innovation-accounting— establish learning milestones/pivot-or-persevere— act on what you learned
What the Lean Startup Framework Actually Delivers
Ries's most important insight isn't the MVP or the pivot — it's the idea that startups should optimize for learning speed, not output speed. The team that runs 10 Build-Measure-Learn cycles before running out of runway has a dramatically higher chance of finding product-market fit than the team that spends the same time building one polished product.
The Lean Startup Skill accelerates the learning loops by making each step disciplined: the hypothesis is explicit, the measurement is pre-defined, the interpretation is structured. You still have to do the building — but you know what you're building and why.
Ready to run your first validated experiment? Get the Lean Startup BookSkill and start with /mvp-designer.