Model Validation: Lean Startup
A team in Guadalajara spends an entire year programming an online tutoring platform without showing it to anyone "so that no one copies it." When they launch, they discover that students wanted something else. Another team, with the same idea, would have found that out in a week. What makes the second team learn fifty-two times faster than the first?
The Build-Measure-Learn Cycle
Now place the cycle around the minimum viable product (MVP). You build the MVP, you measure the customers' reaction, and the phase that closes the cycle is learn: comparing the data with the hypothesis and deciding whether to persevere or pivot. In the figure, that phase appears with a "?" so that you can identify it.
Explain what the Lean Startup methodology is and why it proposes validating ideas with real customers instead of blindly trusting the initial plan.
Explain what the Lean Startup methodology is and why it proposes validating ideas with real customers instead of blindly trusting the initial plan.
The basic Lean Startup cycle is build → measure → learn. Describe what is done in each phase and explain why it is a cycle and not a line that ends in 'launch'.
The basic Lean Startup cycle is build → measure → learn. Describe what is done in each phase and explain why it is a cycle and not a line that ends in 'launch'.
The diagram shows the Lean Startup validation cycle around the minimum viable product (MVP). Identify the missing phase, marked with '?', and explain what two decisions…
The diagram shows the Lean Startup validation cycle around the minimum viable product (MVP). Identify the missing phase, marked with '?', and explain what two decisions can come out of it.
The Minimum Viable Product
A team wants to validate a book-swapping app for neighbors. Propose a minimum viable product (MVP) that does not require coding the full app (for example, a messaging…
A team wants to validate a book-swapping app for neighbors. Propose a minimum viable product (MVP) that does not require coding the full app (for example, a messaging group and a form) and explain what hypothesis it allows them to test.
Testable Hypotheses and Customer Development
Customer development recommends 'getting out of the building' to talk to customers before building anything. Explain how you would conduct ten interviews to validate a…
Customer development recommends 'getting out of the building' to talk to customers before building anything. Explain how you would conduct ten interviews to validate a healthy snack box service and why questions like 'do you like my idea?' should be avoided.
Persevere or Pivot
In Lean Startup, after measuring, you must decide whether to persevere (continue improving in the same direction) or pivot (change strategy while retaining what was…
In Lean Startup, after measuring, you must decide whether to persevere (continue improving in the same direction) or pivot (change strategy while retaining what was learned). Define both options and reason what you would do if data shows that many people try your recipe app, but hardly anyone uses it again.
Vanity Metrics and Actionable Metrics
A startup's monthly report shows: 150,000 cumulative downloads (+25 %), 80,000 social media followers (+40 %), weekly retention rate of 2.3 % (-15 %), and free trial to…
A startup's monthly report shows: 150,000 cumulative downloads (+25 %), 80,000 social media followers (+40 %), weekly retention rate of 2.3 % (-15 %), and free trial to paid subscription conversion rate of 0.8 % (-0.5 points). The marketing director proposes doubling advertising spend, arguing that "the business keeps growing." Technically refute her proposal based on the analysis and interpretation of these data, classifying and justifying each metric according to its methodological typology (vanity or actionable metric), and predict what will happen to the company's financial viability if they maintain this strategy.
Designing a Complete Experiment
You now have all the pieces. A Lean experiment brings together four elements: a hypothesis (who, what they do, how much), an MVP (the cheapest test), a metric (actionable, not vanity), and a decision criterion (a number set before measuring).
A startup spent a year programming its online class platform without showing it to anyone 'so it wouldn't be copied', and upon launching discovered that students wanted…
A startup spent a year programming its online class platform without showing it to anyone 'so it wouldn't be copied', and upon launching discovered that students wanted something else. Diagnose which Lean principles were violated and reconstruct what its first build–measure–learn cycle would have looked like if applied from the first month.
Design a complete validation experiment for a healthy office meal delivery service: state the main hypothesis, the MVP you would use, the metric you would measure, and…
Design a complete validation experiment for a healthy office meal delivery service: state the main hypothesis, the MVP you would use, the metric you would measure, and the numerical criterion you would use to decide whether to persevere or pivot. Justify each choice.
Back to the beginning: the second team learns $52$ times faster because it launches a minimum viable product in a week to measure the market's real response, instead of spending a year carrying out a plan blindly without testing its hypotheses.