Exercises: Model validation: Lean Startup

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 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.

In Lean Startup, ideas are formulated as testable hypotheses. Turn "I think people will like it" into a testable hypothesis for a dog walking app, and explain what makes it testable (who the customer is, what behaviour you expect, how you would measure it).

Define what a minimum viable product (MVP) is and explain why "minimum" does not mean "shoddy": it must have just enough to validate a hypothesis, but work well at that.

Order these actions in the build–measure–learn cycle and explain what follows next: counting how many visitors sign up, launching a page that advertises the product with a sign-up button, deciding whether the data confirms the hypothesis.

Explain what pivoting means in Lean Startup (changing strategy while keeping what has been learned) and give an example of a segment pivot: an app designed for students that ends up aimed at companies.

Explain why customer development orders you to "get out of the building" to talk to customers before building the product, and what two risks anyone who skips that step runs.

A team wants to validate a book-swapping app for neighbours. 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.

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.

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.

Design the minimum viable product of a maths tutoring service between students at the same school without programming any app (noticeboard, form, messaging group) and point out the key metric that would say whether the idea works.

A validation landing page advertises a product that does not exist yet and measures how many people leave their email or click "buy". Explain how this technique works, which hypothesis it validates, what advantages it has over surveys and what practical and ethical limits should be respected.

A meditation app boasts 50 000 accumulated downloads, but only 800 people open it each week. Explain which of the two metrics rules when deciding the product's future, why the other is a vanity metric and what two measures you would try to improve retention.

A healthy breakfast home subscription service gets many first purchases, but only 15% renew in the second month. State what data you would ask for to understand the leakage (reasons for cancelling, perceived price, variety) and reason whether it is time to persevere by improving the product or to pivot.

Write five useful questions to interview potential customers of a party sound equipment rental service, avoiding leading questions ("do you like my idea?") and seeking real past behaviours ("how did you sort out the music at your last party?").

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.

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.

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.

"StudyBite", a fictitious app that generates notes summaries with artificial intelligence, is in the validation phase. a) State its two main hypotheses (value and growth). b) Propose a minimum viable product to test each one. c) Define the metric and the numerical threshold with which you would decide to persevere or pivot in each case.

The founding team of a sports booking platform has spent two years introducing continuous aesthetic improvements and new features ("dark mode", user badges, calendar integration). Despite this, recurring usage does not exceed 4% and the acquisition cost is twice the average revenue generated per user. Management justifies its inaction by stating that they are "iterating with agility". Refute their argument by explaining which methodological trap they have fallen into, diagnose which mistakes they are making regarding Eric Ries's principles, and design a periodic evaluation protocol with quantitative criteria that forces the team to persevere or pivot.

A home 3D-printing company designed a model for individuals to print decorative figurines and household spare parts via subscription. After six months, they observe that individuals do not have recurring demand, but several local auto repair shops and dental clinics have called asking if they manufacture urgent custom parts. Based on this situation, design three distinct pivot strategies for the company (varying the customer segment, the need addressed, or the channel/monetisation) and justify how you would leverage the validated learning from the initial model in each one so as not to start from scratch.

A mobile language-learning application shows the following monthly figures in its funnel:

• Visits to the website and app store listing: 50.000

• Downloads and account registrations: 12.000

• Users completing their first lesson on the day of download: 1.200

• Users still practicing after 30 days: 1.050

• Users inviting a friend: 350

• Users paying for the annual premium subscription: 210

1. a) Map each data point to its stage within the pirate metrics model (AARRR) and calculate the percentage conversion rates between consecutive stages.
2. b) Interpret the results: identify with justification at which stage the main bottleneck is located.
3. c) Formulate an improvement hypothesis to fix this drop-off and design a specific Lean experiment to validate it.

At a panel discussion on entrepreneurship, a speaker states: "In Lean Startup, all that matters is speed: quickly code any shoddy product, launch it to the market, and see what happens; planning and researching is a waste of time". Adopt a critical stance and technically refute their statement. Contrast the rigorous concept of a Minimum Viable Product with a defective product, justify why the Lean method is grounded in the scientific method, and predict the economic consequences for a company that confuses agility with the absence of testable hypotheses and metrics.