MVP & the Build–Measure–Learn loop — learn fast, spend little
🎯 Goal: Learn to build the cheapest test that teaches you the fastest, and always validate the RISKIEST assumption before building big.
An MVP (Minimum Viable Product) is not a “small, ugly product” — it’s the cheapest experiment that teaches you the single most important thing. Eric Ries (The Lean Startup) describes the loop Build → Measure → Learn: build a test → measure real behaviour → extract the lesson → adjust, as fast as possible. The survival trick: attack the riskiest (leap-of-faith) assumption first — the one that, if wrong, collapses the whole project — instead of polishing the easy, safe stuff. Learning one loop fast beats writing pretty code for a product nobody needs yet.
Lesson content
🎯
An MVP is a question, not a product. Every startup stands on a stack of assumptions: “people have this problem”, “they’ll pay”, “we can deliver at cost X”. Build it all and you lose a month and your money if one layer is wrong. The MVP asks the reverse: what must be true for the whole project to survive? Then it finds the cheapest way to ask the market that question. A landing page, a fake demo, pre-order deposits, selling by hand before the product exists — all are MVPs if they answer an assumption.
🧭
The Build–Measure–Learn loop — 5 steps done right: 1) List all the big assumptions of your idea. 2) Pick the riskiest one — wrong means collapse. 3) Set a pass threshold BEFORE testing (e.g. “≥ 20 pre-orders in 1 week”) so you can’t fool yourself later. 4) Build the cheapest test that measures just that assumption (no more). 5) Measure real behaviour → Learn: pass and advance; fail and pivot or retest price/positioning. Loop fast, learning one thing each turn.
🧰
TOOL — Experiment canvas (one row per assumption, set the threshold FIRST):
Assumption
Cheap test
Metric
Pass threshold
People want a breakfast box delivered to the door
Landing page + Pre-order button
% who click order / visits
≥ 8% and ≥ 30 orders
They’ll prepay 50k to reserve a slot
Collect a deposit via e-wallet / QR
People who actually deposit
≥ 15 in 1 week
We can deliver by 7am sharp
Hand-deliver to the first 10
% orders delivered on time
≥ 90% on time
Rule: set the threshold before you see the result. Once the numbers arrive, it’s easy to rationalize “6% is fine too” — a pre-set threshold stops that.
🌏
Case study — pre-selling on Ecomobi/Passio (step by step): 1) Instead of buying a full inventory and praying it sells, many sellers let creators pre-sell a new product. 2) They measure clicks and orders — the real behaviour of buyers. 3) Only if orders clear the threshold do they then source the goods to fulfil; if not, they drop the product without ever paying to hold stock. This is exactly a “test demand before you build” MVP. Contrast — the risky way: borrow money to import 1,000 units on a gut feeling that “it’ll surely sell”, then get stuck with unsold stock. The difference: one lets the market answer first, the other bets all its capital on an unvalidated assumption.
⚠️
5 MVP traps: • Building too many features “to be safe” before showing customers — slow and costly. • Testing the easy assumption, dodging the riskiest one out of fear of hearing “no”. • Not setting a threshold first, then rationalizing a weak result. • Measuring vanity signals (likes, views) instead of real behaviour (orders, payments). • Loving your MVP like your baby, unable to drop it when the data says it’s wrong.
✅
Checklist before running an MVP: ① Am I testing the RIGHT, riskiest assumption? ② Is this test the CHEAPEST way to answer that question? ③ Did I set a numeric pass threshold BEFORE testing? ④ Is my metric real BEHAVIOUR (order/pay), not vanity likes? ⑤ If it fails, do I have a plan to pivot or retest?
Practice exercise
🔬 APPLIED EXERCISE (1 week): (1) Write your idea’s 3 biggest assumptions. (2) Circle the RISKIEST one (wrong = whole project collapses). (3) Fill one Experiment-canvas row for it: the cheapest test, the metric, and a pass threshold set BEFORE testing. (4) If possible, run it for one week, then note: pass/fail and the lesson.
Worked template: FILLED EXAMPLE — a tutor-matching app: riskiest assumption = “parents will pay to find a tutor via an app”. Cheap test: a post + a signup form with a “willing to pay X” field. Metric: parents who actually fill the form. Pre-set threshold: ≥ 20 in one week before you start coding. Result 25 → passes, advance to the next test.
Quick quiz
1. What does an MVP truly mean?
→ The cheapest experiment that teaches the single most important thing
An MVP is the cheapest experiment that teaches the most important thing, not a small, ugly product.
2. In the Build–Measure–Learn loop, which assumption should you attack FIRST?
→ The riskiest one — the one that, if wrong, collapses the whole project
In Build–Measure–Learn, attack the riskiest assumption first — the one whose failure collapses the project.
3. Why is a “landing page + deposit button” a good test before building the product?
→ Because it measures real BEHAVIOUR (click/pay) cheaply and fast, confirming demand
A landing page + deposit button measures real behaviour (click/pay) cheaply and fast, confirming demand before building.
4. Sellers on Ecomobi/Passio “letting creators pre-sell, then sourcing” is an example of:
→ A kind of MVP that tests demand first, cutting inventory risk
Ecomobi/Passio lets creators pre-sell, then source inventory — an MVP testing demand and cutting warehouse risk.
5. Why set a numeric “pass threshold” BEFORE running the test?
→ To avoid rationalizing a weak result after seeing the numbers
Set a numeric pass threshold BEFORE testing so you don’t rationalize a weak result afterwards.
6. A landing page gets 12% “Pre-order” clicks, but almost nobody will deposit 50k. The most correct lesson is?
→ Interest (clicks) differs from willingness to pay; the “will pay” assumption isn’t confirmed, so retest price/positioning
A click is only mild interest; paying is the strong proof of demand — a big gap signals you should retest price.
Advanced
A deeper framework
Customer behaviour comes in commitment rungs, and each rung is a different-strength signal: seeing an ad (weakest) → clicking → leaving an email → free pre-order → actual payment/deposit (strongest). A good MVP measures at a rung as close to money as possible. Many teams celebrate “lots of interest” but measure the wrong, weak rung — likes and views don’t pay the bills.
When a result fails, you have two choices: pivot — keep the lesson, change one big element (segment, problem, pricing model); or persevere — tweak small things and retest. The key is to decide on behavioural data, not on the emotional cost of your effort. The MVP exists so you dare to drop a wrong assumption early and cheaply.
The behaviour-signal ladder (stronger as it nears money)
1,000 ad views
Very weak signal — just attention
120 click “Pre-order”
Interest — but nothing spent yet
15 pay a 50k deposit
Strong evidence — real money spent
The closer to paying behaviour you measure, the more trustworthy the signal. 15 deposits beat 1,000 views.
Common trap: “Vanity metrics”: taking pride in high likes, views, downloads — numbers that look nice but don’t prove demand. Always measure behaviour near money (orders, deposits, repeat buys); otherwise you’re fooling yourself with pretty data.
Advanced questions
1. You have 4 weeks and one of two moves: (1) code a full-featured app first, then give it to customers, or (2) build a landing page + deposit button to measure demand for one week, then decide. By Lean Startup, which is wiser?
→ (2), because it validates the riskiest assumption (is there real demand) cheaply and fast, avoiding building big only to find it’s wrong
Lean Startup prioritizes fast, cheap learning. (2) is an MVP attacking the riskiest assumption head-on — whether anyone truly wants to pay — before spending 4 weeks coding. If nobody orders on the landing page, you save a month and avoid building what nobody needs.
2. Your landing page gets a high 12% “Pre-order” click rate, but almost nobody deposits 50k. The most correct lesson from Build–Measure–Learn is?
→ Interest (clicks) differs from willingness to pay; the “will pay” assumption isn’t confirmed, so retest price/positioning before building
Measure behaviour at multiple levels: a click is only mild interest, while paying is strong proof of demand. The big gap between 12% clicks and ~0% deposits reveals a barrier in price or value — learn and adjust, don’t build big yet.
3. Your team calls the MVP a “success” for hitting 50,000 views and 3,000 likes, but nobody has ordered or deposited. How should you judge it?
→ Those are vanity metrics; no behaviour near money was measured, so real demand isn’t proven
Likes and views are vanity metrics — easy to inflate but no proof of demand. Trustworthy signals live in behaviour near money (orders, deposits, repeat buys). Redesign the test to measure the paying-commitment rung.
🎯 Real-life mission
REAL-LIFE MISSION: Write your idea’s 3 biggest assumptions and circle the RISKIEST. Fill one Experiment-canvas row: the cheapest test (landing page / deposit / hand-selling), a metric that is BEHAVIOUR near money, and a pass THRESHOLD set in ADVANCE. Run it for one week, then note: pass or not, which commitment rung you measured, and whether you’ll pivot or persevere next loop.