RunsheetInnovation

Lean Startup

ComplexityMedium

TimeWochen bis Monate je Lernzyklus

Participants2-8

FormatWorkshop + async

MaturityCanonical

Prerequisite

What needs to be finished first

Complete firstAssumption Mapping

A prioritized list of the riskiest assumptions about problem, solution, market, and business model exists.

Without: Without prioritized assumptions, random MVPs are built and learning per investment time stays low.

Complete firstLean Canvas

A Lean Canvas version or comparable business model sketch describes assumption fields (problem, solution, channels, revenue).

Without: Without business model anchors, successful experiments can create solutions that do not scale economically.

Preparation

What needs to be ready before start

Materials

Assumption backlog with priority; experiment template (hypothesis, method, sample, success criterion); tracking tool for learning reports (Notion, Confluence, Airtable); analytics setup for behavior; budget per cycle.

People / roles

A founder or Product Lead as owner; one to three makers (cross-functional, often one PM, one designer, one engineer); an analyst for measurement setup; a sponsor with pivot/persevere mandate.

Pre-read

Vision and business model; prioritized assumption backlog; available data and target group; budget and time limit per cycle; known constraints (regulatory, technical).

Time needed

Weeks to months per cycle, several cycles total

Setup

Define cycle cadence (for example 2-3 weeks per build-measure-learn). Create learning report template. Put pivot-persevere date in calendar. Secure sponsor mandate in writing.

Core question

The one question this method answers

Which riskiest assumption does the team test in the next cycle with minimal effort, and what measurable behavior confirms or contradicts it?

Flow

Marker: Phase

Step	Duration	Action	Hint
1Phase 1: Hypothesis and success criterion	1-2 days	Formulate the riskiest open assumption as hypothesis: "We believe X does Y because Z. We will know we are right if we observe A in B people." Set success criterion before building the MVP.	Defining success after MVP build is confirmation bias. Define in advance what exactly constitutes confirmation and contradiction.
2Phase 2: Build MVP	3-10 days	Build the smallest possible version that allows measuring behavior. Options: landing page, concierge, wizard of oz, fake door, prototype. Build only essential engineering.	MVPs become products when building is enjoyable. Keep a strict cut: what is minimally needed to test behavior? Everything else waits.
3Phase 3: Measure	1-4 weeks depending on sample	Measure behavior quantitatively: conversion, clicks, payments, return visits. Reach minimum sample before interpreting. Capture raw data, not only aggregates.	Teams who interpret intermediate data often stop too early or bias results. Disciplined waiting for minimum sample is required.
4Phase 4: Learn and report	1-2 days	Check data against success criterion. Write learning report: hypothesis, setup, outcome, interpretation, open questions. Sponsor and team review.	A learning report without explicit outcome label (confirmed/disproved/open) has little effect. Outcome label is mandatory, even when negative.
5Phase 5: Pivot or persevere	Half day	Decision: continue as planned (persevere), change strategy (pivot), or switch learning mode (measure differently). For each decision include rationale and next assumption. Sponsor signs.	A pivot without rationale is capitulation. Persevere without rationale is stubbornness. Every decision needs a learning-report reference, or the method becomes theater.

Artifact

What comes out at the end

Form

Learning report per cycle with hypothesis, MVP description, measurement setup, outcome, pivot/persevere decision, and next hypothesis. Plus ongoing assumption backlog with status (open, in test, confirmed, disproven).

Versioning / ownership

One learning report per cycle with date and cycle number. Assumption backlog grows with status updates. Pivot decisions must refer to learning report; archive prior hypotheses.

Tool alternatives

Notion or Confluence page per learning report
Airtable experiment database
Linear experiment issue type
Google Docs for detailed learning reports

lean-startup-working-template.md

Compact working template for Lean Startup with context, input, output artifacts, and next step.

Lean Startup Working Template

Goal

Build-Measure-Learn cycles to validate business assumptions quickly and affordably.

Context

When and for what do we use this method?

Input

Which data, observations, decisions, or materials are available?

Execution

Short notes along the runsheet.

Output artifacts

Hypothesis list:
MVPs:
Learning reports:
Pivot or persevere decision:

Assumptions and open questions

Decision / next step

Owner, date, and success signal.

Example output

Concrete filled scenario, fictional example

lean-startup-beispiel.md

Concrete filled scenario, fictional example

Learning report cycle 4 — Solo Tax Advisory MVP (18.05.2026)

Hypothesis: We believe solo tax advisors in DACH will pay EUR 19/month for AI document capture because document capture is their largest time sink. Confirmation is when 25 of 50 contacted advisors proceed to confirmation page (50% conversion).

MVP: Landing page with pricing table (9/19/29 EUR), Stripe checkout in test mode (real cards not charged). Promoted via LinkedIn DM outreach (53 people contacted).

Measurement (2 weeks, 53 advisors contacted, 50 clicks):

19 EUR plan: 11 of 50 reached checkout (22%).
9 EUR plan: 27 of 50 (54%).
29 EUR plan: 4 of 50 (8%).

Outcome: Hypothesis disproved for 19 EUR. Confirmed for 9 EUR.

Interpretation: Willingness to pay is lower than assumed. Possible reasons: advisors compare with DATEV Company Online, which is already included in full DATEV plans. Value proposition needs clearer differentiation.

Symptom

Only 8 clicks are measured; any interpretation has ±50% uncertainty.

What to do

Define minimum sample before test (for example 50 or 100). Mark results as open with smaller sample and extend cycle.

Stop criteria

Done signals checkable in under a minute

No clearly formulatable business model or target segment, so assumptions cannot be prioritized.

Requirements are contractually fixed and experiments are not allowed.

No sponsor mandate for pivot decisions, making learning reports unenforced.

Existing data already makes riskiest assumptions sufficiently known.

Team cannot build quantitative measurement and behavior evidence is structurally absent.

Time or budget does not support at least 3 cycles, leaving data points insufficient.

Finished the runsheet?

Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.

Go to profile More methods