A Field Guide for Rapid Experimentation
Alex Osterwalder
Replace gut feeling with evidence — 44 experiments for testing whether your idea has legs before you bet on it.
Most entrepreneurs fall in love with their idea and skip straight to building. Osterwalder gives you a systematic way to stress-test assumptions early, so you spend real money only on ideas that have already earned it.
Everything Osterwalder and Bland want you to walk away with
Too many entrepreneurs execute ideas that look great in presentations and spreadsheets, only to learn their vision was a hallucination. Don't make the mistake of executing without evidence — test thoroughly first.
Desirability asks if customers actually want it. Feasibility asks if you can build and deliver it. Viability asks if you can earn enough money from it. Each risk gets its own hypotheses and experiments.
Use an Assumptions Map to plot every hypothesis by importance and evidence. The top-right quadrant — critical to success but lacking evidence — is where you must focus your testing energy first.
If all hypotheses start with 'We believe that...' you fall into a confirmation trap. Create competing hypotheses that try to refute your assumptions and test them simultaneously, especially when the team disagrees.
Every experiment needs four components: a hypothesis, a description, the metrics you'll measure, and pre-set success criteria. Without criteria defined in advance, you'll rationalize any result.
Incrementally fund teams using a venture-capital style approach based on the learnings they share. Match experiment strength to what you're trying to learn — don't over-invest before you've earned the right to.
Multitasking across projects silently kills progress. Teams need real budgets for experiments, direct access to customers, clear strategic direction, and a facilitative leadership style that leads with questions, not answers.
Don't fall in love with your first idea. Ideate broadly, narrow down with business prototypes, then assess. You'll constantly improve these prototypes with insights from testing in future iterations.
Asking people what they'd do is unreliable. Watching what they actually do when given the chance to sign up, pre-order, or pay tells you the truth. Real paying customers are always different from hypothetical ones.
They move fast, create momentum, and aren't afraid to test disruptive models. They know the bottleneck is always at the top of the bottle — leadership support makes or breaks the testing culture.
These notes are inspired by direct excerpts and woven together into a readable guide you can follow from start to finish.
By David J. Bland and Alexander Osterwalder
Too many entrepreneurs and innovators execute ideas prematurely because they look great in presentations, make excellent sense in spreadsheets, and look irresistible in the business plan—only to learn later that their vision turned out to be a hallucination. Don’t make the mistake of executing business ideas without evidence: test your ideas thoroughly, regardless of how great they may seem in theory.
💡 Key Insight
This book outlines one of the most extensive testing libraries available to help you make your ideas bulletproof with evidence. Test extensively to avoid wasting time, energy, and resources on ideas that won’t work.
To test a big business idea, you break it down into smaller chunks of testable hypotheses. These hypotheses cover three types of risk. First, desirability—that customers aren’t interested in your idea. Second, feasibility—that you can’t build and deliver your idea. Third, viability—that you can’t earn enough money from your idea. You test your most important hypotheses with appropriate experiments. Each experiment generates evidence and insights that allow you to learn and decide. Based on the evidence and your insights, you either adapt your idea if you learn you were on the wrong path, or continue testing other aspects of your idea if the evidence supports your direction.
Types of Hypotheses
| Hypothesis Type | Core Question | Primary Risk |
|---|---|---|
| Feasible | ”Can we do this?” | Can’t manage, scale, or access key resources, activities, or partners. |
| Desirable | ”Do they want this?” | Market too small, too few customers want the proposition, or we can’t reach and retain them. |
| Viable | ”Should we do this?” | Can’t generate more revenue than costs. |
This material serves three audiences. If you are a corporate innovator, you are challenging the status quo and building new business ventures within the constraints of a large organization. If you are a startup entrepreneur, you want to test the building blocks of your business model to avoid wasting the time, energy, and money of the team, cofounders, and investors. And if you are a solopreneur, you have a side hustle or an idea that is not quite yet a business.
If you do not have all of the skills your team needs, evaluate technological tools to fill the void. There are new tools coming to market every day that allow you to create landing pages, design logos, and run online ads — but a strong cross-functional team is still the foundation.
Definition — Cross-Functional Team
A cross-functional team has all the core abilities needed to ship the product and learn from customers. A common basic example consists of design, product, and engineering. Depending on the business, you may also need skills in legal, data, sales, marketing, research, and finance.
Key Insight
A lack of diverse experiences and perspectives on a team will result in baking your biases right into the business. When forming your team, keep diversity top of mind — rather than as an afterthought.
1. Data Influenced. You do not have to be data driven, but you need to be data influenced. Teams no longer have the luxury of burning down a product backlog of features. The insights generated from data shape the backlog and strategy.
2. Experiment Driven. Teams are willing to be wrong and experiment. They are not only focused on the delivery of features, but craft experiments to learn about their riskiest assumptions. Match experiments to what you are trying to learn over time.
3. Customer Centric. To create new businesses today, teams have to know “the why” behind the work. This begins with being constantly connected to the customer — not limited to the new customer experience, but expanding to both inside and outside of the product.
4. Entrepreneurial. Move fast and validate things. Teams have a sense of urgency and create momentum toward a viable outcome. This includes creative problem-solving at speed.
5. Iterative Approach. Teams aim for a desired result by means of a repeated cycle of operations. The iterative approach assumes you may not know the solution, so you iterate through different tactics to achieve the outcome.
6. Question Assumptions. Teams have to be willing to challenge the status quo and business as usual. They are not afraid to test out a disruptive business model that will lead to big results, as compared to always playing it safe.
Dedicated. Teams need an environment in which they can be fully dedicated to the work. Multitasking across several projects will silently kill any progress. Small teams who are dedicated outperform large teams who are not.
Funded. It is unrealistic to expect these teams to function without a budget. Experiments cost money. Incrementally fund teams using a venture-capital-style approach, based on the learnings they share during stakeholder reviews.
Autonomous. Teams need to be given space to own the work. Do not micromanage them to the extent that it slows their progress. Instead, give them space to give an accounting of how they are making progress toward the goal.
Leadership. A facilitative leadership style is ideal — lead with questions, not answers. Be mindful that the bottleneck is always at the top of the bottle.
Coaching. Teams need coaching, especially if this is their first journey together. Coaches — internal or external — help guide teams when they are stuck, particularly in moving beyond interviews and surveys to a wider range of experiments.
Customer Access. The trend of isolating teams from the customer must be reversed. If teams keep getting pushback on customer access, they will eventually just guess and build anyway.
Resources. Teams need enough physical or digital resources to make progress and generate evidence. Constraints are good; starving a team will not yield results.
Strategy. Without a clear, coherent strategy, you will mistake being busy with making progress. Teams need direction on where they will play — adjacent market or new one — to unlock new revenue.
KPIs. Key performance indicators act as signposts so that everyone can tell whether the team is making progress toward the goal.
Before diving in, get the team aligned using these nine steps:
Team Alignment Steps
Shaping an idea is a design loop with three repeating steps: Ideate, Business Prototype, and Assess.
Step 1 — Ideate. Come up with as many alternative ways as possible to use your initial intuition — or insights from previous testing — to turn your idea into a strong business. Do not fall in love with your first ideas.
Step 2 — Business Prototype. Narrow down the alternatives from ideation using business prototypes. When starting out, rough prototypes like napkin sketches are fine. Use the Value Proposition Canvas and Business Model Canvas to make ideas clear and tangible. You will constantly improve these prototypes as insights flow in from testing.
Step 3 — Assess. Evaluate the design of your business prototypes. Ask: “Is this the best way to address our customers’ jobs, pains, and gains?” “Is this the best way to monetize our idea?” “Does this best reflect what we have learned from testing?” Once satisfied, start testing in the field.
You do not have to be a master of the Business Model Canvas to use this framework, but it is invaluable for defining, testing, and managing risk. It breaks a business idea into nine building blocks:
Definition — The Nine Building Blocks
The canvas maps visually as a single page. On the right are customer-facing elements (Segments, Value Propositions, Channels, Relationships) — your desirability building blocks. In the center-left are Key Activities, Resources, and Partners — your feasibility building blocks. At the bottom, Revenue Streams and Cost Structure form your viability building blocks.
The Value Proposition Canvas zooms into two building blocks and examines them in finer detail. It has two sides: the Value Map describing what you offer, and the Customer Profile describing who you are serving.
Definition — Value Map (Your Offering)
Definition — Customer Profile (Your Customer)
The goal is to achieve a strong fit between your Value Map and your Customer Profile — your gain creators and pain relievers should directly address the gains and pains that matter most to your customers.
To test a business idea you first have to make explicit all the risks that your idea will not work. You need to turn the assumptions underlying your idea into clear hypotheses you can test, then prioritize them ruthlessly.
When creating hypotheses, begin by writing the phrase “We believe that…” — for example: “We believe that millennial parents will subscribe to monthly educational science projects for their kids.”
Key Insight — Avoiding Confirmation Bias
If you create all of your hypotheses in the “We believe that…” format, you can fall into a confirmation bias trap — constantly trying to prove what you believe instead of trying to refute it. To prevent this, create competing hypotheses that try to disprove your assumptions: “We believe that millennial parents will not subscribe…” You can even test these competing hypotheses at the same time. This is especially helpful when team members cannot agree on which direction to pursue.
Principle — Characteristics of a Good Hypothesis
These hypotheses address whether customers actually want what you are offering. They map to the customer-facing building blocks of the Business Model Canvas: Customer Profile (Do the jobs, pains, and gains you identified really matter?), Value Map (Do your products actually solve high-value jobs and relieve pains?), Customer Segments (Are you targeting the right segments, and are they large enough?), Value Propositions (Is yours unique enough?), Channels (Can you reach customers?), and Customer Relationships (Can you retain them?).
These hypotheses test whether you can actually build and deliver what you promise — whether you can perform the necessary Key Activities at scale, secure the required Key Resources (technology, IP, talent, capital), and create the Key Partnerships your model depends on.
These hypotheses examine whether your idea can make money — whether customers will pay a specific price, whether you can generate sufficient Revenue Streams, keep your Cost Structure under control, and ultimately earn a profit.
Once you have written your hypotheses, use the Assumptions Map to prioritize them. It is a two-axis, two-by-two grid:
Assumptions Map (How to Use the Quadrants)
| Quadrant | Meaning | Action |
|---|---|---|
| Top Left — Important + Have Evidence | Critical assumptions with observable evidence already available. | Share with the team, challenge evidence quality, and keep monitoring. |
| Top Right — Important + No Evidence | Highest-risk assumptions with little or no proof. | Run experiments first. This is the primary focus area. |
| Bottom Left — Unimportant + Have Evidence | Low-priority assumptions that already have support. | Track lightly; avoid over-investing time. |
| Bottom Right — Unimportant + No Evidence | Low-priority assumptions with limited support. | Defer until higher-priority risks are resolved. |
Definition — The Assumptions Map
Step 1 — Identify: Write each desirability, feasibility, and viability hypothesis on its own sticky note. Use different colors for each type. Keep hypotheses short and precise — no blah blah blah. Discuss and agree as a team.
Step 2 — Prioritize: Plot every hypothesis on two axes. The x-axis is evidence: place a hypothesis on the left if you have relevant, observable, and recent evidence; on the right if you do not. The y-axis is importance: place a hypothesis at the top if it is absolutely critical for your idea to succeed — meaning if proven wrong, the entire idea fails. Place it at the bottom if it is not something you would test first.
Step 3 — Focus on the top-right quadrant: High importance, little evidence. These are your riskiest assumptions — the ones that, if proven false, will cause your business to fail. This is where you start.
Principle
The major focus of this entire framework is on how to test the top-right quadrant of your Assumptions Map: important hypotheses with little evidence. These assumptions, if proven false, will cause your business to fail. Start here — always.
With your most important hypotheses identified, turn them into experiments. Start with cheap, fast experiments to learn quickly. Every experiment you run reduces the risk that you will spend time, energy, and money on ideas that will not work.
A good experiment is precise enough so that any team member can replicate it and generate usable, comparable data. It clearly defines:
Definition — Four Components of an Experiment
Definition — Call-to-Action Experiment
A specific type of experiment that prompts a test subject to perform an observable action. Used to test one or more hypotheses by measuring what people do — not just what they say — making it a source of stronger evidence.
The Test Card captures your entire experiment on a single page before you begin running it:
Test Card Template
Also capture: Test Name, Deadline, Assigned To, and Duration.
Before running any experiment, document these parameters to ensure clarity and replicability:
Experiment Guidelines
Anyone who isn’t embarrassed by who they were last year probably isn’t learning enough. — Alain de Botton
Not all evidence is created equal. The strength of a piece of evidence determines how reliably it helps you support or refute a hypothesis.
Principle — The Evidence Spectrum
Weak evidence comes from opinions and beliefs (“I would…,” “I think… is important”), from what people say in interviews or surveys, from lab settings where people know they are being observed, and from small investments like email sign-ups.
Strong evidence comes from facts and events (“Last week I…,” “I spent $__ on…”), from what people actually do — observable behavior is a good predictor of future action — from real-world settings where people are not aware they are being tested, and from large investments like pre-purchasing a product or putting one’s professional reputation on the line.
In practice, you layer experiments to progressively strengthen evidence. You might start with interviews to gain initial insights into your customers’ jobs, pains, and gains. Then run a survey to test those insights at broader scale. Finally, conduct a simulated sale to generate the strongest type of evidence for customer interest. Your confidence level should rise with each additional experiment you conduct for the same hypothesis.
After running an experiment, capture your learnings in a structured way. The Learning Card is the companion to the Test Card:
Learning Card Template
Also capture: Insight Name, Date of Learning, and Person Responsible.
Experiments only create value when their evidence leads to clear decisions. This chapter covers the rituals and ceremonies that turn raw evidence into action — and keep your team aligned, reflective, and moving forward.
Daily Standup Agenda
Weekly Learning Agenda
This is arguably the most important ceremony. When you stop reflecting, you stop learning and improving.
Biweekly Retrospective Agenda
Principle 1 — Visualize Your Experiments
Make your work visible to yourself and others. Write down your experiments — one per sticky note — and draw a simple experiment board with four columns: Backlog → Setup → Run → Learn. Rank experiments in the Backlog from top to bottom, with the next one to execute at the top. Pull them across as you work on each. If you keep all this work in your head, you will never achieve flow, and your teammates cannot read your mind.
Principle 2 — Limit Experiments in Progress
Multitasking too many experiments leads to trouble. Define work-in-progress limits — for example, a limit of one experiment per column. This prevents the team from pulling a second experiment until the first has moved forward. In practice, this means you run the customer interviews before the survey, instead of trying to do both at once. What you learn from one experiment informs the next.
Principle 3 — Continuous Experimentation
Continue to experiment over time. Identify and visualize blockers — such as an internal department refusing to allow customer contact — because these prevent flow and need to be communicated to stakeholders. As your process matures, you may need to split columns (separating “ready to run” from “still being set up”) to capture the nuances of your workflow.
Key Insight — The Goal of the Entire Process
The goal is not to test and learn for the sake of it. The goal is to decide, based on evidence and insights, to progress from idea to business. Your experiments, ceremonies, and flow principles all serve this one objective. Make sure every cycle through the process brings you closer to a clear pivot, persevere, or kill decision.
With a library of experiments available, the challenge becomes picking the right one. Narrow your choice by asking three questions:
1. What type of hypothesis are you testing? Some experiments produce better evidence for desirability, others for feasibility, and others for viability. Pick experiments that match your major learning objective.
2. How much evidence do you already have? The less you know, the less you should waste time, energy, and money. When uncertainty is high, quick and cheap experiments are most appropriate — they point you in the right direction despite producing generally weaker evidence. As your understanding grows, shift to experiments that produce stronger evidence.
3. How much time do you have? If a major meeting with decision-makers or investors is approaching, you may need fast experiments to generate evidence on multiple fronts quickly. If you are running low on funding, choose experiments that will produce evidence compelling enough to extend it.
Principle — Four Rules of Thumb for Selecting Experiments
Great teams do not treat experiments as isolated events. They build momentum by running sequences that progressively strengthen evidence over time. Every experiment has logical predecessors and successors — experiments you can run before, during, and after. By chaining experiments together deliberately, you move from early, cheap signals to deep, high-confidence evidence faster than running experiments in isolation.
Definition — Common Experiment Sequences
| Context | Recommended Sequence |
|---|---|
| B2B Hardware | Customer Interview → Paper Prototype → 3D Print → Data Sheet → Mash-Up MVP → Letter of Intent → Crowdfunding |
| B2B Software | Customer Interview → Discussion Forums → Boomerang → Clickable Prototype → Presale → Single Feature MVP |
| B2B Services | Expert Stakeholder Interviews → Customer Support Analysis → Brochure → Presale → Concierge |
| B2C Hardware | Customer Interview → Search Trend Analysis → Paper Prototype → 3D Print → Explainer Video → Crowdfunding → Pop-Up Store |
| B2C Software | Customer Interview → Online Ad → Simple Landing Page → Email Campaign → Clickable Prototype → Mock Sale → Wizard of Oz |
| B2C Services | Customer Interview → Search Trend Analysis → Online Ad → Simple Landing Page → Email Campaign → Presale → Concierge |
| B2B2C | Customer Interview → Online Ad → Simple Landing Page → Explainer Video → Presale / Concierge → Buy a Feature / Data Sheet → Partner & Supplier Interview → Letter of Intent → Pop-Up Store |
| Highly Regulated | A Day in the Life → Validation Survey → Customer Support Analysis → Sales Force Feedback → Storyboard / Explainer Video → Brochure → Partner & Supplier Interview → Data Sheet → Presale |
Discovery experiments help you explore the landscape — uncovering customer jobs, pains, and gains, and generating initial signals about your value proposition. They are generally cheaper, faster, and produce weaker evidence than validation experiments, making them ideal for early stages when uncertainty is highest.
Customer Interview
An interview focused on exploring customer jobs, pains, gains, and willingness to pay. Ideal for gaining qualitative insights into the fit between your value proposition and customer segment — and a good starting point for price testing. Not ideal as a substitute for what people will actually do.
Partner & Supplier Interviews
Similar to Customer Interviews, but focused on whether you can feasibly run the business. You will be sourcing and interviewing key partners to supplement the Key Activities and Key Resources that you cannot do — or do not want to do — in-house.
Expert Stakeholder Interviews
Similar to Customer Interviews, but focused on getting buy-in from key players inside your organization whose support is critical to moving forward.
A Day in the Life
A method of qualitative research that uses customer ethnography — observing or working alongside customers — to better understand their jobs, pains, and gains in context. Relatively cheap; you may need to compensate people for their time.
Discovery Survey
An open-ended questionnaire used to collect information from a sample of customers. Ideal for uncovering jobs, pains, and gains at a broader scale. Not ideal for determining what people will actually do — only what they say they will do.
Search Trend Analysis
The use of search data to investigate patterns in what people are looking for online. Ideal for performing your own market research — especially on newer trends — instead of relying on third-party data.
Web Traffic Analysis
The use of website data collection, reporting, and analysis to look for customer behavior patterns in an existing product or site.
Discussion Forums
Mining online discussion forums to uncover unmet jobs, pains, and gains. Ideal for finding unmet needs in your existing product or a competitor’s.
Sales Force Feedback
Leveraging the frontline sales team to surface unmet jobs, pains, and gains. Ideal for businesses that use a group of people to conduct sales.
Customer Support Analysis
Using existing customer support data to uncover unmet jobs, pains, and gains. Ideal for businesses that already have a substantial number of existing customers.
Online Ad
An online advertisement that clearly articulates a value proposition for a targeted customer segment with a simple call to action. Ideal for quickly testing your value proposition at scale with customers online.
Link Tracking
A unique, trackable hyperlink used to gather quantitative data on customer actions. Ideal for testing what people do, not just what they say.
Feature Stub
A small test of an upcoming feature — usually in the form of a button — that includes only the very beginning of the experience. Ideal for rapidly testing the desirability of a new feature within an existing product. Not ideal for testing mission-critical functionality.
404 Test
A faster, riskier variation of the Feature Stub: nothing sits behind the button, generating 404 errors each time it is clicked. Count the errors to measure interest. Do not run a 404 test for more than a few hours.
Email Campaign
Email messages deployed across a specific period of time to customers. Ideal for quickly testing your value proposition with a customer segment. Not ideal as a replacement for face-to-face customer interaction.
Social Media Campaign
Social media messages deployed across a specific period of time. Ideal for acquiring new customers, increasing brand loyalty, and driving sales.
Referral Program
A method of promoting products or services to new customers through referrals, word of mouth, or digital codes. Ideal for testing how to organically scale your business.
Paper Prototype
A sketched interface on paper, manipulated by a person to simulate the software’s reactions to customer interaction. Ideal for rapidly testing the concept of your product with customers. Not ideal as a replacement for proper usability testing.
3D Print
Rapidly prototyping a physical object from a three-dimensional digital model. Ideal for rapidly testing iterations of a physical solution with customers.
Storyboard
Illustrations displayed in sequence to visualize an interactive experience. Ideal for brainstorming scenarios of different value propositions and solutions with customers.
Data Sheet
A one-page physical or digital sheet with specifications of your value proposition. Ideal for distilling your offering down to a single testable page for customers and key partners.
Brochure
A mocked-up physical brochure of your imagined value proposition. Ideal for testing in person with customers who are difficult to find online.
Explainer Video
A short video that explains a business idea in a simple, engaging, and compelling way. Ideal for quickly communicating your value proposition at scale.
Boomerang
Performing a customer test on an existing competitor’s product to gather insights on the value proposition. Ideal for finding unmet needs in an existing market without building anything.
Pretend to Own
A nonfunctioning, low-fidelity prototype placed into a customer’s daily life to test whether the solution fits naturally. Sometimes called a Pinocchio experiment. Ideal for generating your own evidence on the potential usefulness of an idea.
Product Box
A facilitation technique used with customers to visualize value propositions, main features, and key benefits in the physical form of a box. Ideal for refining your value proposition and narrowing in on key features.
Speed Boat
A visual game technique used with customers to identify what is inhibiting their progress — and how it impacts your feasibility. Ideal for going beyond conversations to get a visual representation of what is slowing customers down.
Card Sorting
A technique where a person uses cards with customers to generate insights into jobs, pains, gains, and value propositions.
Buy a Feature
A technique where people use pretend currency to “buy” the features they most want in a product. Ideal for prioritizing features and refining your understanding of customer jobs, pains, and gains.
Invention is not disruptive. Only customer adoption is disruptive. — Jeff Bezos
Validation experiments go further than discovery. They produce stronger evidence by testing whether customers will actually commit — with their time, money, or reputation — to your value proposition. They tend to require more investment to set up but yield evidence you can make real decisions on.
Clickable Prototype
A digital interface representation with clickable zones that simulate the software’s reactions to customer interaction. Ideal for rapidly testing the concept of your product at higher fidelity than paper. Not ideal as a replacement for proper usability testing with customers.
Single Feature MVP
A functioning minimum viable product with the single feature needed to test your core assumption. Ideal for learning whether the central promise of your solution resonates with customers.
Mash-Up
A functioning MVP that combines multiple existing services to deliver value without building custom technology. Ideal for learning whether the overall solution resonates with customers before investing in custom development.
Concierge
Creating a customer experience and delivering value entirely through manual human effort — and the customer knows it. Ideal for learning firsthand about every step needed to create, capture, and deliver value. Not ideal for scaling a product or business.
Life-Sized Prototype
Real-world replicas of service experiences at full scale. Ideal for testing higher-fidelity solutions with customers at a small sample size before deciding to scale.
Simple Landing Page
A simple web page that clearly illustrates your value proposition with a call to action. Ideal for determining whether your value proposition resonates with your customer segment.
Crowdfunding
Funding a project or venture by raising many small amounts of money from a large number of people, typically via the internet. Ideal for validating demand with customers who believe in your value proposition. Not ideal for determining whether your new business venture is feasible.
Split Test
A method of comparing two versions — control A against variant B — to determine which performs better. Ideal for testing different versions of value propositions, prices, and features to see what resonates best with customers.
Presale
A sale held before an item is made available for purchase. Unlike a mock sale, you process a real financial transaction when it ships. Ideal for gauging market demand at a smaller scale before launching to the public.
Validation Survey
A closed-ended questionnaire used to learn whether customers would be disappointed if your product went away or whether they would refer others to it. Captures sentiment at scale, though it reflects what people say rather than what they do.
Wizard of Oz
Creating a customer experience and delivering value manually with people instead of technology — but the customer does not know a person is behind the curtain. Ideal for learning firsthand about every step needed to create, capture, and deliver value. Not ideal for scaling a product or business.
Mock Sale
Presenting a sale for your product without processing any payment information. Ideal for testing different price points and measuring purchase intent without a real financial transaction.
Letter of Intent
A short, written, non-legally-binding contract. Ideal for evaluating key partners and B2B customer segments. Not ideal for B2C customer segments.
Pop-Up Store
A retail store opened temporarily to sell goods and test face-to-face purchasing behavior. Ideal for B2C businesses testing whether customers will really make a purchase. For B2B, consider a conference booth instead.
Extreme Programming Spike
A simple program built to explore whether a technical or design solution is feasible. The term comes from rock climbing and railroads — a necessary pause to determine whether you can continue making progress. Ideal for quickly evaluating software feasibility. Not ideal for scaling: it is typically thrown away and rebuilt properly afterward.
The more success you’ve had in the past, the less critically you examine your own assumptions. — Vinod Khosla, Venture Capitalist
Even disciplined teams fall into common traps when testing business ideas. Here are the eight most dangerous pitfalls — and how to fix each one.
Pitfall 1 — Time Trap: Not dedicating enough time.
You get what you invest. Teams that do not put in enough time to test business ideas will not get great results. Too often, teams underestimate what it takes to conduct multiple experiments well.
The Fix: Carve out dedicated time every week to test, learn, and adapt. Set weekly goals for what you would like to learn about your hypotheses. Visualize your work so that stalled or blocked tasks become obvious.
Pitfall 2 — Analysis Paralysis: Overthinking things you should just test and adapt.
Good ideas and concepts are important, but too many teams overthink and waste time rather than getting out of the building to test and adapt.
The Fix: Time-box your analysis work. Differentiate between reversible and irreversible decisions — act fast on the former, take more time for the latter. Avoid debates of opinion; conduct evidence-driven debates followed by decisions.
Pitfall 3 — Incomparable Data: Messy data that cannot be compared.
Too many teams are sloppy in defining their exact hypothesis, experiment, and metrics. That leads to data that cannot be compared — for example, testing with different customer segments or in wildly different contexts.
The Fix: Use the Test Card. Make the test subject, experiment context, and precise metrics explicit. Make sure everybody involved in running the experiment is part of the design.
Pitfall 4 — Weak Evidence: Only measuring what people say, not what they do.
Teams are often satisfied with surveys and interviews and fail to go deeper into how people actually behave in real-life situations.
The Fix: Do not just believe what people say. Run call-to-action experiments. Generate evidence that gets as close as possible to the real-world situation you are trying to test.
Pitfall 5 — Confirmation Bias: Only believing evidence that agrees with your hypothesis.
Sometimes teams discard or underplay evidence that conflicts with their hypothesis, preferring the illusion of being correct in their prediction.
The Fix: Involve others in the data synthesis process to bring in different perspectives. Create competing hypotheses to challenge your beliefs. Conduct multiple experiments for each hypothesis.
Pitfall 6 — Too Few Experiments: Conducting only one experiment for your most important hypothesis.
Few teams realize how many experiments they should conduct to support or refute a hypothesis. They make important decisions based on one experiment with weak evidence.
The Fix: Conduct multiple experiments for important hypotheses. Differentiate between weak and strong evidence. Increase the strength of evidence as uncertainty decreases.
Pitfall 7 — Failure to Learn and Adapt: Not taking time to analyze evidence and generate insights.
Some teams get so deep into testing that they forget the real goal: to decide, based on evidence and insights, to progress from idea to business.
The Fix: Set aside time to synthesize results, generate insights, and adapt your idea. Always navigate between the detailed testing process and the big picture. Create rituals to ask: Are we making progress from idea to business?
Pitfall 8 — Outsource Testing: Outsourcing what you should be doing and learning yourself.
Testing is about rapid iterations between testing, learning, and adapting an idea. An agency cannot make those rapid decisions for you, and you risk wasting time and energy by outsourcing.
The Fix: Shift resources reserved for an agency toward building an internal team of professional testers.
Leadership in an experimentation culture looks fundamentally different from traditional leadership. Whether you are improving an existing business model or inventing a new one, the principles remain the same: use language that empowers, hold teams accountable for outcomes rather than features, and develop your facilitation skills.
Be mindful of how your words land. Overuse of the leader’s perspective — “I think,” “my experience says” — can unintentionally strip teams of their decision-making authority. They will start waiting for you to assign experiments rather than designing their own. Focus on business outcomes, not features and dates, and create opportunities for teams to give an account of how they are making progress.
Principle — Leadership Language (Existing Models)
Do say: “We, Us, Our” · “How would you achieve this business outcome?” · “Can you think of 2–3 additional experiments?”
Don’t say: “I, Me, Mine” · “Deliver this feature by release date” · “This is the only experiment we should run”
Creating something new demands a “strong opinions, weakly held” mindset. Start with a hypothesis, but remain genuinely open to being proven wrong. If you are merely trying to prove you are right, your cognitive biases will take over.
Principle — Leadership Language (New Models)
Do say: “What is your learning goal?” · “What obstacles can I remove to help you make progress?” · “How else might we approach this problem?” · “What learning has surprised you so far?”
Don’t say: “I don’t trust the data.” · “I still think we should build it anyway.” · “You need 1,000 customers for it to mean anything.” · “This has to be a $15M business by next year.”
Leadership Moves
Meet your teams one half-step ahead. Think of where you eventually want team members to be, then look backward to find the next small nudge. Whether in one-on-ones, retrospectives, or hallway conversations — find opportunities to guide that first step.
Understand context before giving advice. Actively listen until team members are finished speaking. Ask clarifying questions to make sure you understand the full context before offering your perspective.
Key Insight — Say “I Don’t Know”
Practice saying these three words when you do not have the answer. It helps your teams understand that you do not have all the answers — nor should you. Follow it up with “How would you approach this?” or “What do you think we should do?” Saying “I don’t know” models the behavior the leaders you are developing will embrace.
There has been a broad shift from traditional, functionally siloed organization models to more agile, cross-functional team approaches. When testing new business ideas, speed and agility are imperative. Cross-functional teams can adapt more quickly than functionally siloed teams. In many organizations, small, dedicated, cross-functional teams outperform large, siloed project teams.
| Seed | Launch | Growth | |
|---|---|---|---|
| Funding | Less than $50,000 | $50,000–$500,000 | $500,000+ |
| Team Size | 1–3 | 2–5 | 5+ |
| Time per Team Member | 20–40% | 40–80% | 100% |
| Number of Projects | High | Medium | Low |
| Objectives | Customer understanding, context, and willingness to pay | Proven interest and indications of profitability | Proven model at limited scale |
| KPIs | Market size, customer evidence, problem/solution fit, opportunity size | Value Proposition evidence, financial evidence, feasibility evidence | Product/market fit, acquisition and retention evidence, business model fit |
| Experiment Themes | Desirability 50–80%, Feasibility 0–10%, Viability 10–30% | Desirability 30–50%, Feasibility 10–40%, Viability 20–50% | Desirability 10–30%, Feasibility 40–50%, Viability 20–50% |
An innovation portfolio helps you manage multiple bets simultaneously, distributing resources across ideas at different stages of maturity. Just as a venture capitalist does not invest everything in one startup, your organization should spread innovation investments across a portfolio and manage them with the discipline each stage demands.
Definition — The Three Portfolio Stages
Seed (under $50K, teams of 1–3, 20–40% time commitment): Focus on customer understanding, context, and willingness to pay. KPIs center on market size, customer evidence, and problem/solution fit. Experiment mix: 50–80% Desirability, 0–10% Feasibility, 10–30% Viability.
Launch ($50K–$500K, teams of 2–5, 40–80% time commitment): Focus on proven interest and early indications of profitability. KPIs center on Value Proposition evidence, financial evidence, and feasibility evidence. Experiment mix: 30–50% Desirability, 10–40% Feasibility, 20–50% Viability.
Growth ($500K+, teams of 5+, 100% time commitment): Focus on a proven model at limited scale. KPIs center on product/market fit, acquisition and retention evidence, and business model fit. Experiment mix: 10–30% Desirability, 40–50% Feasibility, 20–50% Viability.
Key Insight — Uncertainty and Funding
The relationship between uncertainty and funding moves in opposite directions. Early on, uncertainty and risk are very high while funding is low. As you generate evidence through experimentation, uncertainty drops — and funding should rise accordingly. This is why a venture-capital-style approach to incremental funding works: teams earn the right to more investment by reducing risk with evidence.
Teams need a governing body that funds experiments, removes obstacles, and makes decisions.
Designing the Committee
Committee Working Agreement
Six Areas to Monitor as a Committee
Principle — The Complete Framework
Design your team and shape your idea using the Business Model Canvas and Value Proposition Canvas. Test by hypothesizing, experimenting, learning, and deciding — using Test Cards, Learning Cards, and the Assumptions Map. Choose from an extensive experiment library of discovery and validation experiments, sequencing them for maximum learning. And adopt the mindset of avoiding pitfalls, leading through experimentation, and organizing your teams and portfolio for continuous testing. The goal, always, is to reduce the risk and uncertainty of new ideas with evidence — so you can progress from idea to business with confidence.