Testing Business Ideas in Manufacturing: When generic experiments don't apply

Every guide on testing business ideas assumes you are building software. Run a landing page test. Ship an MVP in two weeks. A/B test your value proposition on 10.000 visitors.

Try that in manufacturing. You cannot A/B test a production line. You cannot iterate weekly on a physical product that requires tooling, certification, and a six-month supply chain. And you cannot run a landing page test for a €2.000.000 industrial system that requires a buying committee of eight people to approve.

I have facilitated over 100 testing business ideas sessions with industrial and manufacturing companies. The methodology works. But the experiments need to be different. The timelines are different. The evidence thresholds are different. And the cost of testing the wrong assumption is not a wasted sprint. It is a wasted production line.

This article covers what changes when you apply business idea testing to manufacturing, and which experiment types actually work in industrial B2B.

For context on the full testing methodology, read the Testing Business Ideas practitioner’s guide. For how to design the business model and value proposition that your experiments will test, see the Business Model Canvas for manufacturing and the Value Proposition Canvas for manufacturing.

Why standard testing approaches fail in manufacturing

The Testing Business Ideas methodology from Strategyzer provides an experiment library of 44 experiment types. It is one of the best resources available. But roughly half of those experiments assume conditions that manufacturing companies do not have.

Here are the four structural constraints that change everything.

You cannot iterate fast on physical products. A software team can ship a new feature on Friday and measure results on Monday. A new industrial product takes 12 to 24 months from concept to certified production. That timeline means every experiment you run carries more weight. You get fewer attempts, so each one has to count.

Your customers do not click buttons. In B2B manufacturing, the purchase decision involves 4 to 8 stakeholders, takes 6 to 18 months, and runs through procurement processes that no landing page can replicate. A “sign up for early access” button tells you nothing about whether a buying committee will commit €500.000.

Capital commits early and stays committed. A SaaS company can pivot after a bad experiment by rewriting code. A manufacturer that invested €380.000 in new tooling based on an untested assumption cannot undo that decision. I worked with one industrial equipment company that made exactly this mistake. The customer segment they tooled up for turned out to be half the size they projected. Testing before tooling would have cost them €15.000 in customer interviews and concept validation. Not testing cost them twenty times that.

Regulatory approval adds a non-negotiable timeline. For manufacturers in food, chemicals, medical devices, or automotive, compliance certification can take 6 to 18 months on top of development. You cannot “launch and learn” when launching requires passing regulatory review. Testing the business model assumptions before entering the compliance pipeline is not optional. It is the only way to avoid burning time and money on a product the market does not want.

Which experiments work for testing business ideas in manufacturing

Not every experiment in the standard library applies to manufacturing. But more experiments work than most industrial teams realize. The key is knowing which ones to use and when.

Customer interviews (with the right people)

This is the single most underused experiment in manufacturing. Not customer satisfaction surveys. Not “would you buy this?” conversations. Structured interviews with the stakeholders who actually make buying decisions.

In manufacturing B2B, that means separate conversations with operations, engineering, procurement, and the budget holder. Each has different jobs, pains, and definitions of value. The operations manager wants uptime. Procurement wants total cost of ownership over 10 years. Engineering wants compatibility with their existing Siemens PLCs.

I facilitated a testing session with a manufacturer of industrial filtration systems who was developing a new product line. The team was convinced the primary customer need was higher filtration capacity. After eight interviews across four potential customers, the actual primary need turned out to be predictable maintenance scheduling. The product specifications they were engineering for were solving the wrong problem. Eight interviews, roughly €3.000 in travel costs, saved them from a €200.000 engineering investment aimed at the wrong target.

How to make it work in manufacturing:

• Interview stakeholders separately, not in a group (people censor themselves in front of colleagues)
• Ask about how they currently solve the problem, not whether they like your idea
• Focus on what they have spent money on before, not what they say they would spend money on
• Get specific: “What did your last supplier switch cost you?” is better than “Is switching costs a concern?”

Letters of intent

In B2B manufacturing, a letter of intent from a potential customer is stronger evidence than any prototype. It means someone with budget authority has put their name on a document saying they would commit to a purchase under specific conditions.

This is not the same as a verbal “yes, we are interested.” Manufacturing companies hear that all the time. It means nothing until procurement is involved.

One manufacturer I work with was testing a new predictive maintenance service for their equipment. Instead of building the full digital platform, they created a two-page concept document describing the service, the pricing model, and the expected outcomes. They presented it to five existing customers. Two offered to sign letters of intent for a pilot program. That evidence, two written commitments from real customers, was enough to justify the next €150.000 in platform development.

What a manufacturing letter of intent should include:

• The specific problem the customer expects the product to solve
• The conditions under which they would purchase (price range, delivery timeline, specifications)
• A commitment to participate in a pilot or place a first order
• A signature from someone with budget authority, not just from the champion

Supplier conversations as experiments

This is unique to manufacturing and rarely discussed in testing guides. Your supplier conversations are experiments in disguise.

When you discuss a new product concept with a potential supplier, you are testing feasibility assumptions: Can this component be manufactured at the required tolerance? What are the minimum order quantities? What is the lead time? What happens to unit cost at different volumes?

I have seen manufacturers treat supplier conversations as procurement activities. They are not. They are feasibility experiments. And the information you get back often changes the business model.

One agricultural equipment manufacturer discovered during supplier conversations that a key component they assumed would cost €45 per unit actually required a minimum order of 10.000 units at €28 per unit or 500 units at €67 per unit. That single data point changed the entire viability calculation. At the lower quantity, the product was unprofitable. At the higher quantity, they needed a customer base three times larger than their initial target. The business model assumptions shifted before a single prototype was built.

How to use supplier conversations as experiments:

• Bring your concept to suppliers early, before final specifications
• Ask specifically about minimum order quantities, lead times, and volume-dependent pricing
• Test whether your assumed cost structure holds at realistic production volumes
• Map single-source dependencies and identify alternative suppliers for critical components

Concept testing with visual assets

You do not need a functional prototype to test whether customers want what you are building. A 3D rendering, a technical drawing, or even a concept video can test the value proposition before you invest in engineering.

The goal is not to test whether the product works. It is to test whether the customer cares enough about the outcome to engage further.

One packaging machinery manufacturer tested a new product concept using three detailed 3D renderings and a one-page specification sheet. They showed it to seven potential customers at a trade show. Five requested follow-up meetings. Three asked for pricing. The team had spent €8.000 on renderings and travel. They would have spent €120.000 on a functional prototype that tested the same question: do customers want this?

When to use concept testing versus prototyping:

The mistake I see most often: jumping straight to functional prototypes. In software, prototyping is cheap. In manufacturing, it is an investment. Test whether the business model works before testing whether the product works.

Pilot programs with lighthouse customers

A pilot program is the manufacturing equivalent of a beta test. But unlike a software beta, a manufacturing pilot involves real equipment, real installation, and real risk for the customer.

That makes pilot customers harder to find. But it also makes their commitment much stronger evidence. A customer who agrees to install your unproven equipment on their production line is making a serious commitment.

How to structure a manufacturing pilot:

• Define what you are testing (the business model assumption, not the product) and what the fail criteria are before starting
• Limit the pilot to 2 to 3 customers maximum. More than that is a soft launch, not an experiment
• Set clear timelines: 3 to 6 months is enough to test most assumptions
• Measure what matters: not just technical performance, but willingness to continue, willingness to pay the target price, and willingness to refer others
• Document learnings systematically. A pilot without documented fail criteria is not an experiment. It is a free trial

The testing sequence for manufacturing

The order of experiments matters more in manufacturing than in software, because each experiment is more expensive. Here is the sequence I use with manufacturing teams.

Step 1: Test desirability first. Before anything else, find out whether customers want the outcome your product delivers. Customer interviews and concept documents are your cheapest tools. If there is no demand, everything else is waste.

Step 2: Test viability second. Once you know customers want it, test whether the business model works. Can you deliver it at a price customers will pay, with margins that sustain the business? Supplier conversations and back-of-napkin calculations with real data points go here.

Step 3: Test feasibility last. Only after you have evidence of desirability and viability should you invest in prototyping, engineering, and production development. This is where the big capital commitments happen. By this point, they should be backed by evidence, not assumptions.

Most manufacturing teams do the opposite. They start with feasibility (can we build it?), then viability (can we sell it?), and only discover desirability problems after the production line is running.

I worked with a chemical company that spent 18 months developing a new formulation before talking to customers. The product worked perfectly. Nobody wanted it at the price required for profitability. Eighteen months of R&D investment, zero revenue. Flipping the sequence, testing desirability first, would have taken six weeks and revealed the pricing problem before a single batch was produced.

Setting fail criteria for manufacturing experiments

The most critical step in testing business ideas for manufacturing, and the step that teams skip most often.

Fail criteria work better than success criteria because you cannot move the goalposts. “28% is close enough to 30%” is the sentence that kills manufacturing innovation projects slowly. Set the fail point. If results fall below it, the hypothesis is false.

For manufacturing, I use three methods to set fail criteria:

Back-of-napkin profitability. Work backward from the investment required. If your tooling costs €500.000 and you need 25% margins to justify the investment, how many customers do you need at what price point? Set your interview fail criterion at the customer interest level that makes those numbers impossible. If you need 20 customers in year one and only 2 out of 15 interviewed companies show genuine interest, that is a fail.

Industry analogs. Find comparable products or services in your industry and use their conversion rates as benchmarks. If the typical conversion from trade show lead to purchase order in your market is 8%, setting your fail criterion at 3% gives you a realistic floor.

Early adopter thresholds. Your first customers will be early adopters who are actively looking for solutions. If even they are not interested, the broader market will not be either. Set higher thresholds for early adopter experiments (60 to 80% positive response) than you would for general market tests.

To assess whether your organization has the structure and culture to run this kind of systematic testing, the Innovation Readiness assessment helps identify what is in place and what is missing. For the assessment adapted to industrial realities, see Innovation Readiness for manufacturing.

For the most common mistakes teams make when testing business ideas, read Business Experiment Mistakes: Why Most Teams Learn Nothing from Testing.

For a deeper look at multi-stakeholder validation, see Testing Business Ideas for B2B: Multi-Stakeholder Validation.

To manage a portfolio of tested manufacturing ideas with proper governance, see Innovation Portfolio Management for manufacturing.