A business idea often contains several untested claims disguised as one confident sentence. Customers have a problem. They care enough to change their behavior. The proposed product solves it. The price supports delivery. The business can reach those customers at an acceptable cost.
Building the product tests only part of that chain. A disciplined experiment identifies the uncertain link that could most change the decision to invest.
Make the assumption explicit
Consider a hypothetical scheduling tool for small service businesses. “Owners need better scheduling” is too broad. A testable claim might be that owners spend several hours each week resolving appointment conflicts and would pay a specified monthly amount to reduce that work.
Now the inquiry has a customer, a problem, a frequency, and an economic proposition. Each can be wrong independently. Some owners may have the problem but lack authority to buy. Others may have an inexpensive workaround that is good enough.
NSF’s I-Corps program uses direct customer and industry discovery to assess the commercial potential of technology. Its emphasis on conversations with potential customers and other stakeholders illustrates why a technical demonstration alone is insufficient. [1]
Ask about behavior before asking for approval
A question such as “Would you use this?” can invite a polite prediction. Ask instead how the problem was handled last time, how often it occurred, who was involved, and what the existing solution cost.
For the scheduling example, request a description of the current process without collecting unnecessary private customer information. The objective is to understand the work, not persuade the interviewee to validate the idea.
The SBA’s market-research guidance includes demand, market size, competing alternatives, and pricing. Those categories help turn an interesting conversation into a more complete picture of the opportunity. [2]
Choose the smallest informative test
If the main uncertainty concerns workflow, a simple prototype may be enough. If the uncertainty concerns willingness to pay, a price discussion or a clearly described paid pilot is more relevant. If it concerns delivery cost, the founder may need to perform the service manually and measure the time involved.
The test should match the uncertainty. A large mailing list does not establish that the product can be delivered profitably. A working prototype does not establish that a buyer will approve it. Each result answers a particular question.
For an illustrative pilot, the founder could specify the service, price, duration, and success measure in advance. Participants should understand what already exists and what is being tested. An experiment should not depend on misleading prospective customers.
Decide what the result will change
Before running the test, write the next decision. For example: proceed to a limited build only if participating firms use the core function repeatedly and delivery costs remain within a defined range.
Numbers should fit the stage and consequence of the decision. A handful of interviews can expose a mistaken assumption, but cannot reliably estimate the preferences of an entire market. A small pilot can reveal implementation problems without establishing long-term retention.
Record conflicting evidence as carefully as supportive comments. If several potential customers describe a different problem, that may justify revising the idea. It should not automatically be treated as a failure of communication.
The purpose of early experimentation is to make the next commitment more informed. A useful test can produce a decision to build, narrow the customer group, change the offer, or stop. Progress is the reduction of a consequential uncertainty, not simply the accumulation of encouraging activity.
Sources
Research, reflection, and a wider perspective.