Choose a narrow starting group
The first few users should have enough in common that their experiences can be compared. If each person has a different job, budget and definition of success, every request can look equally urgent. A narrower group helps a founder distinguish a recurring problem from a one-off preference.
Start with a situation, not only a demographic. People who are preparing the same kind of report or managing the same difficult handoff may share a workflow even if their job titles differ. Describe the trigger that makes them look for help.
Recruit around work that is already happening
Ask a prospective user to show how they handled the problem last time. The documents, delays and workarounds in that account are often more useful than a prediction about future behavior. If the task will happen again soon, there is an opportunity to test the product against something real.
Give the trial a clear shape. Agree on one task, the expected result and a time to review what happened. Avoid asking for an open-ended commitment to ‘try the platform.’ A small, specific commitment is easier to make and easier to learn from.
Keep support from hiding the product
Early founders should be close to users. But personal attention can make a weak product feel useful. Track where a founder is doing work on the user’s behalf and whether the result still holds when that help is reduced.
A simple observation log can separate three things: what the user did, what the product did and what the founder supplied. That distinction helps a team decide whether it has discovered a product opportunity, a service opportunity or a problem that still needs a better solution.
Earn the next introduction
A returning user is useful evidence when the return is connected to a recurring job. An introduction is useful when the next person has the same problem. Neither should be reduced to a vanity metric. Ask why the person returned and what they told the person they referred.
The aim is to find a repeatable account of value: a recognizable person faces a recognizable problem and chooses the product because it helps in a specific way. Once that pattern starts to hold, distribution experiments become easier to design. Before then, adding more people can simply multiply the confusion.
Use the first ten relationships to discover a repeatable reason to choose the product.