A new experiment can feel productive long before it meets a customer. You can spend a week polishing a logo, organizing tools, and tweaking a page. The work may be necessary, but activity alone does not tell you whether the idea is useful.
Write the question first
Choose something the test can actually answer. “Can this become a business?” is too broad for one week. “Will independent consultants discuss paying for a better proposal template?” is a question you can take into a small set of conversations.
Set a time limit and a spending limit before you begin. Define the action you want to observe and the decision you will make afterward. The point is to learn enough to choose the next step.
A scorecard you can copy
Keep signals in perspective
A page view tells you someone arrived. An email signup shows a little more interest. A specific question about delivery or a paid pilot gives you different evidence. Record these separately so a large, easy-to-get number does not hide a small number of meaningful conversations.
You can learn from a test with no sales. Perhaps the audience does not recognize the problem, the offer is hard to understand, or you reached people who will never buy. Write down what you know and what is still an assumption.
After a sale, count delivery
Track direct costs, refunds, and the hours required to complete the work. Revenue is only one part of the story. A service that needs repeated unpaid revisions may need a tighter scope. A product with good initial interest may still need support and distribution work.
At the end of the test, make a decision. Continue when the evidence supports another round. Change a specific part when you have learned what is missing. Stop when the likely value no longer justifies the effort. Each is a useful outcome.
Your next move
Copy the scorecard, choose one question, and set a review date before spending another week on your idea.