A founder spends four months building a service. Friends and contacts were enthusiastic and several said they would definitely use it.
It launches. Two of the enthusiastic people buy. The others were being encouraging, which is what people do when asked a hypothetical question by somebody they like.
The four months were the expensive part. Two weeks of trying to sell the thing before building it would have produced the same information and left her with most of a quarter.
Step 1: describe the thing before you build it
One page setting out what it is, who it is for and what it costs. If you cannot write this clearly, the problem is the idea rather than the page.
Step 2: put a price on it immediately
Demand at no price is not demand. The price is what converts a pleasant conversation into information, which is why it belongs in the first test.
Step 3: try to sell it to twenty people directly
Not a survey. Actual conversations with people who would plausibly buy, ending in an ask. Twenty conversations produce a clear answer within a fortnight.
Step 4: ask for a commitment rather than an opinion
A deposit, a pre order, a booking or a written intention. Anything that costs the other person something is a signal, and anything that does not is politeness.
Step 5: count the people who ask when
Unprompted questions about availability are the strongest positive signal available, and they are considerably more reliable than any stated intention.
Step 6: watch what people offer to pay
Somebody who negotiates has accepted the premise and is discussing terms. Somebody who says it sounds great has not engaged with it at all.
Step 7: run a small paid test where it fits
A modest spend against a page with a real call to action produces conversion data in days, and it tests the message as well as the demand.
Step 8: sell it before you can deliver it
With an honest timescale. Taking real commitments for something that will exist in six weeks is legitimate, informative and how a great many good businesses start.
Step 9: build the smallest deliverable version
Enough to serve the first customers properly. Everything beyond that should wait until real customers tell you what they actually need.
Step 10: listen to the objections, not the compliments
The reasons people give for not buying are the product roadmap. Compliments from people who did not buy are worth nothing at all.
Step 11: test the price as well as the idea
An idea that works at one price and not at another is a different business, and this is discovered cheaply now and expensively later.
Step 12: record every conversation
Who, what they said, what they objected to and what happened. Twenty recorded conversations are a genuine research asset, and considerably more useful than a memory of how the fortnight felt.
Step 13: check the page is doing its job
Before concluding there is no demand, check that the page answers price, applicability and timescale, which is the failure in getting found before anybody knows your name.
Step 14: capture the interest while you are still testing
An assistant on the page that answers the questions, records who asked and captures the name and email of everybody ready to commit, so you can ask for the deposit yourself, turns a test into a list of real customers before anything has been built. Checking it behaves correctly first is covered in how to test your chatbot before going live, and the same pre booking reassurance appears in beard and grooming service questions before booking. See features, start at try, and the sector view sits under startups.