The most expensive way to test a business idea is to build the whole thing and find out whether anyone pays. Census Bureau data on business survival shows roughly one in five new firms closes within its first year and about half by year five — and while causes vary, unvalidated demand sits near the top of every post-mortem list. Validation is the discipline of buying that information cheaply, before the lease, the loan, and the resignation letter. Orer News publishes information, not business advice.
The standard to hold yourself to is uncomfortable on purpose: compliments are not data, survey answers are not data, and waiting lists are weak data. Money from strangers is data.
What does real evidence look like?
Ranked from weakest to strongest: friends saying they would buy; an online poll; email signups from a landing page; refundable pre-orders; non-refundable deposits; repeat purchases. Each step up the ladder costs the customer more commitment, which is exactly why each step is more informative. A landing page that converts visitors to an email list at 5 percent tells you about curiosity; fifty $200 deposits against a delivery date tell you about demand. Aim your tests at the deposit level or above wherever the product allows.
How many customers is enough?
For most local and service businesses, the question is not a percentage but a count: can you find ten unconnected people who pay at a price that covers your costs? Ten paying strangers who found you without personal connection is a real signal; one hundred warm respondents is not. For software and consumer products, the proxy shifts to conversion — what fraction of people who experience the pitch transact — but the underlying discipline is identical: small numbers of committed payers beat large numbers of interested talkers.
What does a staged test look like in practice?
Stage one, two weeks: describe the offer in one page — the customer, the problem, the price, the delivery — and put it in front of the target audience through the channel you would actually sell in. Stage two, two to four weeks: sell manually. Take payment by invoice, fulfill by hand, deliver the service yourself a few times, run the pilot out of a rented kitchen or a spreadsheet. Stage three: repeat at a price that includes your real costs and a margin. Only what survives all three stages in the actual sales channel counts — a test run through your LinkedIn feed tests your popularity, not the market.
What kills otherwise good validations?
Three patterns recur. Leading the witness: describing the idea in glowing terms and asking whether people like it — they will. Friends-and-family rounds disguised as market tests: their money is charity with extra steps. And survivorship of the founder's enthusiasm: ignoring the twelve no's because the thirteenth person was enthusiastic. The discipline is to set the pass criteria before the test — how many paying customers, at what price, by what date — and let the market fire you from the idea without negotiating.
What should you keep from the job while testing?
Employment is the cheapest startup funding that exists: it bankrolls the test, carries your health insurance, and preserves the borrowing power you will need later. Test on nights and weekends within any non-compete and moonlighting constraints your employer imposes, and keep the test small enough that failure costs hundreds of dollars, not the household's savings. The job quits when the validated revenue, not the enthusiasm, says so.
- Evidence ladder: praise < signups < refundable pre-orders < deposits < repeat buys
- Ten unconnected paying strangers beats a hundred survey answers
- Sell manually before building at scale
- Set pass/fail criteria before the test begins
Validation is not about proving yourself right. It is about arranging for the market to correct you at the smallest possible cost — while you still have the salary to fund the correction.
For more context, read How to Scope a First Product You Can Actually Ship.
For more context, read seed pitch deck.
For more context, read Bootstrapping or Seed Round: What Each Actually Costs You.