Validate the idea before you decorate it
A logo can be gorgeous while the business underneath it quietly dies. Validation means proving that a specific customer has a specific problem and will take a meaningful step toward your solution.
Use three kinds of evidence: market evidence, problem evidence, and buying evidence. Do not call the idea validated until the intended customer—not your supportive cousin—takes an action that costs time, reputation, access, or money.
Write one claim you can actually disprove
“A wellness business for busy women” is fog. Try: “Independent hairstylists who lose late-cancellation income will pay $29 a month for automated deposits and reminders.” Now you can identify the buyer, the expensive moment, the alternative, the price, and the behavior that would prove you wrong.
Write down five assumptions: who has the problem, how often it occurs, what it costs now, how people solve it today, and why they would switch. Rank them by danger. The assumption most likely to kill the business gets tested first.
Use market research to avoid interviewing a fantasy
The SBA separates market research from competitive analysis for good reason. Market data can show how many plausible customers exist, where they are, and what they already buy. Competitor research shows the price and inconvenience people already tolerate. Neither proves they will buy from you, but both stop you from inventing a customer who exists only in a Canva mood board.
- Search direct competitors, substitutes, DIY workarounds, and “do nothing.”
- Read one-star and three-star reviews to find switching triggers and unmet expectations.
- For a local business, use Census Business Builder to examine population, income, industry, and competing establishments around the actual location.
- Record real prices, delivery times, guarantees, and customer complaints in one comparison sheet.
Ask about the last time, not the imaginary future
Founders repeatedly describe the same Reddit autopsy: friends loved it, survey respondents said they would buy it, and launch day produced silence. The questions invited politeness. Ask, “Tell me about the last time this happened,” “What did you do next?” “What did that cost?” and “What have you already tried?” A person who has never attempted a workaround may dislike the problem without valuing a solution.
Avoid pitching during the first half of the conversation. Your job is to learn the trigger, current behavior, budget owner, objections, and words customers use. Ten interviews with the right people beat a hundred responses from whoever happened to click.
Move from conversation to an honest offer
Interest is a clue. A buying attempt is evidence. Offer the smallest deliverable you can fulfill responsibly: a paid pilot for a service, a limited preorder with clear dates and refund terms, a deposit-backed booking, a manual concierge version of software, or a product sample sold at a price that resembles the future business.
| Signal | What it tells you | What it does not tell you |
|---|---|---|
| Likes, votes, compliments | The message caught attention | That the problem is urgent or worth paying for |
| Email signup | Someone will exchange contact information | That the eventual price or product will convert |
| Interview + recent workaround | The problem exists in behavior | That your offer wins |
| Deposit, preorder, paid pilot | The offer crossed a real commitment threshold | That delivery, margin, retention, or scale works |
| Repeat purchase or referral | Value survived first use | That the whole market behaves the same way |
Score the idea on evidence, not enthusiasm
Give each claim a 0, 1, or 2: no evidence, indirect evidence, or direct behavior. Score customer access, problem frequency, current spending, offer conversion, delivery feasibility, and unit economics. A weak score does not require a funeral; it tells you what to test next. A strong score justifies a small commitment, never a spending spree.
If nobody in the intended market will discuss a recent example, show a workaround, introduce you to the budget owner, join a relevant pilot, or try to buy, you do not yet have demand evidence. You have a theory.
Your finish line
Stop this stage when you can name the buyer, trigger, current alternative, tested offer, observed action, and next risky assumption. Then run a cheap test. If you are still choosing fonts, return to the customer.