Black Tech Startup 01: Prove the Problem Before You Build the Product

This is Part 1 of 6 in the Black Tech Startup series. Work the sequence in order: each guide assumes you completed the operating decisions in the previous one.

Start with the economic truth, not the idea

Black founders do not need another speech about believing in an idea. The first job is to protect scarce time, cash and ownership by proving that a specific buyer has a problem serious enough to change behavior or spend money. That discipline matters in any startup, but it matters even more when outside capital is less predictable. Federal Reserve small-business research continues to document uneven financing conditions across owner groups, while Census data show Black entrepreneurship is large in absolute numbers but still underrepresented among employer firms. The practical response is not pessimism; it is evidence. Make the business earn the right to consume your money. A technology idea becomes a startup opportunity only when four things line up: a painful problem, a reachable buyer, an urgent reason to act and a plausible way for your product to get paid.

Write a problem thesis that can be disproved

Before interviews, write a one-page problem thesis. Name one buyer, one recurring situation, one costly consequence and one existing workaround. “Small businesses need AI” is useless because it cannot be tested. “Independent dental practices with two to five providers lose scheduled revenue when same-week cancellations are not refilled, and office managers currently handle recovery manually by phone and text” is testable. Add what you believe the problem costs in dollars, time, risk or missed revenue. Then write what evidence would prove you wrong. If ten qualified buyers tell you the problem is rare, already solved or not worth paying to fix, you should be willing to narrow or kill the idea. A thesis that cannot lose is not research; it is self-protection.

Interview behavior, not opinions

NSF I-Corps is built around customer discovery because founders routinely overestimate what people will do from what they say they like. Do not ask, “Would you use this?” Ask what happened the last three times the problem occurred. Who noticed it? What did they do? What did it cost? Who approved the workaround? What happens if nothing changes? What software, contractor or employee is already involved? When did they last spend money trying to fix it? The best interview is not a pitch. You are reconstructing a real workflow and locating the budget, authority and urgency inside it. Aim first for 20 to 30 conversations with people who match the same buyer profile. If you keep changing industries, company sizes and job titles, you will collect anecdotes instead of a market signal.

Build an evidence ledger, not a pile of notes

Create a simple evidence ledger with one row per interview. Track buyer type, problem frequency, economic consequence, current workaround, current spend, decision maker, buying trigger, objections and whether the person agreed to a concrete next step. Separate facts from your interpretation. Score each conversation on pain, urgency, authority and willingness to commit. The commitment ladder matters: giving advice is weak evidence; introducing you to the budget owner is stronger; sharing internal workflow data is stronger; agreeing to a paid pilot is stronger still. Your goal is not a high average “interest” score. Your goal is to see whether the same pattern repeats often enough that you can describe the buyer and problem without hand-waving.

Map the real competition: the budget you must displace

Your competitor is not only another startup. It can be an employee doing the work manually, a spreadsheet, a consultant, a feature inside software the customer already owns, or the decision to tolerate the problem. For every interview, record what would have to stop, shrink or change for your product to get budget. Then compare alternatives on total cost, speed, risk, switching effort and measurable outcome. This creates a sharper value proposition and reveals whether you are entering a budget category that already exists or asking the buyer to invent one. Selling into an existing budget is usually easier. If no budget exists, your evidence must be strong enough to show that the cost of doing nothing is larger than the cost and disruption of adopting you.

Use public data to test whether the niche is big enough to matter

Customer interviews tell you whether the pain is real; public data helps tell you whether there are enough plausible buyers. Use Census Business Builder, County Business Patterns and data.census.gov to estimate how many establishments fit your target industry and geography. Use NAICS codes carefully: the right code is a starting point, not proof that every firm is a customer. Build three market views. Bottom-up: number of reachable buyers multiplied by an evidence-based annual price. Beachhead: the segment you can realistically sell in the first 24 months. Expansion: adjacent segments that share the same workflow after you have proof. Avoid giant top-down “AI market” numbers. Investors and experienced operators can see immediately when a founder has multiplied an industry headline by an arbitrary percentage.

Run a paid-problem test before a product test

The fastest way to learn whether a software idea has economic value is often to deliver the outcome manually before automating it. If the product is supposed to qualify construction leads, run the qualification service yourself using forms, spreadsheets and existing tools. If it is supposed to automate compliance evidence, assemble one evidence package manually. Charge if you can. A paid concierge pilot teaches you what the customer actually values, where information is missing, what exceptions destroy the neat workflow and who must approve a purchase. It also prevents a common founder mistake: spending six months engineering features for a process nobody has agreed to buy. Manual delivery is not the business model; it is an instrument for discovering the business model.

Set kill, narrow and go criteria before emotion takes over

At the start of discovery, define thresholds. A reasonable early gate might require repeated evidence of the same painful workflow from at least 15 qualified interviews, at least five buyers willing to take a concrete next step, and at least two willing to test a paid or contractually defined pilot. Your numbers can differ, but decide them before you fall in love with the solution. “Kill” means the problem is weak, rare or unreachable. “Narrow” means the problem is strong but only for a smaller customer group or use case. “Go” means pain, authority, urgency and commitment are converging. Do not confuse compliments with progress. The company begins when the market starts pulling work out of you.

The Black founder advantage is specificity, not a race-based assumption

A Black founder may see problems, networks or customer behavior that larger incumbents have ignored. That can be a real informational advantage, but it still has to survive customer discovery. Do not assume that a community connection automatically creates willingness to pay, and do not limit yourself to a “Black market” unless the problem is actually defined that way. Use lived experience as a source of hypotheses and access—not as a substitute for evidence. The strongest position is when you can say: we understand this workflow unusually well, we can reach the buyer efficiently, and we have proof that solving it creates measurable value. That is a business advantage any investor, customer or partner can understand.

Your 30-day deliverable

By day 30 you should have a written problem thesis, a defined buyer profile, 20 to 30 documented interviews, a market-size worksheet, a map of current alternatives, a commitment ladder and a decision memo: kill, narrow or proceed. If proceeding, name the single outcome your first product must deliver and the smallest proof that a customer would pay for it. Do not incorporate a complicated company, hire a development team or buy a large technology stack just to feel like a founder. Guide 2 begins only after the problem has earned the next investment.

Black Tech Startup series

Your six-part path to an operating company

Source desk

Research behind this guide

Use the primary sources below to verify current rules, eligibility and program details before acting. Program terms can change.