Opportunity Discovery: A Practical Guide for Builders Who Keep Building the Wrong Things
The Problem in Concrete Terms
You have the skills to build. You've shipped features, maybe whole products. But you keep hitting the same wall: you don't know what to build next that people will actually pay for.
This shows up in specific ways:
- You have a notebook full of ideas but no framework to decide which one deserves your next three months
- You've built things that got "cool!" reactions but zero paying customers
- You spend weeks researching a market only to realize the problem isn't painful enough, or the budget doesn't exist, or you'd be the tenth entrant with no differentiation
- You default to scratching your own itch, but your problems turn out to be too niche or too idiosyncratic
- You see competitors raising money in a space and can't tell if that means "validated opportunity" or "bubble about to burst"
The core issue isn't creativity. It's lack of a repeatable process to separate signal from noise before you commit significant time.
Why the Usual Fixes Fail
"Just talk to customers"
Everyone says this. Almost nobody explains which customers, what to ask, or how to interpret answers. You end up with polite encouragement ("Yeah, I'd use that!") that evaporates when you ask for a credit card.
"Build an MVP and iterate"
This works if you're already in the right ballpark. If you're solving a non-problem for a non-buyer, iterating just produces better versions of something nobody wants. You waste months optimizing the wrong thing.
"Follow your passion"
Passion sustains you through the grind, but it doesn't create demand. Many passionate founders build for audiences that don't exist or can't pay.
"Look at trending topics on Twitter/Product Hunt/Reddit"
By the time a trend is visible on these platforms, the easy opportunities are gone. You're seeing the survivors, not the graveyard of similar attempts.
"Use a framework like Jobs-to-be-Done or Value Proposition Canvas"
Frameworks are useful after you have raw data. They don't tell you where to look or how to get that data. Most people fill them with assumptions, not evidence.
Step-by-Step Solution
This process takes 2–4 weeks of part-time work. It requires no budget, no special tools, and no permission. It produces a ranked shortlist of opportunities with evidence behind each.
Phase 1: Define Your Constraints (Day 1)
Before hunting opportunities, define your personal filter. This prevents you from chasing things you can't or won't execute.
Write down honest answers to:
- Time horizon: How many months can you work without revenue? (Be specific: "6 months" not "a while")
- Capital: How much of your own money will you put in before requiring external funding or revenue?
- Skills: What can you actually build well right now? Not "what could I learn." What can you ship in 4–8 weeks?
- Distribution advantage: Do you have access to any specific audience, community, or channel? (Newsletter, Slack group, former coworkers, industry forum, etc.)
- Deal-breakers: What business models, markets, or work types will you not do? (Enterprise sales, consumer apps, hardware, regulated industries, etc.)
Keep this document. You'll use it to kill ideas fast.
Phase 2: Generate Candidate Problem Areas (Days 2–5)
Don't brainstorm solutions. Collect observed problems. Aim for 30–50 raw entries.
Source 1: Your own work history (10–15 problems)
List every repetitive, manual, error-prone, or hated task you've seen in past jobs. Include tasks you automated poorly with scripts/spreadsheets. Note the role, company size, and industry.
Source 2: Adjacent roles (5–10 problems)
Talk to 3–5 people in roles adjacent to yours (sales if you're engineering, support if you're product, etc.). Ask: "What's the most annoying part of your week that involves software or data?" Write down verbatim quotes.
Source 3: Public complaint mining (10–15 problems)
Search these patterns on Reddit, Twitter, LinkedIn, niche forums:
- "I hate having to [manual task] every [time period]"
- "There's no good way to [specific workflow]"
- "We pay $[amount] for [tool] but still have to [workaround]"
- "Why does [process] still require [manual step]?"
Capture: exact quote, platform, date, apparent role/industry of poster.
Source 4: Tool review mining (5–10 problems)
Go to G2, Capterra, Product Hunt reviews for tools in spaces you understand. Filter for 2–3 star reviews. Look for: "Missing [feature]," "Doesn't integrate with [X]," "Too expensive for [use case]," "Support is slow for [critical issue]."
Output: A spreadsheet with columns: Problem Description, Source, Role/Industry, Frequency Hint, Your Constraint Fit (Y/N/Maybe).
Phase 3: Rapid Filtering (Days 6–7)
Apply your constraints document. Mark each row Y/N/Maybe on Constraint Fit. Kill anything that's a hard N.
For Maybes and Ys, score 1–5 on:
- Pain clarity: Is the problem specific and acknowledged? (1 = vague dissatisfaction, 5 = "I lose $X/week because of this")
- Budget evidence: Is there proof money moves? (1 = "would be nice," 5 = "we pay $X for a bad workaround")
- Frequency: Daily/weekly = 5. Monthly = 3. Quarterly = 1.
- Your build ability: Can you ship a v1 in your time horizon with current skills? (1 = need to learn stack, 5 = can build this weekend)
Keep only rows scoring ≥12/20. You should have 5–10 candidates.
Phase 4: Evidence Collection (Days 8–21)
For each remaining candidate, do three things. Do not skip. Do not batch—finish one candidate before starting the next.
1. Find 5 people who have this problem right now
Not "might have it." People actively experiencing it.
- Use your network, LinkedIn search, relevant Slack/Discord communities, subreddit members
- Message template: "I'm researching [specific problem] for [role/industry]. Not selling anything. Would you spare 15 minutes to walk me through how you handle it today? Happy to share what I learn."
- Aim for 5 conversations per candidate. If you can't find 5 people in a week, the market is too small or too hidden—kill it.
2. Run a structured interview (15–20 min each)
Script. Do not improvise.
1. "Walk me through the last time you encountered [problem]. What happened?" (Listen for steps, tools, workarounds, frustration points) 2. "What did you try to solve it? What happened?" (Reveals existing solutions, switching costs) 3. "How much does this problem cost you—time, money, errors, missed opportunities?" (Force quantification) 4. "If this problem were solved tomorrow, what would change for you?" (Reveals value ceiling) 5. "Who else on your team/in your network deals with this?" (Referral path) 6. "What would make you *not* buy a solution?" (Uncovers deal-breakers: compliance, integration, price, trust) Record (with permission) or take detailed notes. Tag quotes with: Pain, Workaround, Budget, Dealbreaker, Referral.
3. Test willingness to pay before building
At the end of each interview, if the problem is real and painful:
"I'm exploring building a solution that [one-sentence value prop]. It would cost roughly $[price] per [month/seat/transaction]. If this existed today, would you put down a $50 deposit to get early access and lock in that price?"
This is the only signal that matters. Not "would you use it." Not "is this interesting." A deposit (refundable, held in Stripe or even a Google Form + manual follow-up) proves intent.
Track: Candidate, Interviews Done, Deposits Collected, Common Dealbreakers, Price Resistance Point.
Phase 5: Synthesis & Selection (Days 22–25)
Create a one-pager per surviving candidate:
| Field | Content |
|---|---|
| Problem (one sentence) | |
| Target buyer (title, company size, industry) |




