Opportunity Discovery: A Practical Guide for When You're Stuck Looking for Problems Worth Solving
The Problem in Concrete Terms
You're sitting at your desk with a notebook open, or a blank Notion page, or a whiteboard covered in half-erased scribbles. You want to build something—a product, a service, a business—but every idea you write down feels either obvious and crowded, or vague and unvalidated. You've read about "finding your niche" and "scratching your own itch" and "talking to customers," but when you actually try to do it, the process dissolves into noise.
Concretely, the stuckness looks like this:
- You have a list of 20+ ideas but no framework for killing 19 of them
- You've interviewed three people who said "that sounds cool" and you have no idea if that means anything
- You keep noticing problems in your daily life but can't tell which ones are business problems versus annoyance problems
- You've built two MVPs that nobody used and you're gun-shy about starting a third
- You spend more time researching "how to find ideas" than actually testing them
The core issue isn't creativity. It's that opportunity discovery is treated as a lightning-strike moment when it's actually a systematic investigation process. You're trying to have insights instead of running experiments.
Why the Usual Fixes Fail
1. "Scratch your own itch"
This works if your itch is shared by thousands of people who have budget and urgency. Most itches aren't. You end up building a tool for a problem only you have, or one that people solve with a spreadsheet and don't care enough to pay for.
2. "Talk to customers"
Generic advice. Talk to which customers? About what? With what script? Most people ask "would you use this?"—a question that produces polite lies. They don't know how to distinguish signal from politeness, or how to structure conversations that reveal actual behavior rather than hypothetical intent.
3. "Look for problems in communities"
You join subreddits, Discords, Slack groups. You see complaints. But complaints ≠ demand. People complain about things they'll never pay to fix. You collect a list of grievances with no way to prioritize or validate them.
4. "Follow trends / emerging tech"
AI, Web3, no-code, whatever's next. This is solution-first thinking. You're holding a hammer looking for a nail. The opportunity isn't the technology; it's the constraint the technology removes for a specific group of people.
5. "Build in public / ship fast"
Shipping fast is good. Shipping the wrong thing fast is just efficient waste. Without a discovery process that filters before you build, you're just generating more failed MVPs.
Step-by-Step Solution: The Opportunity Discovery Loop
This process takes 3–6 weeks if you're doing it part-time. It produces a shortlist of 1–3 opportunities with evidence, not guesses.
Phase 1: Define Your Search Space (Days 1–3)
Don't start with ideas. Start with constraints.
Write down your non-negotiables. These are filters, not preferences:
- Time: Hours per week you can realistically commit (be honest)
- Capital: Cash you can put in without stress
- Skills: What you can build/sell/do now, not what you could learn
- Network: Who you can reach in 48 hours with a warm intro
- Interest areas: Domains you'll still care about in 18 months
- Dealbreakers: B2C, hardware, regulated industries, long sales cycles, etc.
Output: A one-page "founder fit" document. Every opportunity you evaluate gets scored against this. If it fails two constraints, discard it immediately.
Phase 2: Map Problem Territories (Days 4–10)
Pick 3–5 territories to explore. A territory is a context, not a problem. Examples:
- "Independent insurance adjusters handling catastrophe claims"
- "Shopify merchants doing $100K–$1M/year with custom products"
- "Facilities managers at multi-site restaurant chains"
- "Senior engineers at Series B SaaS companies managing on-call"
Criteria for good territories:
- You can name 20+ specific roles/companies in it
- They spend money on software/tools/services already
- You can reach them (cold email, warm intro, community, event)
- The work is repetitive, regulated, or revenue-critical
For each territory, do this:
- List 20 specific people (names, titles, companies, LinkedIn URLs). Not "product managers"—"Sarah Chen, PM at Datadog, leads incident response tooling."
- Find where they congregate: Specific subreddits, Slack groups, newsletters, conferences, Twitter lists, YouTube channels.
- Collect 50+ raw observations: Spend 2 hours per territory lurking. Copy-paste exact quotes. Screenshot workflows. Note tools they mention. Capture complaints, workarounds, "how do I..." questions, vendor complaints, spreadsheet screenshots.
Output: A territory dossier for each—20 names, 5 watering holes, 50 observations. No analysis yet. Just raw material.
Phase 3: Run Problem Interviews (Days 11–25)
Goal: Find problems that are urgent, frequent, expensive, and unsolved.
Recruit: Email your 20 names per territory. Template:
Subject: Quick question about [specific workflow you noticed]
Hi [Name],
I'm researching how [title] at [company type] handle [specific process]. Saw you've been at [Company] for [time]—figured you'd have perspective.
Not selling anything. 15 minutes, I'll send you a $50 gift card for your time.
Worth a quick call?
Interview script (stick to this):
- Context: "Walk me through the last time you did [process]."
- Tools: "What tools do you use? What do you hack together?"
- Pain: "What's the worst part? When did it last go wrong?"
- Cost: "How much time/money does that problem cost you? Your team? The company?"
- Attempts: "What have you tried to fix it? Why didn't it stick?"
- Budget: "If someone solved this reliably, who signs off on spending $X/month?"
- Trigger: "What would have to be true for you to switch tomorrow?"
Rules:
- Record (with permission). Transcribe later.
- Zero pitching. Zero "would you use..."
- Ask "tell me more" 3x per interview.
- 5 interviews per territory minimum. Stop when you hear the same 3 problems repeated.
Output: Interview transcripts + a problem scoring sheet per territory.
Phase 4: Score and Filter (Days 26–30)
For each problem discovered, score 1–5 on:
| Dimension | Question |
|---|---|
| Urgency | Is this a "hair on fire" problem or a "nice to fix"? |
| Frequency | Daily? Weekly? Monthly? Quarterly? |
| Cost | What's the fully loaded cost (time × salary + tool spend + error cost)? |
| Budget ownership | Does the person feeling the pain control budget? |
| Workaround quality | How bad is the current solution? (Spreadsheet = good. Legacy vendor = great.) |
| Switching trigger | Is there a clear event that forces change? (Audit, growth, regulation, vendor sunset) |
| Founder fit | Does this match your constraints from Phase 1? |
Threshold: Keep only problems scoring ≥28/35 and ≥4 on Urgency, Cost, and Budget Ownership.
Output: 1–3 validated problem statements with evidence:
"[Role] at [company type] lose $[amount]/month because [specific failure] during [trigger event]. They currently [workaround]. They've tried [failed solutions]. Budget sits with [title]. We can reach them via [channel]."
Phase 5: Micro-Validation (Days 31–40)
Before building, prove demand with one of these:
Option A: The Concierge Test
- Find 3 people from your interviews
- Offer to solve the problem manually for 2 weeks
- Charge 50% of your target price
- Deliver the outcome yourself (spreadsheets, scripts, phone calls, whatever)
- If they pay and renew, you have a business. If they
Top comments (0)