Opportunity Discovery: A Practical Guide for When You're Stuck
The Problem in Concrete Terms
You have the capacity to build. You have time, maybe some savings, perhaps a skill set that people pay for. What you don't have is a clear signal on what to build. Every idea feels either too small to matter, too crowded to enter, or too vague to validate. You've read about "finding your niche" and "scratching your own itch," but your own itches are either already solved by existing tools or too idiosyncratic to support a business. You've tried listing problems you encounter daily, but the list either runs dry after five items or balloons into hundreds of minor annoyances with no clear path to revenue. You're not lacking creativity—you're lacking a reliable filter that separates "this could work" from "this is a waste of six months."
The stuckness shows up in specific ways: you open a notebook or Notion page titled "Ideas" and stare at it. You scroll Product Hunt or Indie Hackers and feel a mix of envy and paralysis—everything seems taken. You've maybe built two or three side projects that died because you lost conviction halfway through, not because the code was hard. You can't tell if you're being too picky or not picky enough. The result is inertia disguised as research: reading more newsletters, bookmarking more frameworks, waiting for a "eureka" moment that doesn't arrive.
Why the Usual Fixes Fail
1. "Scratch your own itch" assumes your itches are marketable. Most developers' personal problems are either already solved (password managers, note-taking apps, habit trackers) or too niche to sustain a business (a CLI tool for your specific Kubernetes workflow). The advice works for a subset of people whose professional context exposes them to unsolved, high-value problems. If you're not in that subset, it's a dead end.
2. "Talk to customers" presumes you know who they are and what to ask. Cold outreach without a hypothesis wastes everyone's time. You end up having polite conversations that yield no actionable insight because you're fishing in the wrong pond with the wrong bait. The advice also ignores that access to the right people is often the hardest part—if you could easily reach 20 VPs of Engineering at Series B companies, you'd already have a network that surfaces opportunities.
3. "Look for complaints on Reddit/Twitter/forums" surfaces noise, not signal. People complain about everything. Distinguishing "I'd pay to fix this" from "I'm venting" requires context you don't have. The loudest complaints are often about problems with no budget (consumer annoyances) or problems that are symptoms of organizational dysfunction no software can fix.
4. "Follow trends" (AI, Web3, no-code, etc.) makes you a late entrant. By the time a trend is legible enough to write a blog post about, the low-hanging fruit is gone. You're competing with funded teams who started six months earlier. Trend-chasing also produces shallow ideas—wrappers around APIs that the platform will eventually absorb.
5. Frameworks like "Jobs to Be Done" or "Value Proposition Canvas" are analysis tools, not discovery tools. They help you structure an opportunity you've already found. They don't tell you where to look. Applying them to a blank page produces performative busywork: filling boxes with guesses.
Step-by-Step Solution
Phase 1: Map Your Asymmetric Advantages (Days 1–2)
Opportunity discovery isn't about finding a generic "good idea." It's about finding an idea you are uniquely positioned to execute and defend. Start by documenting your specific advantages—assets that would take a competitor months or years to replicate.
Do this now:
List every professional role you've held for 6+ months. For each, write: the core workflow, the tools used, the internal jargon, the metrics that mattered, the complaints that never got fixed, the manual workarounds people built in spreadsheets or scripts. Don't filter. A junior QA role at a logistics company might reveal more than a senior PM role at a SaaS unicorn.
List every industry you have secondhand access to. Spouse, siblings, close friends, former classmates, parents. What do they do all day? What software do they hate? What processes involve printing PDFs, re-entering data, or emailing spreadsheets? You don't need to know the industry—you need a trusted informant who does.
List your technical differentiators. Not "I know React." Instead: "I've built three Chrome extensions that scrape dynamic sites," "I've optimized Postgres queries for 100M+ row tables," "I've integrated with five different ERPs." Specific implementation experience beats general competence.
List your distribution channels. Newsletter subscribers, Twitter followers, Slack/Discord communities you're active in, former coworkers now at other companies, conference contacts. Even 50 relevant people is a channel. 10,000 random followers is not.
-
Score each asset on three dimensions (1–5):
- Depth: How much insider knowledge do you really have?
- Access: Can you get 5–10 conversations with decision-makers in two weeks?
- Pain intensity: Do you know of a problem people actively complain about and spend money/workaround on?
Keep only combinations scoring 10+. This is your search space. Everything outside it is a distraction.
Phase 2: Run Structured Discovery Conversations (Days 3–14)
You're not "chatting." You're running a repeatable interview process to test whether a problem is real, urgent, and funded.
The protocol:
Recruit 8–12 people per opportunity area. Use your access list. Ask for 20 minutes. Frame it: "I'm researching [specific workflow] for a potential tool. Not selling anything. Your experience would save me months of wrong turns." People say yes to helping, not to being pitched.
-
Use the same 7-question script every time. Deviating makes patterns invisible.
- "Walk me through the last time you did [workflow]."
- "What tools or workarounds did you use?"
- "Where did it break or slow down?"
- "What did you do when it broke?"
- "Have you looked for a better tool? What happened?"
- "If this problem were solved, what would change for you/your team?"
- "Who else should I talk to?"
Record and transcribe. Otter.ai, Fireflies, or manual notes. You will not remember accurately.
-
After each batch of 3–4 interviews, write a one-page memo:
- Problem statement in their words
- Current solutions and why they fail
- Evidence of budget (existing spend, headcount allocated, failed build attempts)
- Decision-maker title and org size
- Your confidence score (1–10) that this is a real, paid problem
Kill or continue. If confidence < 7 after 8 interviews, archive the area. Move to the next asset combination. Do not "pivot the idea"—pivot the search.
Phase 3: Validate Willingness to Pay (Days 15–28)
Conversations reveal pain. Only commitment reveals value.
Do this for each surviving opportunity (max 2):
Define the MVP scope ruthlessly. One workflow. One user type. No settings, no integrations, no admin dashboard. What is the smallest thing that delivers the "aha" moment? Write it as a user story: "As a [role], I want to [action] so that [outcome]."
Build a clickable prototype. Figma, Framer, or even a Loom video walking through mockups. No code. One day max.
Run 5–8 "pre-sale" calls. Same people or new ones from your network. Show the prototype. Ask: "If this existed today at $X/month per seat, would you put it on a company card?" Note: not "would you use it"—would you pay. $X should be your target price (start at $50–200/mo for B2B).
Collect commitments, not compliments. A commitment is: "Send me the invoice," "Introduce me to my manager," "I'll pilot with my team of 5." A compliment is: "This looks great, let me know when it launches." Only commitments count.
Threshold: 3 paid commitments or 2 pilot agreements. If you hit it, build. If not, return to Phase 1 with what you learned. The interviews weren't wasted—the
Top comments (0)