Opportunity Discovery: A Field Guide for When You’re Stuck Guessing
The Problem in Concrete Terms
You have capacity—time, capital, or a team sitting idle—but every direction looks either saturated, vague, or like a distraction. You aren't lacking ideas; you are drowning in unvalidated ones. You have a Notion page with 47 "potential markets," a spreadsheet of TAM calculations you don't trust, and a growing anxiety that you’re building something nobody will pay for.
The stuckness manifests in three specific loops:
- The Research Trap: You spend weeks reading Gartner reports, Reddit threads, and competitor pricing pages. You feel productive, but you cannot articulate why a specific buyer would switch today.
- The "Cool Tech" Trap: You prototype a feature using the new LLM API or a no-code stack. It works technically. You have zero conversations lined up to show it to.
- The Signal-to-Noise Paralysis: You hear "talk to customers," so you book five calls. They are friendly, vague, and contradictory. You leave with zero decisions made and a calendar full of follow-ups that go nowhere.
You don't need more ideas. You need a mechanism to kill bad ones fast and isolate the one constraint a specific group of people will pay to remove right now.
Why the Usual Fixes Fail
1. "Just talk to 50 customers."
This advice assumes you know who to talk to and what to ask. If you target the wrong segment, 50 conversations yield 50 different feature requests and no pattern. If you ask "Would you use this?", people lie to be polite. You get false positives, not signal.
2. "Scratch your own itch."
Your itch is rarely a market. You are an outlier in technical tolerance, budget authority, and workflow context. Building for yourself usually produces a tool you like but cannot sell because the buyer’s constraints (compliance, procurement, legacy integrations) are invisible to you.
3. "Analyze the TAM/SAM/SOM."
Top-down market sizing is a fundraising exercise, not a discovery tool. A $10B TAM tells you nothing about the $50k wedge you can actually win this quarter. It encourages targeting "everyone in marketing" instead of "Series A B2B SaaS founders struggling with outbound reply rates."
4. "Build an MVP and iterate."
Shipping code before you have a commitment (pre-sale, LOI, paid pilot, or at minimum a scheduled demo with a decision-maker) burns the only scarce resource: your runway. Code creates sunk cost bias. You start defending the feature instead of hunting the problem.
Step-by-Step Solution: The Wedge Discovery Protocol
This process takes 2–3 weeks if you treat it as your full-time job. It produces a Validated Wedge: a specific persona, a specific painful workflow, a specific budget owner, and a verbal commitment to pay for a solution.
Phase 1: Define the Hunting Grounds (Days 1–2)
Action: List your Unfair Advantages.
Open a blank document. List every asset you have that is not evenly distributed:
- Network: Who exactly do you have direct access to (1st degree) who controls budget? List names. Not "CTOs." Names.
- Domain Knowledge: What workflow have you executed 100+ times professionally? (Not "marketing." "Setting up HubSpot attribution reporting for multi-touch funnels.")
- Technical Assets: Code libraries, data sets, API access, regulatory knowledge, hardware.
- Channel Access: Newsletter audiences, Slack communities you mod, podcast guest slots, cold email infrastructure that actually delivers.
Action: Generate 10 "Persona x Problem" Hypotheses.
Cross your advantages against plausible pains. Format: [Specific Role] + [Specific Workflow] + [Specific Pain Metric].
Bad: "Marketers hate reporting."
Good: "RevOps Managers at Series B SaaS (50–200 emp) spend 15 hrs/week manually stitching CRM + CSV data for board decks because Snowflake is too heavy and BI tools break on dirty data."
Stop at 10. Force specificity. If you can't name the workflow, the metric, or the trigger event, delete the row.
Phase 2: The "Problem Interview" Sprint (Days 3–10)
Action: Recruit 3 interviews per hypothesis (30 total max).
Do not post on LinkedIn asking for chats. Use your Unfair Advantages.
- Network: "Hey [Name], I’m researching how [Specific Role] handles [Specific Workflow]. You’re the best person I know for this. 15 mins, no pitch."
- Cold Outreach (if network dry): Target the exact title/company size. Subject: "Question about [Specific Workflow]." Body: 3 sentences. Context -> Specific Friction Question -> Ask for 10 mins. Send 50 to get 3.
Action: Run the Script. Do not deviate.
- Context: "Walk me through the last time you did [Workflow]."
- Tool Stack: "What tools touch this process? Where does the data live?"
- The "Magic Wand" (Constrained): "If you could wave a wand and fix one step without changing the rest of the stack, which step?"
- Current Cost: "How do you handle that step today? Internal script? VA? Manual? What does that cost in hours/$?"
- Trigger: "When did this become a 'must fix now' vs 'annoying'? What changed?" (Funding round? Audit? New hire? Tool price hike?)
- The Commitment Test: "If a solution existed that solved only that step and integrated with [Current Tool], what’s the process to buy a $500/mo pilot?" → Listen for: "I need boss approval," "Security review takes 8 weeks," "I can swipe a card today."
Action: Score immediately after the call.
Use a 1–5 scale on three dimensions only:
- Pain Urgency: Is there a recent trigger event forcing action? (1 = "Nice to have," 5 = "Getting fired if board deck is late again")
- Budget Access: Can this person spend money or champion a vendor this quarter? (1 = "Influence only," 5 = "Card holder")
- Workflow Specificity: Can you draw the exact screen/click/API call where the break happens? (1 = "Vibe," 5 = "I can write the spec")
Kill Rule: If a hypothesis averages < 10/15 across 3 calls, archive it. Do not "pivot the question." Kill the hypothesis. Move to the next row.
Phase 3: The Solution Simulation (Days 11–14)
You have 1–2 surviving hypotheses. Now you test the solution without code.
Action: Build the "Concierge MVP" Proposal.
Create a 1-page PDF (Notion export is fine) for the surviving hypothesis.
- Header: "Automating [Specific Step] for [Specific Role]."
- The Workflow Diagram: 4 boxes. Current State (Manual) -> Your Service (Human/API hybrid) -> Output (Clean CSV/Deck/Alert) -> Destination (Their Tool).
- The Promise: "We deliver [Output] every [Frequency] with [SLA: e.g., <2% error rate, delivered by 8 AM]."
- The Price: "$X/mo for [Volume]. 30-day pilot. Cancel anytime."
- The Ask: "Reply 'Yes' and I’ll onboard you Monday."
Action: Send to the 3–5 interviewees who scored highest on Budget Access.
Send personally. "Based on our chat, I built this. Does this match what you described?"
Action: Measure Commitment, Not Compliments.
- Pass: They sign/LOI/pay/introduce boss. → Build.
- Soft Pass: "Looks great, let me check with [Boss]." → Schedule the 3-way call immediately. If they ghost, it’s a fail.
- Fail: "Cool, keep me posted." → Archive hypothesis. Return to Phase 1.
Phase 4: The Wedge Definition (Day 15)
If you have a Pass: Write the Wedge Spec. This is your only strategy doc.
- Persona: Title, Company Stage, Tech Stack Constraint (e.g., "Must work on top of HubSpot + Snowflake").
- Trigger Event: The specific moment they become a buyer (e
Top comments (0)