DEV Community

Alan Descure
Alan Descure

Posted on Originally published at smtm-store.pages.dev

Opportunity Discovery: A Practical Guide for When You're Stuck

Opportunity Discovery: A Practical Guide for When You're Stuck

The Problem in Concrete Terms

You have time, skills, and maybe even capital—but you cannot point to a specific problem worth solving. Every idea feels either too small (a vitamin, not a painkiller), too crowded (ten funded competitors), or too vague ("I'll build a better CRM"). You sit down to research, open twenty tabs, close them all, and end the day with nothing but a vague sense of anxiety.

This isn't "writer's block for founders." It's a process gap. You're trying to have an insight instead of running a process that produces insights. The difference matters: insights are unreliable; processes are repeatable.

You know this intellectually. But when you try to "just talk to customers," you don't know who to call, what to ask, or how to distinguish polite nods from genuine desperation. When you try "scratch your own itch," your itches are either already solved or too niche to build a business on. When you scan Product Hunt or Indie Hackers, you see finished products—not the messy middle where opportunities actually live.

The result: weeks or months of paralysis, or worse, building something nobody buys.


Why the Usual Fixes Fail

Customer interviews without a hypothesis. "Talk to users" is useless advice without a specific question you're trying to answer. You end up collecting opinions, not evidence. Opinions are free; evidence costs time and changes your mind.

Market research reports. Analyst reports describe past markets. Opportunities live in the gap between what people do today and what they'll do tomorrow. That gap doesn't show up in TAM calculations.

Trend-chasing. AI, Web3, no-code, creator economy—by the time a trend has a name, the asymmetric opportunities are gone. You're competing on execution against people who started six months earlier.

Brainstorming sessions. Divergent thinking without convergent filters produces lists, not decisions. You need a way to kill bad ideas fast, not generate more of them.

Waiting for the "aha" moment. Insight feels like lightning, but it's actually pattern recognition across dozens of data points. You don't get the pattern without the data points.


Step-by-Step Solution

Phase 1: Define Your Search Space (Days 1–2)

You cannot search everywhere. Constrain the problem until it's tractable.

1.1 List your unfair advantages. Write down every asset you control: technical depth (specific languages, domains), distribution access (newsletters, communities, past customers), domain expertise (industries you've worked in), capital, time, location, network. Be specific. "Python" is not an advantage. "Building low-latency trading systems in Python for three years" is.

1.2 List your constraints. Capital, runway, team size, regulatory barriers, geographic limits, energy levels. Honesty here prevents wasted cycles.

1.3 Define your ideal customer profile (ICP) hypotheses. Not "small businesses." "Dental practices with 3–8 chairs doing $1.5M+ revenue who outsource billing." Write three distinct ICP hypotheses. Each must have: firmographics, a specific workflow you understand, and a way to reach 50+ of them within two weeks.

1.4 Pick one ICP to start. Use this filter: which group can you reach this week for conversations? Start there. You can pivot ICPs later; you cannot pivot from zero conversations.


Phase 2: Map the Workflow, Not the Pain (Days 3–10)

People buy workflow improvements, not abstract pain relief. Your job is to document the current workflow in excruciating detail.

2.1 Recruit for workflow mapping, not "feedback." Message: "I'm researching how [ICP] handles [specific workflow] today. 20 minutes, no sales pitch, I'll share the annotated workflow map when done." Offer a $50 gift card or a donation to their charity. Cold email, LinkedIn, warm intros—whatever gets you on calls.

2.2 Run the workflow mapping interview. Script:

  • "Walk me through the last time you did [workflow] start to finish."
  • "What tools, people, spreadsheets, or workarounds are involved at each step?"
  • "Where does it break, stall, or require manual cleanup?"
  • "What happens if this step goes wrong? Quantify the cost: time, money, risk, reputation."
  • "What have you tried to fix it? Why didn't it stick?"
  • "If magic solved this tomorrow, what would change for you?"

Record. Transcribe (Otter, Whisper, manual). Do not take notes during—stay present.

2.3 Build the annotated workflow map. For each interview, create a flowchart: steps → tools/people → break points → quantified cost → current workarounds → desired state. Use Miro, FigJam, or paper. One map per interview.

2.4 Synthesize across 8–12 interviews. Look for: steps where everyone complains, high quantified costs with no adequate tool, workarounds that look like product features (custom scripts, VA-heavy processes, spreadsheet-to-spreadsheet copy-paste). Ignore "nice to have." Keep "I lose $X/week if this breaks."


Phase 3: Score and Filter (Days 11–12)

You now have data. Apply a ruthless filter.

3.1 Score each opportunity candidate on five dimensions (1–5 each):

  • Urgency: Does the prospect actively seek a solution today? (1 = "someday," 5 = "I emailed three vendors this week")
  • Frequency: How often does the workflow run? (1 = quarterly, 5 = daily)
  • Quantified pain: Can you put a dollar number on the cost? (1 = "it's annoying," 5 = "$40k/year in lost revenue + compliance risk")
  • Budget authority: Does your contact control spend or influence the decision-maker directly? (1 = "I need VP approval," 5 = "I have a corporate card")
  • Build fit: Does this align with your unfair advantages from 1.1? (1 = "I'd need to learn everything," 5 = "I've built 80% of this before")

3.2 Calculate weighted score. Weights: Urgency 30%, Quantified Pain 25%, Frequency 20%, Budget Authority 15%, Build Fit 10%. Multiply and sum. Max score = 5.0.

3.3 Keep only candidates scoring ≥ 3.5. If zero candidates clear the bar, return to Phase 2 with a new ICP hypothesis. Do not lower the bar.

3.4 For each surviving candidate, write a one-pager: ICP, workflow step, quantified pain, current workaround, why now, why you, rough solution sketch, go-to-market hypothesis. This is your decision artifact.


Phase 4: Validate Willingness to Pay (Days 13–21)

Ideas are free. Commitment is signal.

4.1 Build a "concierge MVP" proposal. Not a product. A service delivery of the solution: "I will manually do [workflow step] for you for 30 days using [your method/tools]. Price: $X. If it doesn't deliver [specific outcome], you pay nothing." X = 10–20% of their quantified annual pain.

4.2 Pitch to 5–10 prospects from your interviews. Use the exact language from their workflow maps. "You mentioned losing $3k/month on manual reconciliation. I'll run it for you for $500/month. First month free if it doesn't cut that by half."

4.3 Track responses in a simple CRM (Notion, Airtable, spreadsheet):

  • Contact | Pitch sent | Response | Objection | Next step | Commitment level (0–3)
  • 0 = ghosted, 1 = "interesting, keep me posted," 2 = "send me details/proposal," 3 = "yes, let's start"

4.4 Decision rule: If you get ≥ 2 Level-3 commitments, you have an opportunity. Build the product. If you get 0–1 Level-3 after 10 pitches, return to Phase 3, pick the next candidate, or return to Phase 2 with a new ICP.


Phase 5: Start Building (Day 22+)

You now have: a specific workflow step, quantified pain, budget authority, and 2+ paying pilot customers. Stop researching. Start building the smallest thing that delivers the outcome for those pilots.

Set a 4-week deadline for v1. Ship to pilots. Charge. Iterate.


Checklist

  • [ ] Listed 5+ unfair advantages with specificity
  • [ ] Listed 5+ honest constraints
  • [ ] Wrote 3 distinct ICP hypotheses with reach plans

Start here

Top comments (0)