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 Looking for Problems Worth Solving

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 or blank document, trying to find "a good opportunity." Maybe you want to start a business, build a side project, or propose a new initiative at work. The advice you've heard is vague: "scratch your own itch," "talk to customers," "look for problems people complain about." But when you actually try to do it, you hit specific walls:

  • You can't distinguish between a real problem and a minor annoyance. Someone complains about their CRM. Is that a business opportunity or just venting? You don't know how to tell.
  • Your own problems feel too niche or too obvious. You hate how your team handles standups. Is that a SaaS idea or just a process tweak? You have no framework to evaluate.
  • Customer conversations go nowhere. You book calls, ask "what's your biggest pain point?", get generic answers like "time management" or "communication," and leave with nothing actionable.
  • Market research feels like guessing. You look at G2 reviews, Reddit threads, Twitter complaints. Everything looks like noise. You can't separate signal from volume.
  • You keep finding "solutions looking for problems." You get excited about a technology or approach, then realize you're forcing it onto a market that doesn't care.

The result: weeks or months of spinning, a graveyard of half-researched ideas, and the creeping sense that maybe you just "don't have the entrepreneur gene."


Why the Usual Fixes Fail

1. "Scratch your own itch" assumes your itch is representative.

Most builders have atypical workflows, technical tolerance, and budget authority. What bothers you often doesn't bother a paying market—or the people who do have the problem can't pay to solve it.

2. "Talk to customers" without a hypothesis is waste.

Open-ended discovery calls produce anecdotes, not evidence. Without a specific, falsifiable assumption about who has what problem under what constraints, you're just collecting stories.

3. "Look at competitors" leads to feature parity, not opportunity.

Analyzing existing products shows you where the market was, not where it's going. The gaps in competitor offerings are often gaps for a reason: too small, too hard, or too low willingness-to-pay.

4. "Validate with a landing page" tests demand for a specific solution, not the problem.

A waitlist for your imagined product tells you nothing about whether the underlying problem is urgent, frequent, and valuable enough to sustain a business.

5. Frameworks like Jobs-to-be-Done or Design Thinking are taught as workshops, not daily practice.

You learn the canvas, fill it out once, then revert to intuition when the real work starts. The framework becomes theater.


Step-by-Step Solution: A Repeatable Discovery Process

This process takes 2–4 weeks part-time. It requires no budget, no team, and no special tools—just a spreadsheet, a calendar, and discipline.

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

Don't start with ideas. Start with constraints.

Create a one-page "search brief" answering:

Question Your Answer
What type of buyer can I realistically reach? (e.g., "B2B marketing managers at 50–500 person companies," not "small businesses")
What budget authority do they have? (credit card, department budget, requires VP approval)
What domain do I understand well enough to evaluate problems? (e.g., "devops tooling," "clinical trial ops," "creator monetization")
What outcome am I optimizing for? (venture-scale, lifestyle business, internal tool adoption, portfolio piece)
What time horizon can I commit to before needing revenue/traction?

Why this works: It kills 80% of bad ideas before you waste time on them. If you can't reach the buyer, don't understand the domain, or need revenue in 3 months but the sales cycle is 12, the opportunity is dead on arrival.

Output: A written brief you can reference every time a shiny idea appears.


Phase 2: Map the Workflow, Not the Pain (Days 3–7)

Pick one specific role in your target buyer segment. Not "marketers"—"content marketing managers at B2B SaaS companies who publish 4+ blog posts/month."

Find 3–5 people in that exact role. Use LinkedIn, Twitter, your network, cold DMs. Offer a $50 gift card for 30 minutes. Say: "I'm researching how [role] handles [specific workflow]. Not selling anything. Just want to understand your process."

Run a workflow mapping interview. Do not ask about problems. Ask them to walk you through a specific instance of the core workflow:

"Walk me through the last time you [published a blog post / ran a sprint retrospective / onboarded a vendor / closed monthly books]. Start from when you knew it needed to happen. End when it was done. Tell me every tool, handoff, decision, and delay."

Take timestamped notes. Capture:

  • Steps (what they did, in order)
  • Tools (what they used at each step)
  • Handoffs (where work moved between people/systems)
  • Delays (where they waited, re-did work, or stalled)
  • Workarounds (spreadsheets, Slack messages, manual exports, "I just...")

Do this 3–5 times. Pattern-match across interviews. You're looking for:

  • Steps that appear in every interview but use different tools
  • Handoffs that consistently cause delays or errors
  • Workarounds that exist because no tool handles the step well
  • Steps the interviewee complains about unprompted (mark these)

Output: A workflow map (spreadsheet or Miro board) with frequency counts for each step/tool/pain signal.


Phase 3: Score Candidate Problems (Days 8–10)

From your workflow map, extract every distinct problem signal. Each becomes a row in a scoring sheet. Score each 1–5 on:

Criterion 1 = Low 5 = High
Frequency Happens monthly or less Happens daily/weekly
Urgency "Nice to fix" "Blocker / fire drill"
Cost of status quo Minor annoyance Measurable $ loss, churn risk, compliance risk
Willingness to pay No budget, no authority Budget exists, buyer has authority
Reachability Can't find/access buyers You can book 5 calls this week
Differentiation Crowded, mature solutions Few/no good options, or solutions miss the workflow context
Founder fit You'd hate building this You have unfair advantage (domain, network, tech)

Keep only problems scoring ≥25/35. You'll typically have 2–4 survivors.

Why this works: It forces explicit tradeoffs. A problem that's urgent but unreachable dies. A problem you're perfectly suited for but nobody pays for dies. You stop arguing with yourself and start comparing apples to apples.


Phase 4: Run a "Concierge Test" on the Top 1–2 Problems (Days 11–21)

Do not build. Do not design. Do not write code.

For each surviving problem, find 3 people who have it right now (not "would have it"). Offer to solve it manually for them for 2 weeks—free or paid, depending on complexity.

Define the engagement:

  • What you'll do (specific deliverable, not "help")
  • What they must provide (access, data, feedback time)
  • Success criteria (measurable: "reduce time from 4hr to 1hr," "eliminate 3 manual handoffs")
  • Duration (2 weeks max)

Execute like a service provider. Use their tools. Follow their constraints. Document every edge case, exception, and "oh by the way."

Track:

  • Time you spend per unit of value delivered
  • Where you struggle (this reveals hidden complexity)
  • What they actually care about vs. what they said
  • Whether they ask to continue/pay at the end

Output: A one-pager per problem: "Concierge Results — [Problem Name]" with metrics, quotes, and a go/k

Start here

Top comments (0)