DEV Community

Alan Descure
Alan Descure

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

Opportunity Discovery: A Field Guide for When You’re Stuck Guessing

Opportunity Discovery: A Field Guide for When You’re Stuck Guessing

The Problem in Concrete Terms

You have capacity—time, capital, or a team—but you don’t know where to point it. Every direction looks equally viable or equally risky. You’re not lacking ideas; you’re lacking a filter. The backlog is full of "maybe later" features, the market research folder has three contradictory reports, and the last three experiments you ran produced flat lines.

Stuck opportunity discovery looks like this:

  • You spend Monday researching a niche, Tuesday building a prototype, Wednesday realizing the market is saturated, and Thursday starting over.
  • You cannot articulate why you rejected the last five ideas other than "it didn't feel right."
  • You are waiting for a "signal" (a viral tweet, a competitor pivot, a customer email) to tell you what to build next.
  • Your evaluation criteria change weekly: sometimes it’s TAM, sometimes it’s speed to build, sometimes it’s founder fit.

The core issue isn't a shortage of opportunities. It’s that you are trying to evaluate infinite options with finite criteria, usually inside your own head.


Why the Usual Fixes Fail

1. "Talk to more customers"
This advice assumes you know who to talk to and what to ask. If you are in the discovery phase, you don’t have a customer segment yet. Cold outreach to random personas yields polite noise, not signal. You end up building what the loudest person asked for, not what the market pays for.

2. "Scratch your own itch"
This works for side projects. It fails for businesses because your problems are rarely representative of a paying market. You optimize for your specific workflow, ignoring the constraints (budget, compliance, legacy stack) that define the actual buyer.

3. "Analyze TAM/SAM/SOM"
Spreadsheet modeling is a projection exercise, not a discovery exercise. You can make a $1B market look like $10M or $100M by changing two assumptions. It creates false precision. You end up chasing a market size that exists only in your model.

4. "Build an MVP and iterate"
Building is expensive. If you build to discover, you are paying engineering rates for market research. You also fall in love with the code, making it harder to kill a bad direction when the data says stop.

5. "Follow the trend" (AI, Web3, No-Code, etc.)
Trends are crowded. By the time a trend is obvious in Twitter feeds or TechCrunch, the alpha is gone. You compete on hype cycles, not value creation.


Step-by-Step Solution: The Constraint-First Loop

This process replaces "idea generation" with "constraint discovery." You stop looking for the right idea and start mapping the boundaries of what is possible for you right now.

Phase 1: Define Your Non-Negotiables (The Filter)

Before looking outward, lock down your internal constraints. Write these down. Do not skip this.

  1. Capital Constraint: What is the absolute maximum cash you can deploy before revenue must arrive? (e.g., $5k, $50k, $0).
  2. Time Constraint: How many focused hours/week can you genuinely commit for the next 90 days? (Not "hope to," will).
  3. Skill/Asset Constraint: What specific unfair advantages do you possess today? (Audience, proprietary data, specific technical depth, regulatory access, sales network, domain expertise). List only assets you have now, not ones you can learn.
  4. Risk Tolerance: Are you optimizing for learning, cash flow, or equity value? Pick one primary goal for this cycle.
  5. Dealbreakers: What will you not do? (e.g., No cold calling, no hardware, no regulated industries, no 24/7 support, no venture backing).

Output: A one-page "Constraint Card." Every opportunity must pass this card. If it fails one, it’s dead. No debate.

Phase 2: Map Adjacent Problems (The Search)

Don't brainstorm solutions. Hunt for expensive, urgent, frequent problems inside your constraint boundaries.

Source A: Your Purchase History (Last 12 Months)
Export your credit card/business expenses. Filter for:

  • Tools you pay for but complain about.
  • Services you hired humans for (consultants, agencies, freelancers).
  • Spreadsheets you built to replace software.
  • Action: For each, write: "What is the core job this solves? Why does the current solution suck? What would I pay to fix the suck?"

Source B: Your "Expert" Network (Not Customers)
Message 10–15 peers who operate inside your constraint boundaries (same industry, same role, same stack). Do not pitch. Ask:

"What is the one operational task you did manually this week that made you want to quit?"
"What report does your boss ask for that takes you 4+ hours to build?"
"What vendor did you fire recently and why?"

Record verbatim. Look for patterns in language, not just topics.

Source C: Negative Reviews of Incumbents
Go to G2, Capterra, TrustRadius, Reddit, niche Slacks/Discords. Filter 1- and 2-star reviews for the top 5 tools in your adjacent space.
Search terms: "export," "integration," "support," "pricing," "slow," "bug," "workaround."
Action: Copy the exact phrasing of the top 20 complaints into a doc. Group by workflow step (onboarding, reporting, collaboration, billing).

Source D: "Glue" Work Observation
If you have access to a target workplace (even your own), watch for:

  • Copy-pasting between Tool A and Tool B.
  • CSV exports → Excel cleanup → Import to Tool C.
  • Notion/Google Docs acting as the "source of truth" because the software fails.
  • Recurring calendar invites for "manual reconciliation."

Phase 3: Score & Kill (The Evaluation)

Take your long list of problem statements (aim for 20–30). Score each 1–5 on only these four dimensions:

  1. Pain Urgency: Is the user actively losing money/reputation/time right now? (5 = "I lose my job if this breaks," 1 = "Nice to have").
  2. Frequency: Daily/Weekly/Monthly/Quarterly? (5 = Daily, 1 = Quarterly).
  3. Constraint Fit: Does solving this fit your Capital, Time, Skill, and Dealbreaker cards? (5 = Uses my unfair advantage, costs $0 to validate, 1 = Requires hiring, regulated, 6-month sales cycle).
  4. Willingness to Pay (Proxy): Is money already changing hands for a bad version of this? (5 = Competitors charging $500+/mo, 1 = Free open-source alternatives exist).

Calculation: Total = (Urgency * 2) + Frequency + (Constraint Fit * 2) + (WTP * 1.5)
Weighting favors urgency and your ability to execute.

The Cut: Kill everything below the 70th percentile. You should have 3–5 survivors. If you have more, your constraints are too loose. Tighten them and re-score.

Phase 4: Micro-Validation (The Proof)

Do not build. Do not design. Do not write code. For each of the 3–5 survivors, run one of these tests this week. Pick the cheapest valid test per idea.

Test A: The "Manual Service" Sale
Define the exact output the user wants. Offer to deliver it manually (concierge/Wizard of Oz) for a fixed price.

  • Script: "I’m solving [Problem]. I can deliver [Result] by [Date] for $[Price]. Zero software, just the outcome. Interested?"
  • Pass: They pay (Stripe link, invoice, Venmo).
  • Fail: "Sounds cool, let me know when it's ready."

Test B: The "Clickable Prototype" Pre-sell
Build a 3–5 screen flow in Figma/Framer/Replit (no backend). Show the happy path only.

  • Action: Send Loom video + Calendly link to 10 target users (from Phase 2 sources).
  • Call: "Walk me through this. Would you pay $[Price] for this today? I’ll give you 50% off for life if you commit now."
  • Pass: They book the call and attempt to pay.
  • Fail: "Nice UI," "I'd use the free tier."

Test C: The "Integration/Plugin" Probe
If the problem lives inside a platform (Shopify, Salesforce, Slack, Fig

Start here

Top comments (0)