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 the capacity to build. You have time, maybe some savings, perhaps a skill stack that includes coding, design, writing, or sales. But every time you sit down to answer "what should I build?" or "what problem should I solve?", you hit a wall.

The wall looks different for everyone:

  • The blank page problem: You open a notebook or document and nothing comes. No ideas feel worth pursuing.
  • The "already exists" trap: Every problem you think of has ten solutions on Product Hunt, three funded competitors, and a Reddit thread explaining why it's a bad market.
  • The passion mismatch: You can list problems intellectually, but none of them make you want to work on them for two years.
  • The validation paralysis: You have three ideas but can't decide which to test first, so you test none.
  • The audience ghost town: You don't have a community, newsletter, or Twitter following. You don't know who you'd sell to even if you knew what.

This isn't "writer's block for founders." It's a systematic gap in how you're gathering and filtering signal. You're trying to generate opportunities from inside your own head instead of extracting them from the world.


Why the Usual Fixes Fail

1. "Scratch your own itch"
Works if your itch is shared by thousands of people with budget and urgency. Fails when your itch is niche, low-budget, or already solved well enough that switching costs exceed the pain.

2. "Talk to customers"
Generic advice. Who? Where? Asking what? Cold DMs get ignored. Surveys get biased responses. "Would you use this?" gets lies. Most people don't know how to run a discovery interview that yields actionable signal.

3. "Look at trending topics on Twitter/Reddit/Product Hunt"
You're seeing solutions, not problems. By the time a solution trends, the opportunity window for new entrants has narrowed. You're also seeing survivor bias—successful launches, not the graveyard of similar attempts.

4. "Pick a growing market and find a wedge"
True but useless without a method for finding the wedge. "B2B SaaS" is a market, not an opportunity. "Dental practice management" is a vertical, not a problem. You need a way to go from broad to specific without guessing.

5. "Just start building, you'll figure it out"
Expensive way to learn. Code is the most expensive validation tool. Building first commits you to a solution before you understand the problem, the buyer, the budget, and the switching trigger.


Step-by-Step Solution

Phase 1: Build Your Signal Sources (Week 1)

Goal: Create a reliable, ongoing stream of raw problem data from people who pay for solutions.

1.1 Identify 3-5 target buyer groups you can actually reach
Don't pick "SMBs." Pick "independent insurance adjusters in Texas," "Shopify merchants doing $100K-$1M/year in home decor," or "engineering managers at Series A B2B SaaS companies." Specificity determines reachability.

Action: List 5 groups where you either (a) have direct contacts, (b) can find active communities, or (c) can cold-email with a relevant hook. Score each 1-5 on access, budget evidence, and your genuine interest. Pick top 3.

1.2 Map where each group actually talks
Not "LinkedIn." Specific Slacks, Discords, subreddits, Facebook groups, forums, newsletters, podcasts, conferences, association directories.

Action: For each group, find 3-5 watering holes. Join. Lurk for 48 hours. Note: what questions repeat? What complaints repeat? What tools get recommended or cursed? What workflows get described with frustration?

1.3 Set up automated signal capture
Action:

  • Create a private Notion database or Airtable base: columns for Source, Date, Raw Quote, Problem Category, Urgency Signals, Budget Signals, Frequency Count
  • Set up Google Alerts for "[role] + struggle/challenge/problem/frustration/hate/wish"
  • Subscribe to 2-3 newsletters per target group
  • Use a tool like GummySearch (Reddit) or PhantomBuster (LinkedIn groups) to export recent discussions monthly

Phase 2: Run Structured Discovery Conversations (Weeks 2-4)

Goal: Convert raw signal into validated problem hypotheses with evidence.

2.1 Recruit 15-20 conversations per target group
Not "interviews." Conversations. No scripts. No "would you pay." Your only goal: understand their world well enough to simulate their decision process.

Recruitment templates that work:

  • Warm intro: "Hey [name], [mutual connection] suggested I talk to you. I'm researching how [role] handles [workflow]—not selling anything, just learning. 20 mins?"
  • Cold but relevant: "Saw your post about [specific pain]. I'm building a map of how [role] solves [problem] today. Your perspective would sharpen it. 15 mins, no pitch."
  • Community ask: "Hi all, I'm talking to 15 [role]s about [workflow]. Not a sales call. Happy to share the anonymized summary. Who's open to 20 mins?"

2.2 Use the "Day in the Life" + "Last Time" protocol
Structure:

  1. "Walk me through a typical Tuesday." (Context, tools, handoffs, interruptions)
  2. "Tell me about the last time you [did the workflow related to your suspected problem]." (Specific instance, not generalization)
  3. "What tools or workarounds did you use?" (Reveals current solution stack)
  4. "What sucked about it?" (Unprompted pain)
  5. "Did you look for alternatives? What happened?" (Reveals buying behavior, switching costs, budget authority)
  6. "If magic wand—what would change?" (Unconstrained desire, but grounded in their reality)

2.3 Capture verbatim, not summaries
Action: Record (with permission) or take raw notes. After each call, spend 10 minutes dumping exact quotes into your database. Tag: Problem Area, Current Workaround, Urgency Language ("I lose hours," "clients complain," "we're losing deals"), Budget Language ("we pay $X for," "budget approved," "boss would kill for").


Phase 3: Synthesize and Score (Week 5)

Goal: Move from anecdotes to a ranked shortlist of testable opportunities.

3.1 Cluster problems by pattern
Action: Review your database. Group quotes into problem clusters. A cluster needs:

  • 3+ distinct people describing the same core friction
  • At least 2 using urgency language
  • At least 1 mentioning budget spent or allocated on current workaround

Name each cluster with a verb-noun phrase: "Reconciling Stripe payouts with bank deposits," "Onboarding clients without manual data entry," "Debugging flaky CI/CD pipelines."

3.2 Score each cluster on the OPPORTUNITY framework

Criterion Question Score 1-5
Owner Is there a clear single decision-maker?
Pain Is the problem described as urgent, costly, or blocking?
Paying Is money already spent on tools, contractors, or workarounds?
Occurrence Does it happen daily/weekly (not monthly/quarterly)?
Reach Can you identify and access 100+ similar buyers in 30 days?
Technical Can you build a meaningful v1 in 6-8 weeks with your skills?
Urgency Is there a trigger event (growth, regulation, tool sunset, hiring)?
Niche Is the problem specific enough to avoid horizontal competition?
Insight Do you have a non-obvious angle from your conversations?
Timing Is the problem growing (more users, more complexity, new constraints)?

Action: Score honestly. Clusters scoring <30/50 get parked. Top 3 move forward.

3.3 Write a one-pager for each top cluster
Template:

  • Problem statement (one sentence, buyer's language)
  • Current workaround (tools + manual steps + time cost)
  • Buyer profile (title, company stage, team size, trigger)

Start here

Top comments (0)