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 no clear vector. You aren't lacking ideas; you are drowning in them. Or worse, you have zero ideas that survive first contact with reality.

The stuck state looks like this:

  • The "Shower Thought" Trap: You generate high-level concepts ("Uber for X," "AI for Y") that evaporate when you try to define the first user.
  • The Research Rabbit Hole: You spend weeks reading market reports, TAM/SAM/SOM breakdowns, and competitor funding announcements, yet you cannot name a single specific person who would pay you tomorrow.
  • The Validation Theater: You build landing pages, run cold email campaigns, or post in communities asking "Would you use this?" People say "Yes, looks cool." Nobody buys.
  • The "Scratch Your Own Itch" Dead End: You solve a problem you have, only to realize you are the only person who has it, or the only one unwilling to pay for the solution.

You are stuck because you are treating opportunity discovery as a brainstorming activity instead of an investigation activity. You are looking for a "good idea" when you should be looking for evidence of unmet demand.


Why the Usual Fixes Fail

1. Keyword Research & Trend Reports

Tools like Ahrefs, Semrush, or Gartner reports tell you what people search for or what analysts think markets will do. They do not tell you why a specific workflow breaks today. High search volume often equals high competition and low urgency. "Project management software" has massive volume; "How to stop subcontractors from uploading wrong revision PDFs to Procore" has near-zero volume but massive urgency.

2. "Talk to Customers" (Generic Edition)

The advice "talk to 20 users" fails when you don't know who the user is yet. If you recruit randomly, you get polite noise. If you recruit friends, you get support, not truth. Without a specific hypothesis about a specific workflow, interviews become therapy sessions for the interviewee, not data for you.

3. Building an MVP to "Test the Market"

Writing code (or no-code workflows) before you have a manual concierge sale is expensive procrastination. You aren't testing the market; you are testing your ability to build. The market doesn't care about your stack; it cares about the delta between current pain and your proposed relief.

4. Competitor Analysis

Copying a competitor’s feature set assumes they know what they are doing. Most features in successful products are vestigial—built for a client who churned two years ago. Chasing parity puts you in a feature war you cannot win. Opportunity lives in the gaps they ignore.


Step-by-Step Solution: The "Workflow Audit" Method

This process requires zero budget, no code, and no audience. It requires access to people who do work. The output is a ranked list of specific, chargeable problems.

Phase 1: Define the Hunting Grounds (Days 1–2)

Do not start with "SaaS ideas." Start with expensive, repetitive workflows.

  1. List 5–10 Industries/Functions you can access.
    • Constraint: You must be able to get a 30-minute Zoom call with a practitioner this week via cold LinkedIn DM, warm intro, or community ask.
    • Examples: "Commercial HVAC estimators," "Dental office managers," "Series A CTOs," "Freight forwarders," "Non-profit grant writers."
  2. Filter for "High Friction + High Cost."
    • Keep only roles where: (A) The person manages a budget or revenue line, (B) The work involves moving data between tools (Excel, email, ERPs, portals), (C) Errors are expensive (rework, compliance fines, lost revenue).
  3. Pick ONE target. Not three. One. You will go deep.

Phase 2: The "Shadow the Work" Interview (Days 3–7)

Do not ask "What problems do you have?" People rationalize problems away. Ask them to show you the work.

The Outreach Script:

"I'm researching how [Target Role] handles [Specific Workflow, e.g., 'change order management' / 'month-end close' / 'candidate scheduling']. Not selling anything. I just want to watch you work for 20 minutes via screen share to understand the friction. Happy to pay $50 for your time / donate to your charity / share the friction map I create."

The Interview Protocol (Record with permission):

  1. Context (2 min): "Walk me through a typical Tuesday."
  2. The Target Workflow (15 min): "Let's zoom in on [Workflow]. Can you share your screen and show me the last time you did this?"
    • Watch for: Alt-tabbing, copy-pasting, VLOOKUPs, PDF scraping, "I have to check with...", "We usually just...", screenshots sent to Slack/Email.
    • Ask: "What happens if this step goes wrong?" "How long does this take?" "Who yells at you if it's late?"
  3. The "Magic Wand" Close (3 min): "If you could wave a wand and remove one step without breaking compliance/quality, which step dies?"

Critical Rule: You are mapping steps, tools, handoffs, and failure points. You are not pitching.

Phase 3: Build the Friction Map (Day 8)

Immediately after each call (do 3–5 calls), transcribe the workflow into a Swimlane Diagram (Miro, Excalidraw, paper).

  • Lanes: Actor (Estimator, PM, Admin, Client).
  • Nodes: Tools (Procore, Excel, Outlook, Phone).
  • Edges: Data movement (Export CSV -> Clean -> Upload -> Email).
  • Red Flags (Mark in Red):
    • Manual Bridge: Human moving data between System A and System B.
    • Tribal Knowledge: "Only Sarah knows which tab to use."
    • Error Recurrence: "We mess up the tax code 20% of the time."
    • Time Sinks: "Takes 4 hours every Friday."
    • Shadow IT: Unapproved scripts, personal Gmail for work files, massive shared spreadsheets.

Phase 4: Quantify the "Cost of Status Quo" (Day 9)

For each Red Flag, calculate the Monthly Cost of Inaction (MCI) with the interviewee's numbers (or your estimates based on their salary).

MCI = (Hours Wasted × Loaded Hourly Rate) + (Error Rate × Cost Per Error) + (Delay Cost × Frequency)

Example: "You spend 6 hrs/week fixing PDF exports. Loaded rate $75/hr. That’s $1,800/mo. Plus 2 errors/mo at $5k rework = $10k. Total MCI: $11,800/mo."

If MCI < $500/mo, discard the problem. It’s a vitamin, not painkiller. If MCI > $2,000/mo, highlight it. This is your opportunity shortlist.

Phase 5: The "Concierge Close" (Days 10–14)

You now have 1–3 high-MCI problems. You do not build. You sell the manual solution.

  1. Define the Service: "I will run the [Workflow] for you manually for 30 days using [Script/VA/Sheet]. You pay $X/mo (≈ 10–20% of MCI). If it doesn't save Y hours, you pay nothing."
  2. Pitch the 3–5 people you interviewed. "Remember that PDF export nightmare? I've built a manual process to handle it. Want in for the pilot?"
  3. Execute manually. Use Zapier, Sheets, Python scripts, Upwork VAs, your own hands. Deliver the result, not the software.
  4. Measure: Did they pay? Did they renew? Did they complain when you were late?

Outcome:

  • Paid + Renewed: You have a validated opportunity. Now you build software to replace the manual labor.
  • Ghosted/Churned: The pain wasn't urgent enough, or your manual process sucked. Return to Phase 2 with a new target. You lost 2 weeks, not 6 months.

Checklist: Are You

Start here

Top comments (0)