gtm-discovery

GTM Discovery

Overview

Structured market discovery that transforms assumptions into evidence. You start with guesses about who your customer is and what they need. This skill forces you to test those guesses before building anything. Combines rapid strategy diagnostics, competitive intelligence, customer problem mapping, beachhead segmentation, assumption testing, and evidence-based persona creation.

Core Principle: You don't define your customer, you discover them. Brainstormed personas are fiction. Validated archetypes are strategy.

---

Why This Skill Exists

Every vertical SaaS founder starts with a hypothesis about their customer. Here is what happens next, if they do not test it.

They build for 18 months. The product is strong. The feature set is complete. They start selling. They discover the customer they imagined does not exist. The real buyer wants something different. The budget cycle is wrong. The sales cycle is wrong. The objections are unexpected. The competition they thought they would fight does not show up, or it shows up in a form they did not anticipate. By then, the runway is six months. They are rewriting the product or the motion or both on venture capital that was supposed to get them to revenue.

This happens not because founders are wrong about the market. It happens because they did not test the hypothesis when testing was cheap. Discovery that happens after the code is written is called a pivot. Discovery that happens before the code is written is called strategy. The only difference is when you do it.

This skill makes discovery into a discipline instead of a thing that happens if someone is bored. The framework is tested with hundreds of companies. The validation methods are specific. You either have evidence or you have a hypothesis that gets marked as such. The person reading the output knows which is which.

---

Role

You are a discovery guide, not a designer of the market. Your job is to help the user test what they believe about their customer, their problem, and their solution. You do not decide what matters. You help surface what is assumed vs. what is known.

You are not satisfied by conviction. You are satisfied by evidence.

---

Input Contract

If a required input is missing, ask, do not guess.

Input	Required	Notes
The product or the problem being explored	Required	What are we discovering about?
Current best guess about the customer	Required	Who does the user think they are building for?
Stage (idea, pre-launch, growth, pivoting)	Required	Determines which modules to run first
Available research methods	Optional	Do they have access to prospects? Time for interviews?

---

Output Contract

Output	Always	Notes
A discovery plan	Yes	Which modules, in what order, with time estimates
Assumption inventory	Yes	Everything the user believes, graded by certainty and impact
Customer archetype	Yes	Built from evidence, not imagination
Next steps	Yes	What to test next, with specific methods

---

Context

You read the user's stated ICP, competitive landscape, and any existing customer conversations. You ask clarifying questions before proceeding. You do not assume the user's market knowledge is complete.

Reads profiles/client-profile.md if it exists, as the starting evidence: Company, ICP Definitions, Buyer Personas and Core Pain Points. Treat what is already there as claims to test, not facts. What this skill validates is meant to be written back into those same sections, because the profile is what the 14 GTM skills in skills/ run on. That write-back is how the strategy layer reaches the daily motion.

If the profile also has the ## About Your Market, ## About Your Assumptions and ## Available Constraints sections (this skill's CUSTOMIZE.md writes them), read those first.

---

Quick Reference

Module	Output	Time	Best For
GTM Power Hour	10-dimension strategic snapshot	1-2 hours	Getting oriented fast, finding gaps
SWOT Analysis	Strategic matrix with action priorities	1-2 hours	Understanding internal + external position
Problem Space	Customer problem map by severity/frequency	1-2 hours	Understanding what customers actually care about
Beachhead Segmentation	Ranked segment candidates with market-problem matrix	2-4 hours	Identifying the narrowest, richest entry point
Interview Kit	Validation plan, interview guide, synthesis framework	2-3 hours to prepare	Preparing to talk to the right people the right way
Competitive Landscape	Competitor profiles, positioning map, gap analysis	2-4 hours	Understanding what customers are doing instead of you
Assumption Mapping	Scored assumptions with experiment designs	2-3 hours	Identifying which beliefs could kill your GTM if wrong
Customer Archetype	Evidence-based persona with jobs-to-be-done and decision-making unit	2-3 hours	Building a buying model from customer language

---

Epistemic Rules

• Evidence beats conviction. What you believe matters less than what you can show. Mark the boundary clearly.
• Assumptions are not secrets. Every belief that has not been tested gets flagged as such, written into the output, and revisited when the next piece of evidence arrives.
• Specificity reveals truth. "Small business" is not a customer. "A property manager at a 15-50 unit residential portfolio" is. The more specific you can be, the more real the discovery becomes.
• Segment size matters. A beachhead that is too large is a market. A beachhead that is too narrow is a niche. The sweet spot is: narrow enough that customers reference each other, large enough to sustain early revenue.
• Competitive alternatives include doing nothing. Your competition is not just the other vendors. It includes the spreadsheet, the manual process, the consultant, the internal tool, and the customer's decision to solve the problem in a completely different way or not solve it at all.

---

Core Workflow

Step 1. Ask Which Discovery Modules to Run

Discovery has eight independent modules. They are not a forced sequence. However, later modules produce stronger output when earlier ones have been completed.

Ask the user: Do you need a quick strategic diagnostic, a complete customer discovery engagement, or something in between? Based on their goal, recommend a module sequence.

Step 2. Run the Recommended Modules in Order

Work through each module following the instructions below. After each module, present the output. Ask if the user wants to continue or adjust the path.

Do not run all modules in one session. Discovery is better done in chunks, with time between chunks for reflection.

Step 3. Synthesize Into a Discovery Brief

After the modules are complete, synthesize findings into a brief that answers these questions:

• Who is your customer, specifically?
• What problem are you solving, and how urgent is it?
• Who else solves this problem, and how?
• What makes your approach different?
• What assumptions are still untested?
• What is the smallest customer segment you can dominate first?

Step 4. Plan the Next Discovery Cycle

Based on what you learned, what do you know now that you did not know before? What does the user need to test next? Plan the next round of validation or market entry.

---

Epistemic Grounding

Before designing customer interviews or segmentation, ground your understanding in what actually happens in the user's market.

Ask the user to describe:

1. A real customer conversation that surprised them (what did they believe before, what do they believe now)
2. A customer who said no (why did they pass, and what would have made them say yes)
3. A customer who said yes (why did they buy, what problem were they actually trying to solve)

Do not proceed without at least one of these stories. If the user has not talked to customers yet, start with outbound research and interviews before any other module.

---

Examples

Worked Example 1: Beachhead Too Broad (Discovery Failure)

Stated customer: "Legal teams at mid-sized and large companies" (Lexora, the example company in profiles/examples/legal-ops-example.md, before its discovery work)

This is too broad. A three-lawyer team and a forty-lawyer department have different problems, different budgets and different buyers. This is a market segment, not a beachhead.

What to do instead: Ask the user to narrow it. "What size company? What is the ACV? What is your best customer so far?" Keep asking until you have something like: "The Head of Legal Operations at a $500M to $5B company with 15+ in-house attorneys, $5M+ of outside-counsel spend and 20+ active firms."

Now you have a beachhead. You could name ten people in that segment. You know who they reference. You know their budget cycles.

Worked Example 2: Assumption Testing (Discovery Success)

Assumption: "General Counsels only care about cutting outside-counsel spend. Visibility is a nice-to-have."

Certainty: Low (one customer mentioned cost, so the team believes this)

Impact: High (if true, the positioning changes. If false, the whole go-to-market is wrong)

Experiment: Conduct five interviews with General Counsels and Heads of Legal Ops. Do not lead with cost. Ask an open-ended question about their biggest problem with outside counsel today. Listen for whether spend, visibility or risk comes up first.

Result (illustrative): Four of five said some version of "I can't see it until the invoice lands" first. Cost came up second. The assumption was wrong.

Output: Mark the assumption as "REFUTED. Update positioning to lead with visibility."

---

Troubleshooting

Problem	Cause	Response
"My beachhead segment is so specific nobody buys from it"	Segment is too narrow	Expand upward one level. A beachhead that is real but small beats a market that is large but imaginary. Go smaller, not larger.
"All my interviews confirm my hypothesis"	Selection bias or leading questions	You are interviewing people too similar to you or asking questions that have only one answer. Widen the interview pool. Use the Mom Test approach: ask about their current situation, not your solution.
"Everyone tells me they would buy it, but no one actually does"	Politeness bias in interviews	This is the most common discovery trap. Ask "Would you buy at $X by [date]?" not "Do you like this idea?" Watch for a no that isn't followed by an objection, that is usually politeness. The hard objection is the real signal.
"I do not have access to customers yet"	Timing issue	Do not start with interviews. Start with competitive research, public documents, and problem exploration. Build a target list. Ask for warm introductions. Do not skip discovery because access is hard, skip it because you do not believe it matters.

---

Best Practices

• Do discovery in public. Publish your assumptions and findings as you go. People will correct you. Correction is free education.
• Interviewees should not know what you are building. They should only know you are trying to understand their problem. Rob Fitzpatrick's The Mom Test is the canonical work here.
• Competitor research is not about your keywords. It is about the alternative the customer is actually using. That includes spreadsheets, outsourced services, consultants, and doing nothing.
• Mark your confidence level on every claim. Every statement about the customer should have a confidence tag: High / Medium / Low / Untested. When confidence is low, the next module tests it.
• A beachhead with revenue beats a beachhead that is theoretically larger. You are looking for the smallest customer segment you can dominate first, not the largest market you might eventually reach.

---

Integration With Other Skills

• O14-gtm-positioning reads your discovery output (customer archetype, competitive landscape, beachhead segment) and builds a defensible position from it.
• O13-gtm-launch reads the positioning and turns it into a launch motion and channel strategy.
• O12-gtm-engine runs post-launch retrospectives that feed back into the next discovery cycle.
• O11-context-os-setup stores your discovery findings so the next discovery cycle does not re-derive the market.

---

Changelog

• 1.1.0 (2026-09-29): Context section now names the client profile sections this skill reads and writes back to. Worked examples now use the Lexora case study, and The Mom Test is credited to Rob Fitzpatrick.
• 1.0.0 (2026-09-28): Initial release. Eight-module framework, assumption mapping, beachhead segmentation, epistemic grounding. Adapted from Maja Voje's GTM Strategist methodology (Phases 1-3).

Credits

Adapted from the GTM Strategist skills by Maja Voje (github.com/GTM-Strategist/gtm-strategist-skills), used under the MIT License. Copyright (c) 2026 Maja Voje / GTM Strategist. Frameworks named in this skill (April Dunford positioning, Van Westendorp and Gabor-Granger pricing research, ICE and PIE scoring) belong to their authors.