Market Map Builder

Build a structured, sourced market map of a sector or category: a defensible segmentation, the named players in each segment with public data on scale and momentum (every figure dated), who's entered and exited recently, and where the whitespace might be. The output is a memo with a table — the kind of artifact a strategy analyst, product marketer, business development lead, or investor produces before a positioning discussion, a build-vs-buy decision, a partnership sweep, or a first call in a new space.

Why this skill exists

Market maps are how teams turn "AI for logistics is interesting" or "we should look at the home-health software space" into a real thesis, but they're brutal to build honestly. Public data is scattered across press releases and company sites, figures conflict across sources, dead companies are still listed as active, and there's a constant temptation to fill an empty segment with a half-remembered name. A map with one fabricated company or a three-year-old funding round presented as current is worse than no map. This skill front-loads the discipline: real companies only, every figure dated and sourced, conflicts surfaced instead of averaged away.

Inputs

Required:

• The sector, category, or theme to map (e.g., "vertical AI for insurance claims", "industrial robotics for warehouses", "patient-intake software for dental practices", "managed print services in the Midwest")

Optional but high-value:

• The user's lens: purpose (competitive positioning, partnership sourcing, build-vs-buy, sales territory planning, investment), geography, company size or stage band — shapes which players matter
• A working segmentation hypothesis to test rather than build from scratch
• Companies already known — the user's own company, existing partners, customers, portfolio or business units in the space (to position, and to avoid awkward omissions)
• Time horizon for "recent" (default: last 24 months)

Missing-input behavior: If the sector is too broad to map usefully ("fintech", "healthcare IT", "marketing agencies"), ask once for a narrower cut or a lens; if the user declines, pick the 2–3 most active sub-segments, map those, and label the choice. For all other missing inputs, proceed with defaults stated at the top of the memo.

How it works

1. Define the market boundary first. One sentence on what's in and what's out, with the edge cases named ("includes claims-automation vendors; excludes core policy-administration systems"). A map without a boundary is a list.
2. Build the segmentation from how the market actually divides — by customer, by workflow stage, by business model, or by technology approach. Pick the single axis that best predicts who competes with whom, and say why. 3–6 segments; more means the axis is wrong.
3. Populate segments via web search. For each segment, find real, verifiable companies. For each: one-line description, HQ/geography if relevant, the best available public scale signal (last known funding round with stage and amount, disclosed revenue or headcount, notable customers, ownership — whichever the sources actually support), and the date of the source for that data. Prioritize primary-ish sources: company announcements, credible trade press, regulatory or corporate filings.
4. Handle data quality explicitly:
   • Scale or funding data older than ~18 months: tag [stale — as of {date}]
   • Conflicting figures across sources: report both with sources — never average or pick silently
   • Unverifiable companies (no findable site, no press, no filings): exclude; do not pad segments
   • Possibly dead companies (site down, no activity in 2+ years): tag [activity unclear]
5. Track entrants and exits within the recency window: notable new companies, product launches by incumbents moving into the space, acquisitions (acquirer + date), shutdowns, and pivots into or out of the space.
6. Identify whitespace — and earn it. A whitespace claim needs reasoning: an underserved customer segment, a workflow step nobody covers, a geography gap, a price point no one serves. Distinguish "I found no company doing X" (absence of evidence, clearly labeled) from "X is structurally unserved because…" (an argument).
7. Position the user's known companies (if provided) on the map neutrally — same columns, same flags, no favorable treatment.

Output format

# Market Map — [Sector]
Prepared [date] | Lens: [purpose/geo/size band or "none specified"] | Recency window: [X months]

## Market Boundary
What's in, what's out, edge cases.

## Segmentation Logic
The axis chosen and why it beats the alternatives considered.

## The Map
| Segment | Company | What They Do (one line) | Scale Signal | Last Known Funding / Ownership | As-of / Source | Flags |
|---|---|---|---|---|---|---|
(Flags: [stale], [conflicting data], [activity unclear])

## Conflicting / Uncertain Data
Each conflict spelled out: figure A (source, date) vs. figure B (source, date).

## Recent Entrants & Exits ([window])
Entrants: company, segment, what signal it sends.
Exits: company, type (acquisition/shutdown/pivot), acquirer, date, read-through.

## Whitespace Observations
Each: the gap, the reasoning, and whether it's evidenced absence or structural argument.
Honest counter-case: why the whitespace might be empty for a reason.

## Sources
Numbered list: publication/site, title or description, date accessed/published.

## Limitations
What this map can't see: stealth companies, unannounced deals, private revenue data, regional players with no web footprint.

Choosing the segmentation axis

The axis decision makes or breaks the map. Candidate axes, with when each wins:

• By customer segment (enterprise vs. SMB; hospital system vs. independent practice; carrier vs. broker) — wins when go-to-market, not product, determines who competes
• By workflow stage (intake → processing → resolution; design → fabrication → installation) — wins in process-heavy verticals where companies own a step, not the whole flow
• By business model (software vs. tech-enabled service vs. marketplace vs. franchise) — wins when economics, margins, and exit comparables differ sharply across models
• By technology approach (rules-based vs. ML-native vs. agentic; on-premise vs. cloud) — wins early in a platform shift; ages fastest, so date it
• Avoid axis salad. If the map needs two axes, use one for segments and mention the second as a column attribute — a 2x3 grid of micro-segments with one company each means the boundary was too narrow or the axis is wrong.

Test the chosen axis: do companies in the same segment actually lose deals to each other? If not, the segmentation is taxonomy, not competition.

Guardrails

• Never invent a company. Every named company must come from a verifiable source found during research. An empty segment cell is a finding, not a failure.
• Every figure carries an as-of date and source. No date, no figure.
• Never reconcile conflicting data silently — surface both numbers.
• Funding ≠ quality, and size ≠ winning. Don't rank companies by capital raised or headcount or imply the best-funded player wins; note it as one signal among several.
• Whitespace humility: "no company found" is a search result, not a market fact. Always include the counter-case (the gap may exist because the economics don't work).
• Private company data is inherently incomplete and lagged — the Limitations section is mandatory, not decorative.
• No recommendations on what to buy, build, partner with, or invest in; the map informs a decision, it isn't one.

Quality checks

• Boundary sentence present and segments mutually exclusive on the stated axis?
• Every company verifiable, every figure dated and sourced?
• Stale/conflicting/unclear tags actually used (a map with zero flags is suspicious)?
• Entrants/exits within the stated window only?
• Each whitespace claim has reasoning plus a counter-case?
• The user's own company or known players positioned with the same rigor as everyone else?
• Sources list complete enough that a colleague could re-verify any cell?

Example prompts

"Map the AI-for-construction space for me — preconstruction through field operations, North America. We sell scheduling software into mid-size general contractors and I want to know who's emerged since 2024 and whether there's anything left in field-safety that we should build or partner for. Flag anything where the data looks old."

"We're a regional hospital system evaluating patient-intake and digital front-door vendors. Build the map, test our three-segment hypothesis (scheduling / intake forms / payments), and tell me if the segmentation holds before we write the RFP."

"Quick landscape: who's building agentic AI for insurance claims? Early-stage through growth, global. I need it before a first call tomorrow — flag everything you couldn't verify."

"Our agency is considering an acquisition in the influencer-marketing-platform space. Map the players by business model — SaaS, managed service, marketplace — with any disclosed revenue or ownership, and tell me who's been acquired in the last two years and by whom."

Works well with

Drill into a single name with private-company-tear-sheet; if a mapped player is public, switch to company-first-pass. For learning the sector itself first, use teach-me.