Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Orchestrate a complete competitive analysis across six steps, from landscape to strategic direction. Use when you need the full picture, not a single scan or card.
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 96% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 105% | 0% |
Orchestrate rigorous, repeatable competitive analysis across six analytical steps, producing four actionable output types (battle cards, comparison matrices, positioning counter-moves, threat assessments). This is the umbrella: each step names its purpose, its frameworks, and the skill that does the work — the way discovery-process orchestrates the discovery skills. Run all six for a strategy cycle; run one when a specific question walks in the door. The steps build on each other, but the sequence bends to the decision.
Works best with: the market or competitor set in scope, and the decision this analysis supports — a roadmap bet, a positioning refresh, a market entry, deal defense. The decision determines which steps get depth and which get a pass. Also useful: any existing intelligence — a prior landscape scan, snapshots, battle cards, or company-intel output — so steps consume rather than repeat it.
Input supplied inline with the invocation — text after the skill name, a pasted context dump, or an appended ARGUMENTS: line — counts as answers already given. Use it; don't re-ask.
Arriving empty-handed? That works too. The workflow opens by asking what decision is on the table and what you already know, then recommends which steps to run and in what order.
Example invocation: Run the competitive analysis process on the mid-market payroll space — decision: whether we defend down-market or push up. We have last quarter's landscape scan.
matrices, positioning moves, threat assessments) are what sales, execs, and roadmap owners actually consume. A step that doesn't feed an output is a detour.
This file holds the sequence, the frameworks, and the decision points — not the research mechanics.
autonomous-investigation protocol (Fact/Inference/Assumption labels, real URLs, confidence stacking), drawing signals from the channels in intelligence-collection-disciplines.
the Kano model, Jobs-to-be-Done, the Business Model Canvas, Ries & Trout positioning, Hamel & Prahalad strategic intent. The frameworks are defined where used; agents shouldn't assume prior knowledge.
and it's the step most teams skip, which is why they're perpetually surprised.
Step 0 — Scope. Confirm the decision, the competitor set (or run market-landscape-scan to find it), and which steps matter. Offer the six steps as a numbered menu; recommend a subset when the decision doesn't need all six.
Who's in the game, where, and how they operate — market saturation, fragmentation, disruptive entrants.
competitors by GTM model, innovation pattern, or regional depth); market fragmentation indexing.
market-landscape-scan — segments, player map,dynamics, whitespace with the dead-zone test.
Beyond feature lists: value delivery, architecture, interoperability, lifecycle maturity.
Product model (core product vs. the ring of services and integrations deals actually hinge on); JTBD-based capability maps (compare by customer job, not by feature name).
competitive-research-snapshot on theselected players — its comparison matrix is this step's artifact.
checklist hides: two "identical" features serving different jobs badly.
How well they serve real customers — outcomes, support quality, ecosystem benefits — not tech specs.
Proposition Canvas (their promise vs. the jobs/pains/gains that matter).
voice-of-customer-miner across the sources yourbuyer reads — need themes, quoted verbatims, competitor weak points, and switching triggers, each with a source-bias note.
signal; curated success stories are marketing.
The financial, organizational, and market baseline: structure, revenue streams, growth levers, channels.
breakdown; sales channel profiling (direct vs. OEM vs. VAR vs. platform).
company-intel per deep-dive competitor — its eleven-sectionoutput (financials, offers, org signals) is this step's artifact; FININT sources per intelligence-collection-disciplines do the heavy lifting.
companies lie less where lying is a felony.
How buyers perceive you versus them — in promise, emotion, and mental association, not capability.
mapping (price vs. value-perception plots; find the uncontested space); Blue Ocean strategy canvas.
review mining); draft or revise with positioning-statement.
say they don't cash. That gap is a positioning counter-move waiting to be claimed.
Where they're going, not where they are — the step that lets you preempt instead of react.
Innovation Ambition Matrix (core/adjacent/transformational mix of their bets).
intelligence-collection-disciplines — TECHINT (patent clusters, trademark filings, preprints), HUMINT (hiring surges, leadership moves), FININT (earnings language, M&A themes), MASINT (supply chain, facility buildout) — fused via the confidence-stacking rule. competitive-intel-watch keeps this step running on a cadence after the initial pass.
subdomain cert is a committed bet. Stack before you brief.
Close by building what the audience consumes — offer as a numbered menu:
battle-card-builderThen set the maintenance cadence: competitive-intel-watch on the schedule the artifact mapping in intelligence-collection-disciplines prescribes (weekly SIGINT layer for cards, quarterly deep passes for direction).
A copy/paste engagement tracker for all seven steps lives in template.md.
Scoping to the decision (fictional): the decision is deal defense against one rival in enterprise deals. The right subset is Steps 2, 3, and 5 on that one competitor, straight into a battle card — Steps 1, 4, and 6 add days, not win-rate. The orchestrator's value is permission to skip: "Based on your decision, I recommend Steps 2→3→5 only. Run the full six?
Step 6 catching what Steps 1-5 missed: the present-state read shows a competitor stable and mid-pack. The direction pass finds a patent cluster in a new classification, four platform-engineer postings, and an earnings call where the CFO declined to break out R&D — three disciplines, one story: a platform pivot in progress. The threat assessment brief re-ranks them from "monitor" to "act," a quarter before their announcement would have forced a reactive scramble.
See examples/sample.md for a complete worked engagement (fictional FSM-software market) showing the orchestration decisions — scoping menu, recorded skips, and step-to-step compounding — with links into the delegated skills' own worked artifacts. examples/sample-industrial.md shows the umbrella bending — steps merged, one run out of order, cadence slowed — with each call's reasoning stated.
analysis ships after the decision got made without it.
buys you a quarter of lead time. Teams that skip it are perpetually surprised on launch day.
classifying a single feature. A framework earns its mention by producing a judgment.
isn't optional — research is only done when it names the artifact it changes.
installing the watch cadence, or it didn't end — it just stopped.
market-landscape-scan (Workflow) — Step 1competitive-research-snapshot (Workflow) — Step 2voice-of-customer-miner (Workflow) — Step 3company-intel (Workflow) — Step 4positioning-statement (Component) — Step 5competitive-intel-watch (Workflow) — Step 6 cadence + maintenancebattle-card-builder (Workflow) — Step 7 outputintelligence-collection-disciplines (Component) — signal channels throughoutautonomous-investigation (Workflow) — evidence discipline throughoutquality theory; Moore, Crossing the Chasm (Whole Product); Ulwick, Outcome-Driven Innovation; Osterwalder & Pigneur, Business Model Generation; Ries & Trout, Positioning; Hamel & Prahalad, Competing for the Future (HBR, 1994); Nagji & Tuff, Managing Your Innovation Portfolio (HBR, 2012)
Other measured skills in the registry, with their headline benchmark lift.