# Competitive Analysis Process — Worked Example (Industrial)

All companies, products, URLs, and figures are **fictional**. Industrial sibling of
[`sample.md`](sample.md). Where the SaaS example showed a full six-step pass, this one shows the
umbrella's *other* muscle: **scoping down and sequencing around a channel-shaped market** —
including a step run out of order for a reason the orchestrator states out loud. The delegated
artifacts live in the suite skills' own `sample-industrial.md` examples.

**Invocation:** `Run the competitive analysis process on retrofit automation for mid-size
manufacturers — decision: whether Northfield productizes down-market, and with what wedge. Budget:
six weeks, one PM plus an analyst half-time.`

---

## The orchestration transcript (condensed)

**Step 0 — Scope.** Decision is two-part (productize? which wedge?), competitor set unknown,
budget tighter than the SaaS example's. The orchestrator recommended:

> With six weeks and 1.5 people, here's my recommended scope:
>
> 1. **Steps 1, 2, 6 now; 3 and 5 folded into a single voice-and-positioning pass; skip 4 except for one company (Recommended)** — the channel structure in this market means perception (Step 5) lives with integrators, not end-buyers; a full separate pass would interview the same people twice
> 2. Full six-step pass — fits the decision but not the budget; adds ~3 weeks
> 3. Steps 1 and 2 only, decide, backfill — fastest; you'd be betting the wedge choice on structural evidence alone
> 4. Something else
>
> Choose a number, or adjust the budget and I'll re-scope.

User chose 1.

**Step 1 — Landscape** → `market-landscape-scan`
([industrial artifact](../../market-landscape-scan/examples/sample-industrial.md)). Deep-dive
players: Helix Motion Systems, Meridian Freight Systems. *Orchestration note:* the scan's
channel finding (integrators own the recommendation) restructured the remaining steps — it's why
Step 5 merged into Step 3, and why every later artifact treats the integrator as an audience.

**Step 6 — Strategic direction, run EARLY (out of order, on purpose).** With an adjacent entrant
(Meridian) suspected, direction evidence gates everything: if their entry were imminent, the
productization question changes shape. The fused read — certification filing (MASINT-adjacent
registry) + customs volumes + an early-access form ([tracker artifact](../../pricing-packaging-tracker/examples/sample-industrial.md))
— dated their entry 12-36 months out. **Verdict: proceed; install tripwires.** The SaaS example
ran 1→6 in order because nothing threatened the sequence; here the threat picked the order. The
umbrella bends to the decision — that's the skill's own teaching, demonstrated.

**Step 2 — Product comparison** → `competitive-research-snapshot`
([industrial artifact](../../competitive-research-snapshot/examples/sample-industrial.md)).
Verdict shaping the wedge: lead with software for mixed-vendor floors; Foresight's single-vendor
scope is structural, not temporary (corroborated later by the packaging split in the tracker).

**Steps 3+5 merged — customer needs and perception, one pass** → `voice-of-customer-miner`
([industrial artifact](../../voice-of-customer-miner/examples/sample-industrial.md)) plus
integrator win/loss debriefs (HUMINT — the perception layer in a channel market). Finding: the
"four dashboards" pain corroborates the mixed-vendor wedge from the buyer side; the compliance
wedge is real but single-venue — support, don't lead. Perception verdict: in this market,
positioning is won in the integrator's head; the uncontested space is "the vendor that makes the
integrator the hero."

**Step 4 — Business baseline, one company only** → `company-intel` on Helix. Corvid was skipped
(post-split distraction already evidenced at snapshot depth; a full baseline wouldn't change the
wedge decision) and Meridian deferred to the tripwires. Skips recorded with reasons.

**Step 7 — Outputs.**

| Output | Built | Owner / consumer | Fed by steps |
|---|---|---|---|
| Battle card: Northfield vs Helix Foresight | ✅ ([artifact](../../battle-card-builder/examples/sample-industrial.md)) | integrator-channel sellers | 2, 3+5 |
| Threat assessment: Meridian entry, dated + tripwired | ✅ | leadership | 6 |
| Positioning counter-move: "make the integrator the hero" | ✅ | PMM + partner program | 3+5 |
| Executive comparison matrix | ✅ (Step 2's matrix) | exec staff | 2 |

**Maintenance cadence installed:** quarterly `competitive-intel-watch`
([artifact](../../competitive-intel-watch/examples/sample-industrial.md)) — quarterly, not
monthly, because filings and registries move at quarterly speed; `pricing-packaging-tracker` on
the same cycle; Meridian tripwires (distributor listings, certification progress) event-driven.

---

## Why this example works

- **The umbrella's judgment calls are the content.** Merging Steps 3+5 (because perception lives
  in the channel), running Step 6 early (because an entry threat gates the decision), and
  skipping most of Step 4 (because it wouldn't change the wedge) — each with its reasoning
  stated. Compare the SaaS sibling's full ordered pass: same skill, different market, different
  shape.
- **Cadence matched evidence speed** — quarterly, where the SaaS example went monthly. The
  fusion-cadence table isn't decoration; it set this schedule.
- **Every delegated artifact exists** — follow the links to see each step's work at full depth in
  the same fictional universe, including the cross-corroborations (tracker + watch both reading
  Meridian's staging).
- **The two siblings together teach the meta-lesson:** the six steps are a menu and a grammar,
  not a ritual. Neither example runs them the same way, and both say why.