Competitive Analysis Process — Worked Example (Industrial)
All companies, products, URLs, and figures are fictional. Industrial sibling of
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:
- 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
- Full six-step pass — fits the decision but not the budget; adds ~3 weeks
- Steps 1 and 2 only, decide, backfill — fastest; you'd be betting the wedge choice on structural evidence alone
- 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). 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) — 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).
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) 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) | 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) — 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.