# Voice-of-Customer Miner — Worked Example (Industrial) All companies, quotes, URLs, and figures are **fictional**, including the verbatims (declared fiction, per the sibling example's rule). Industrial sibling of [`sample.md`](sample.md), and the harder case on purpose: industrial buyers barely post reviews. This run shows what honest mining looks like when public voice is *thin* — smaller numbers, wider bias notes, and a heavier lean on the one channel where practitioners actually talk. **Invocation:** `Mine voice-of-customer for Helix Foresight and retrofit monitoring generally — informs whether Northfield's integrator pitch leads with coverage or with compliance paperwork. Open sweep.` --- # Voice-of-Customer Snapshot ## 1. Scope **Products mined:** Helix Foresight; retrofit monitoring category **Decision supported:** whether the integrator pitch leads with mixed-vendor coverage or compliance documentation **Sources swept:** plant-maintenance forum (skew: hands-on maintenance leads, candid, small-N), two B2B review sites (skew: thin in this category — 14 reviews total across vendors), a controls-engineering subreddit (skew: integrators and engineers, vendor-cynical), trade-association webinar Q&A logs (skew: corporate engineering, polite) **As-of date:** 2026-11-24 ## 2. Need Themes ### Theme: one screen for a floor full of nameplates - **Frequency:** recurring — the forum's most-replied monitoring thread plus repeated subreddit mentions; total distinct voices ~12 (thin category; stated honestly) - **Verbatim:** "we have four vendors' gear on one line and four dashboards. nobody looks at four dashboards" — [maintenance forum, URL] - **Verbatim:** "the Foresight demo was slick until I asked about our non-Helix presses" — [subreddit thread, URL] - **Who says it:** maintenance leads and integrators — **Fact** (forum roles displayed) - **Reading:** mixed-vendor coverage isn't a feature preference, it's the daily reality the category ignores — **Inference**; corroborates the snapshot's structural read from the buyer side. ### Theme: prove the retrofit won't stop the line - **Frequency:** recurring — appears in every channel, including the polite webinar Q&A - **Verbatim:** "my last vendor's 'two-day install' cost us a weekend shift. never again" — [forum, URL] - **Verbatim:** (webinar Q&A) "what's the realistic commissioning window for a running line?" — [Q&A log, URL] - **Who says it:** plant managers and maintenance leads — **Fact** - **Reading:** installation risk outranks feature comparisons; case-study evidence of non-disruptive commissioning may be worth more than any capability claim — **Inference** ### Theme: the compliance binder nobody has time to build - **Frequency:** concentrated — one long forum thread on the new documentation rules, high engagement but a single venue - **Verbatim:** "the new docs requirement means somebody's weekend disappears every time we touch a control circuit" — [forum thread, URL] - **Who says it:** maintenance leads at smaller plants — **Inference** (self-described, unverified) - **Reading:** supports the documentation-automation wedge, but single-venue concentration caps confidence — pressure-test in integrator interviews before it leads the pitch. ## 3. Competitor Weak Points - **Helix Foresight:** mixed-vendor blindness, in buyers' words ("asked about our non-Helix presses") — recurring across two channels — [URLs] - **Sensor startups (category):** "great dashboard, then the pilot ended and nobody bought sensors for the other nine lines" — pilot-to-scale gap named by an integrator — isolated but vivid; watch, don't weaponize — [subreddit, URL] ## 4. Switching Triggers - **Push:** a botched install or a line-down event attributed to the vendor — the forum's switch stories all start with downtime — **Inference** (5 narratives) - **Pull:** the integrator's recommendation — no switch story in the sweep begins with an end-buyer's independent research — **Inference** (absence, noted) ## 5. So What? - **Opportunity hypotheses:** 1. Plants consolidate on whoever shows one screen across all nameplates — **Inference**, confidence: medium-high (recurring, multi-channel, small-N) 2. Non-disruptive commissioning evidence beats feature claims in this market — **Inference**, confidence: medium 3. Compliance documentation is a lead-worthy wedge — **Inference**, confidence: low (single-venue concentration) - **Battle-card-ready weaknesses:** 1. Foresight mixed-vendor blindness in customer words — evidence quality: medium (thin N, but consistent with their own docs — triangulated Fact) 2. Sensor-startup pilot-to-scale gap — evidence quality: low (one voice; do not put on a card yet) - **Assumptions to validate in real interviews:** 1. The four-dashboards pain drives budget, not just complaints 2. Commissioning-risk fear is decision-level for the *corporate* buyer too, or only the plant 3. The compliance thread generalizes beyond one forum's demographics --- ## Why this example works - **Thin voice, handled honestly.** ~12 distinct voices would be a rounding error in SaaS review mining; here it's the market. The example counts voices, names venue concentration, and caps confidence accordingly — instead of inflating a sparse sweep into false certainty. - **Answer to "which decision?" came back reversed from the temptation.** The user hoped compliance could lead the pitch; the mining says coverage leads (multi-channel, recurring) and compliance supports (single-venue) — public voice re-ordering the pitch is the skill earning its run time. - **Triangulation upgraded one weakness.** The mixed-vendor complaint is thin on its own, but it agrees with Foresight's own compatibility docs — customer voice plus vendor documentation is what makes it battle-card-grade despite the small N. - **The channel structure showed up in the switching triggers** — no switch story starts without an integrator. Even the VoC layer keeps pointing at the same strategic fact the whole industrial chain orbits.