Service 02 · Operations · Manufacturing Efficiency

Lift throughput, drop defects, keep the craft.

Bench-level workflow redesign, lean and JIT implementation, and defect-rate reduction for jewelry manufacturing — casting, finishing, setting, and QC — without flattening the craft into a factory line.

ENGAGEMENT 6-week floor diagnostic, then ongoing
SETTING Bench, casting, finishing, QC
VOLUME RANGE High-mix, low-to-mid volume craft
FORMAT On-floor + Zoom reviews
01Why this matters

Three places the jewelry floor leaks margin you can't see.

Jewelry manufacturing is high-mix, low-volume craft. Standard lean playbooks miss the point — but the underlying disciplines absolutely translate, with the right adjustments.

01 · Throughput

Wait time is invisible until it's measured.

Pieces spend 70–85% of their time queued between stations — not being worked on. Until WIP and wait time are mapped, every floor manager underestimates how much capacity is already there.

02 · Defects

Rework is the silent margin killer.

A 4% rework rate doesn't look bad on the report. But rework consumes prime bench time, breaks flow, and slips delivery dates the customer will remember. The compounding cost is 3–4× the visible one.

03 · Yield drift

Metal and finishing yields drift quietly.

Casting yield, polishing loss, and stone-set yield all move with batch size, alloy, operator, and supplier. Without a baseline and a scorecard, a one-point drift becomes a margin event a quarter later.

02Methodology

Walk the floor, map the flow, fix the bottleneck.

A five-phase system that starts at the bench, not the boardroom. Each phase has a defined exit criterion — measurable, named, and signed off — before the next begins.

01
Floor walk & diagnostic
Days on the floor observing actual flow — casting, finishing, setting, QC. Time-stamped photos of WIP queues. First read on the three or four highest-leverage interventions before any change is recommended.
Week 1–2
02
Throughput map
Current-state value-stream map with measured cycle times, batch sizes, wait times, and yields by station. The map is the artifact every later phase argues against.
Week 2–3
03
Pilot cell
A single product family run through a redesigned cell — pull-based replenishment, smaller batches, visual WIP control. The pilot proves the principle before any floor-wide change is made.
Week 3–6
04
Defect root-cause
Defect taxonomy installed, root-cause sessions run weekly with the bench leads, top three causes resolved before the program scales. Quality is built in, not inspected in.
Week 4–8
05
Scale & sustain
Rollout to remaining product families. Monthly KPI review against throughput, defect rate, yield, and on-time delivery. Operator training playbook handed to your supervisors so the system survives the engagement.
Ongoing
Throughput lift
15–25%
Typical capacity gain once WIP is controlled and batching is right-sized.
Defect drop
40–60%
Reduction in rework and final-inspection rejects after root-cause work lands.
Yield improvement
2–4 pts
Recovery on metal, finishing, and stone-set yield once baseline and scorecard exist.
03Deliverables

Working artifacts your supervisors can run.

Every deliverable is something a bench supervisor or a floor manager will actually use — not a slide deck.

Current-state value-stream map

A single-page map of cycle times, wait times, batch sizes, and yield by station — the artifact every later improvement argues against.

Pilot cell redesign plan

The physical layout, batch sizes, pull signals, and visual controls for the pilot cell — with a defined success criterion and rollback plan.

Defect taxonomy & RCA protocol

The named defect catalog, root-cause-analysis template, and weekly review cadence your bench leads will run after the engagement.

Floor KPI scorecard

Throughput, defect rate, yield by station, and on-time delivery on a single weekly page — the document your floor managers actually look at.

Operator training playbook

Station-by-station training notes for new and existing operators — written for shop-floor use, in your house language, not consultant-speak.

Yield baseline & recovery plan

The measured baseline for metal, finishing, and stone-set yield — with named recovery targets and the changes required to land them.

04Who it's for

Built for four kinds of jewelry operation.

The disciplines are the same; the right intervention depends on whether you are integrated or specialized, and on the volume mix you run.

A · Integrated

Multi-bench manufacturing floors

Vertically-integrated operations running casting, finishing, setting, and QC under one roof — where the leverage is in flow between stations.

B · Casting

Casting houses and specialist suppliers

Lost-wax casting operations focused on yield, alloy consistency, and surface quality before pieces leave for finishing.

C · Finishing

Finishing & polishing operations

Standalone finishing or setting houses where throughput is constrained by operator hours and the leverage is in station design.

D · Brand-owned

Brand-owned production facilities

Vertically-integrated brands running their own factory — where manufacturing and merchandising calendars need to lock together cleanly.

05Engagement

Floor diagnostic, then sustained partnership.

Manufacturing improvement is not a single project — it's a discipline. The diagnostic establishes the system; the retainer makes sure it survives the next quarter, the next year, the next product change.

Diagnostic

Six-week floor diagnostic.

Fixed-scope · single deliverable, no commitment beyond

  • Days on the floor — observed flow, measured WIP, time-stamped queues
  • Current-state value-stream map and prioritized intervention list
  • Pilot-cell scope and rollback plan, ready to run
  • Debrief with the owner, ops director, and bench leads
Book the diagnostic
06FAQ

Questions floors ask first.

The orthodox automotive playbook does not — but the underlying disciplines (small batches, pull replenishment, visual control, root-cause analysis) translate cleanly. The adaptation is in how, not whether. Jewelry needs craft-aware lean; we install that, not the textbook version.
No. The work is most effective when run alongside your operations leadership — your manager carries the floor relationship; we install the system, the artifacts, and the cadence. If your manager is the system, the system leaves when they leave. Our job is to outlive that.
The diagnostic is observation-only — no disruption. The pilot cell runs on one product family, contained, with a defined rollback. Floor-wide rollout happens only after the pilot has proven the gain. Production output is protected at every step.
Yes. Specialist casting houses, finishing operations, and setting houses each get the slice of the methodology that fits their station. Integrated floors get the full system. The work scales to where you are.
None required to start. The early phases run on direct floor observation, photos, and a working spreadsheet. If your operation justifies floor-tracking software later, we help scope and select it — but we do not lead with a software purchase.
Mutual NDAs are signed before any access. Floor photos and yield data stay with the engagement. Where we work with multiple operations in similar categories, we maintain clean information walls and decline engagements where conflicts cannot be honestly managed.
Ready when you are

Walk the floor with a second set of eyes.

One conversation is usually enough to identify the two or three highest-leverage moves on a jewelry manufacturing floor. Bring your hardest operational question first.

News from the practice

One note a month. Written by hand.

A short, useful note from the trade — what we’re working on, what’s changed in jewelry sourcing, operations, and brand. Sent only when there’s something worth sending. No tracking pixels, no upsells.