Short answer: find candidates through referral first and trade networks second, check financial stability and insurance before anything valuable moves, run a small documented trial job to test the process rather than the craft, and agree scope, price, date and insurance responsibility in writing before the first piece leaves. Then judge on dates met, not on the sample they showed you.
Where vendors actually come from
The trade's traditional channels are word of mouth, trade shows and associations. They work, and referral in particular remains the strongest signal — because a recommendation from another jeweler carries information about reliability, communication and how someone behaves when a job goes wrong, none of which appears in any listing.
The limitation is reach. Referral finds you the people your immediate network already knows, which means capacity gets discovered slowly and locally. A shop that needs a platinum setter this week, or a caster who will take a small run, often has no way to find one beyond asking around and hoping.
Trade directories and networks widen the field. They are better at discovery and worse at reliability signal, which is why the vetting below matters more when the candidate came from a listing than when they came from a jeweler you trust.
The checks worth running
Before anything valuable moves, five checks. None takes long and the first is the one most often skipped.
Financial stability. The Jewelers Board of Trade exists for this, and checking it before entrusting material to a new counterparty is standard practice in the trade for good reason. The advice appears explicitly in guidance on choosing refiners — check financial stability through the JBT — and the logic is identical for any vendor holding your goods. A firm in trouble is a firm whose creditors may end up interested in the parcel sitting on their bench.
Insurance. Who covers the goods while they are in the vendor's care, and to what limit. Do not assume. The general position with goods held by someone else is that a policy covering your own inventory does not automatically extend to goods in a third party's possession, and the same logic runs in reverse — the vendor's general liability may not cover the value of what you sent. Ask what they carry and for how much.
Longevity. Torry Hoover's advice on refiners generalises well: stick with companies that have been in the business a long time. It is not a guarantee, but survival correlates with the behaviours you want.
References at your size. Ask for two, from shops roughly your volume and in your segment. A vendor excellent with a large manufacturer's standing orders may handle a five-piece job from an independent quite differently.
Scope honesty. What will they not take on? A vendor who tells you plainly that something is outside their capability is worth more than one who takes everything. Andy Holwell of Johnson Matthey's line about refining applies to any trade service: if we think a lot is too small for our own process, we'll always be able to recommend someone else who we trust. That behaviour is a good sign.
The trial job
Never make a new vendor's first job an important one. But do make it representative.
Send something typical. A piece that resembles what you would actually send once the relationship is established. A simple job tells you they can do simple jobs.
Document it completely. Full written spec — metal, stones, sizing, finish, special instructions — plus photographs of the piece before it leaves. This is the standard you want the relationship to run at, and starting anywhere else sets a different expectation.
Agree a price and a date up front. Both in writing. A vendor who will not commit to a date on a small trial job will not commit on a large one.
Then evaluate the process, not just the piece. Did they confirm receipt? Did they call with a question before starting, or guess? Did the piece come back on the date? Did the invoice match the quote? Craftsmanship is necessary but rarely the differentiator — most trade benches can do the work. What varies is everything around it.
Terms to agree before the first piece leaves
Five things, in writing, every time:
- Scope — what work, to what standard, on what piece.
- Price — including how changes are priced if the job turns out different.
- Date — a specific return date, not a duration.
- Insurance and risk of loss — who covers the goods in transit and in their care, and to what value.
- What happens if something is damaged — particularly stones, and particularly customer-owned stones.
Point five is the one nobody wants to raise at the start of a relationship and everybody wishes they had raised at the end of a bad one. It is also the point at which you find out whether the vendor has thought about it, which is itself informative.
Don't build a single point of failure
A single caster or setter for a capability on your critical path is a risk you have chosen without noticing. If they get busy, get sick, or take on a large account, every job of yours that needs them stops.
Two vendors per critical capability is the practical minimum, with a third for overflow at peak. It costs a little in relationship depth and volume leverage, and it buys you the ability to say yes to a customer in December.
It also changes the pricing conversation. A vendor who knows they are your only option prices accordingly, usually without meaning to.
Judging them over time
The useful judgement is not formed on the first job. It emerges across dozens, and only if you record enough to see it.
Four measures carry most of the signal:
- Dates met against dates promised. Not average turnaround — the hit rate against what they said.
- Rework rate. How often a piece comes back needing something fixed.
- Questions before starting. A vendor who calls to clarify rather than guessing is protecting your margin as well as their time.
- Invoice against quote. Consistent drift is a pricing problem dressed as an administrative one.
Patterns emerge over time that are invisible in any three jobs, and they make both sourcing and pricing conversations materially better. But only if the history is somewhere retrievable — which is the point at which this stops being a vetting question and becomes a record-keeping one.
Where Jewelia fits
The finding is still referral and the trade's own directories. The judging is where software earns its place, and in Jewelia it sits in the platform rather than in a folder.
The judging sits in Production & Repairs, where outwork to a vendor is a tracked stage with its own due date and stays on your board rather than disappearing, and in analytics, where dates met, turnaround and throughput accumulate into a record rather than an impression. The CRM holds the vendor account with its terms and history, so the relationship belongs to the business rather than to whoever set it up. And the Planner turns the pre-deadline check-in into a task.
Plans start at $179 a month. The wider treatment is in our guide to managing multiple jewelry vendors and the buyer's guide to jewelry vendor management software.
Bottom line
Referral first, networks second. Check the JBT and check insurance before anything valuable moves. Run a small, representative, fully documented trial job and judge the process as much as the piece. Agree scope, price, date, insurance and damage terms in writing. Keep two vendors for anything critical. Then let the record — dates met, rework, invoice accuracy — do the judging over time, because the first impression is the least reliable data you will ever have about a vendor.
Frequently asked questions
- Where do jewelers find new trade vendors?
- Historically through word of mouth, trade shows and trade associations, which is why capacity tends to be discovered slowly and locally. Trade networks and directories widen the field, but referral remains the strongest signal because it carries information about reliability that no listing does.
- What should I check before sending a new vendor valuable goods?
- Financial stability through the Jewelers Board of Trade, insurance covering goods in their care, how long they have been trading, references from shops your size, and what they will and will not take on. The JBT check is the one most often skipped and the one that matters most when the goods are valuable.
- How should a trial job be structured?
- Small, representative and fully documented. Send a piece typical of what you would actually send, with a complete written spec and photographs, an agreed price and an agreed date. The point is to test the process — communication, documentation and date-keeping — not just the craftsmanship.
- What terms should be agreed before the first job?
- Scope, price and date in writing every time, plus who insures the goods while they are in the vendor's care, what happens if a stone is damaged, expected turnaround, and how changes are approved. Verbal quotes are where margin quietly disappears and where 'I thought you said' arguments begin.
- How many vendors should I keep for each capability?
- More than one for anything on your critical path. A single caster or setter is a single point of failure for every job that needs them, and it removes any leverage in pricing conversations. Two is usually the practical minimum; a third for overflow at peak.
- How do I judge a vendor after the first few jobs?
- On dates met against dates promised, rework rate, whether they call with questions before starting rather than guessing, and whether their invoices match their quotes. Craftsmanship is necessary but rarely the differentiator — most trade benches can do the work. Reliability is what varies.
Stop reading, start clicking.
Open a live Jewelia company with real inventory, memo, jobs and reports already in it. One work email, no card, no sales call, nothing to install.




