Written for the person who has to make this work after the contract is signed. Vendor-neutral where it can be, explicit about where Ring Facet takes a position.
Anyone can make a single round solitaire look acceptable. Melee stones in a halo or a pavé band are where cheap rendering shows — they smear into a grey ridge, or every stone lights identically because they are one mesh. Rotate slowly and watch whether individual stones catch light independently. Then switch to a step cut: an emerald that sparkles like a brilliant means the engine is not modelling the cut.
Most bridal traffic is mobile. Test time to first interaction, whether the option panel is usable one-handed, and whether switching setting five times leaves the page sluggish. Ask what actually downloads when a shopper changes metal — if it is a new model file rather than a shader parameter, every colour swap costs a round trip.
Pick the least common shape and walk every setting. If the vendor multiplies shapes by settings without a compatibility map you will find assemblies that render wrong — a marquise floating above a basket built for a round. Then check the reverse: are options you cannot sell simply hidden, or greyed out and clickable?
If a metal is +10% and a setting is +10%, is the answer +20% of base or +21%? Both are defensible; only one matches what a merchant means when they type it. Ask the question — the answer tells you whether anyone has thought carefully about the pricing model. Then ask what happens if two options are added in a different order.
If the configurator posts its computed total to the cart, yes. The safe pattern is that the browser sends selections and the server computes money — a Shopify Function, or a server-side pricing endpoint on other platforms. Ask which one, specifically.
Ask for a live URL for each one that matters to you — diamond feeds especially. Roadmap items are legitimate; roadmap items presented as shipped are what turns a launch date into a professional-services quote. Ring Facet's own answer is on the diamond feeds page.
And if so: does it require re-exporting with particular layer names, is there a per-model setup fee, and does an unfamiliar stone cut render correctly without bespoke work? Our answer.
Saved builds outlive catalogue changes. If a link naming a discontinued setting shows an error or an empty step, expect support tickets from your most engaged customers.
Anything that requires forking your theme becomes your maintenance problem at the next platform release. App blocks and iframes do not.
Someone has to make the ring. Check that every selection reaches the order in text a person can read without opening a 3D viewer.
A dedicated design-your-ring page ranks and converts differently from a builder embedded in every product page. Most stores want both: a landing page for paid and organic traffic, and an entry point from the collection a customer is already browsing.
The instinct is to enable everything. Resist it. Every option you cannot deliver at your quoted lead time becomes a support conversation and a refund.
Configure five rings across your range and price them the way you price them today. Differences are almost always a missing add-on rule, and they are far cheaper to find now than in a cart.
Configure, add to cart, check out, and look at what arrives in your admin and your inbox. Then hand it to whoever makes the ring and ask whether they could work from it.
The highest-value use of a configurator in the first month is usually in-store, not online. Staff who use it daily will also find your catalogue gaps faster than any QA pass.
The glossary covers the jewellery terms and the 3D commerce terms in one place, which is useful when a vendor call has both a bench jeweller and a developer on it.