Your team wants interactive 3D assets for a catalog sourced from China. A phone scan appears quick, so the initial budget multiplies scan time by the number of products. That estimate leaves out an important question: how much work stands between a captured object and an approved product-page asset?
ONCE recommends buying a representative pilot before committing the catalog to one production route or unit price. The pilot should expose the difficult cases, rather than demonstrate only the easiest product. This is a proposed procurement method, not a report of an ONCE client test.
Scan duration is not delivery duration
Shopify’s 3D scanner guidance, checked October 31, 2026, estimates 10–15 minutes to scan and build a model using its compatible devices. It also describes capture difficulties associated with reflective, smooth, flexible, fine or open structures, and acknowledges that not every product will produce a high-quality result.
That describes a specific tool’s operation, not an all-inclusive commercial turnaround. It does not establish the cleanup, approval or integration time for your catalog.
One observable commercial alternative is Custivo’s per-product modeling quote, supported by photos, dimensions and material information. This is evidence of that supplier’s offer, not proof of an industry-wide pricing shift or a guarantee that modeling always beats scanning.
Select samples by difference, not at random
Three matte boxes cannot validate a catalog containing glass, mesh and upholstery. Inventory the differences that could change the production work:
- Surface: matte, polished, transparent, patterned or carrying important print detail.
- Structure: thin parts, openings, occluded areas, moving components or flexible forms.
- Use: rotating product-page inspection, room placement, close-up advertising or future color changes.
Select a typical item within each meaningful group, then add a challenge sample where an error would undermine the selling point. There is no universal sample count. A polished object with fine structural details may challenge more than one part of the workflow.
Take the pilot all the way to acceptance
Specify the output files, intended viewer, critical product details and reference information before capture. Agree tolerances and viewing conditions for the intended use; do not borrow an unverified universal accuracy threshold.
Retain the initial capture, the cleaned version, an explanation of repairs and a review record in the intended environment. The brand checks product facts, the production partner explains the work, and the implementation owner checks deployment. A file opening in modeling software is not evidence that the product-page experience has passed.
Shopify’s merchant modeling guidance also calls for review by both modeler and merchant. Keep a clear distinction between a captured asset and an approved deliverable.
Let the difficult sample change the route
Consider a hypothetical range containing a matte organizer, a polished kettle and an upholstered cushion. Success on the organizer does not establish the workload for the other two. The kettle needs scrutiny of its outline and reflective appearance; the cushion needs an agreed physical state before a static model can be judged.
This does not mean those products can never be scanned. Additional reference photography, reconstruction, material work or a hybrid route may be appropriate after testing. Any proposed physical preparation, such as coating or marking a sample, needs the brand’s prior agreement. A geometry-capture treatment must not silently become the represented sales finish.
If the required result is outside the current budget, retain photography or video for that group and defer the model. Producing a file is not the commercial objective; helping customers understand the actual product is.
Ask for a batch estimate that explains the work
ONCE proposes separating pilot validation, shared preparation by group, capture or modeling by SKU, cleanup and materials, agreed deployment checks, and authorized additional rework. Work already paid for in the pilot or setup should not be charged again without explanation.
Ask what genuinely repeats. Templates may be reusable; silhouettes, print details and repairs may still require individual work. Even a color variant needs confirmation that the finish and markings are otherwise equivalent.
Where repair effort is unknown, commission a bounded pilot, then agree group prices, exception handling and stop conditions. No unit rates, average labor times or success percentages are assumed here. This scope can sit alongside a product content production brief, with separate approval for the 3D outputs.
Keep the sample version, failure locations, repair actions, use case and final acceptance. Those records may support repeatable service design, but future margins still depend on actual effort and willingness to pay. Start the next quotation with a catalog overview and representative samples, rather than a single promised price for every material.
Source scope: the official Shopify pages and supplier offer above were checked October 31, 2026; publication dates are not inferred. This is evergreen procurement research. The sample method, budget structure and scenarios are ONCE analysis, not measured outcomes.
