3D as a service pricing: why we charge per product per month

Most 3D suppliers quote per model. You pay a few hundred for an asset, you get a file, and everything afterwards is your problem: hosting it, converting it for iOS, keeping it working when the fabric range changes, integrating it into a product template, and re-commissioning the whole thing when the model turns out to be the wrong size.

We price differently, per product per month, and this article explains why and what it costs. If you are comparing quotes, the comparison is not really about the price of a model.

The prices

Basic covers 3D asset creation, AR and 3D web integration, CDN hosting, ongoing updates and maintenance, replacement of up to 30 percent of your variants each year, and implementation.

Value adds everything in Basic plus CDN-optimised hosting, custom dimensions, virtual try-on, annotations, five rendered product images, a 360 render viewer, a 3D configurator with colour and texture selection, and custom implementation.

Enterprise is everything in Value with a tailored scope, custom integrations, dedicated support and a commercial setup negotiated around your volume. The current list is on our prices page and that is the authority if this article ages.

Why monthly rather than per asset

A 3D model is not a deliverable that finishes. Four things happen to it after it is built.

It has to be hosted and served. Two files per product, GLB and USDZ, on a CDN, with correct MIME types and cache headers, at whatever traffic your product pages get. That is an ongoing cost whoever pays it.

It has to keep working. Browsers change, iOS changes what Quick Look accepts, Android changes Scene Viewer's limits, your platform changes its theme structure. Assets built three years ago need attention. Under a per-asset contract, that attention is a new invoice or it does not happen.

The catalogue changes. Fabrics get discontinued, a leg finish is added, a size is dropped. Under per-asset pricing, every one of those is a quote and a two-week wait. The 30 percent annual variant replacement in the Basic tier exists because that churn is normal and pretending otherwise produces a stale catalogue within a year.

The integration is the hard part. Getting the viewer onto the product template, wired to your variant selector so the model matches the selected colour, loading fast enough on mobile, and not breaking when your platform updates. A file in a Dropbox folder does none of that.

The monthly model prices all four in, and it aligns the incentive. We are paid while the thing works, not when we hand over a file.

When in-house is cheaper, honestly

There is a real threshold, and it is worth naming rather than pretending it does not exist.

Build in-house when: you have a large, slow-changing catalogue, at least one full-time 3D artist you can keep busy, engineering capacity to own the viewer and the asset pipeline, and enough volume that a salary divided by product count beats a per-product fee. Roughly, a few thousand stable products with an existing design team that already works in CAD. At that scale the tooling is free, the expertise is the cost, and you will want control.

Use a service when: your catalogue is in the hundreds, it changes seasonally, you do not have a 3D artist, and your engineering team has a roadmap that does not include maintaining a WebGL viewer. This describes most furniture, homeware and speciality retailers we work with.

The hybrid, which is common and sensible: the service builds and maintains the library, your designers work in CAD and hand over source geometry, which lowers the modelling cost and raises the accuracy at the same time.

The mistake in both directions is the same. Buying models one at a time from a freelancer is the cheapest line item and the most expensive outcome, because nothing after delivery is anyone's job.

What to compare when you get quotes

• Does the price include hosting, or is that separate and metered by bandwidth?
• Does it include both GLB and USDZ, or only one?
• Does it include the integration on your platform, or a file and a link to documentation?
• What happens when a variant changes? Included, or quoted?
• Do you get the source files, so you are not locked in?
• Is the model dimensionally accurate to your specification sheet, or built by eye from photographs?
• Who fixes it when Safari breaks something?

A per-model quote that answers "no" or "extra" to most of those is not cheaper than a monthly fee. It is a different, smaller product.

Is it worth it at all

That depends on whether better product pages sell more, and the evidence there is reasonable. Shopify's 2020 merchant data found that interactions with products carrying 3D or AR content converted 94 percent more often than the same kind of products without. Rebecca Minkoff reported a 65 percent purchase lift on AR-enabled product pages and shoppers 44 percent more likely to add an item to the cart after seeing it in 3D. Those are observational, not controlled, so treat them as direction rather than a forecast.

The arithmetic you should actually run is your own. Take one product at 10 dollars a month, 120 dollars a year. How many extra units, or how many avoided returns, does it take to cover that? For most furniture and large-goods retailers the answer is one, which is why the pricing is structured this way.

The way to check is to put one on a page. Send us a product with photos and dimensions and we will build the model for free, hosted, so you can measure it before signing anything. Free sample here. What is included at each tier is on solutions, and the current prices are on prices.

Sources

• CharpstAR, pricing.
• Shopify, merchant data on 3D and AR content, September 2020.
• Shopify, AR shopping: Rebecca Minkoff results.