Augmented Reality And 3D Configuration Are Perfect To Boost Sales

Configurator, AR preview, both or neither

A 3D product configurator connects the choices a product really offers to a visual of the result. A customer picks a finish, a size or a component, and the picture changes to match. An AR preview does a different job: it places a digital version of the product into the camera view of a real room or setting, so the customer can judge size and fit in context.

A project can use either, both or neither. The right route depends on the product, the options that are actually valid, the decision the customer is trying to make, the devices they use and who will build and run the experience. A set of good stills is sometimes enough.

Neither tool is a sales guarantee. A better question to start with is this: what decision should the customer be able to make with more confidence, and how will the business know whether that happened? The rest of this article helps you answer it.

3D product configurator or AR preview: which do you need?

The table compares five ways to show a product online. It is a planning aid, not a list of features any one project will have.

Image-layer configurators and real-time 3D configurators are different builds: one swaps prepared renders, the other draws the product live in the browser. An AR preview is optional on top of either, and which phones and browsers can show it depends on how it is delivered.

Compare interactive formats for a longer look at viewers and configurators, or test configurator feasibility against your own product range.

What to send before anyone builds

Most configurator problems start with missing information, not with the 3D. Before a studio or developer can scope the work, they need:

• Product revision and dimensions. Which version of the product, and its real sizes. If the product is still changing, say so.
• Finish references. Physical samples, finish codes or approved photographs for every material and colour.
• The option list. Every choice the customer can make, written out in full.
• Compatibility rules. Which options can and cannot be combined. The product owner sets and approves these. A picture of a combination is not a manufacturing specification.
• Audience and devices. Who will use it, and on what: phones, desktops, a showroom screen or a sales laptop.
• The intended output. A configurator, an AR view, a set of stills or a mix.
• Platform constraints. The site or shop platform, its rules on scripts and file sizes, and who can change it.
• Ownership and maintenance. Who builds the interface, who hosts it, and who adds new options when the range changes.

Ask one more question separately: does the chosen configuration need to go anywhere, such as a sales conversation, a quote or a basket? If it does, that connection is its own piece of work for whoever runs that system. Nothing here assumes a particular shop, CRM or CPQ connection.

Agree who builds and maintains each layer before production starts. If you want XO3D to build it, see our 3D product configurator service, and test configurator feasibility first if the option list is long.

A worked example (fictional)

This example is made up to show the decision. It is not a client project.

A furniture maker sells a lounge chair in a few fabrics and leg finishes. Customers keep asking two things: what a fabric looks like with a particular leg, and whether the chair will fit by the window.

The team lists what it has. It has CAD for the chair, approved fabric swatches and finish codes for the legs. It does not have a rule for one fabric that is not offered with one leg finish, so the product manager writes that rule down first.

Then the team tests three routes against the two questions:

1. A 360 viewer shows the chair from every side, but only in one finish. It does not answer either question on its own.
2. An image-layer configurator answers the first question for every allowed combination, because each one is rendered. The banned pairing is simply never shown.
3. An AR preview answers the second question, but only on phones that support it, and only once the product team has checked the model’s scale.

The team decides to build the configurator first and add AR later, once it has checked which devices its customers actually use. Your answer could be different. The point is that each route is judged against a real customer question and real data.

A colourway sweep from XO3D’s jewellery work: the camera and light stay fixed while the stone colour changes. An image-layer configurator relies on the same discipline, so every option can be compared fairly.

What XO3D’s own projects show

Two published XO3D projects show configuration work and who did what. Both are configuration examples. Neither is an AR deployment, and neither page records a sales or returns result.

• Acumen. XO3D built an image-layer furniture configurator from prepared renders. Visitors choose timber grade, table edge, wood finish, leg shape and steel-frame finish, and the selector swaps between rendered layers rather than drawing a live 3D model. XO3D’s web team handled integration into the client’s website. Later option updates were not part of the project. See the Acumen image-layer demonstration.
• Rollink Flex 360. XO3D made colourway imagery of the suitcase in eight finishes. Rollink’s own team built the colour selector on its website from that imagery. Read the Rollink Flex 360 case study.

The configurator feasibility guide covers both and explains why an asset commission alone does not settle who builds, hosts and updates the experience. For the wider question of who owns the platform and the visual asset, see Thomas Howcroft’s essay on choosing between a configurator platform and a CGI studio.

How to judge whether it worked

Earlier versions of this article promised more sales, fewer returns and more visitors. Those are not things any tool can promise. What you can do is set up a fair way to look:

1. Name the decision or behaviour. For example, choosing a finish, or checking that a product fits a space.
2. Pick the audience and the baseline. Who you are looking at, and what happened before the change.
3. Record everything else that changed. Prices, stock, campaigns and page design all move the numbers.
4. Read the results with the data owner. A controlled comparison can help. A rise that happens at the same time is not proof that the configurator caused it.

Any analytics or tracking added to a configurator or AR view needs a privacy and consent review by whoever owns the site. What is collected depends entirely on how the experience is built; there is no general rule.

Where to go next

• Product AR: preparing AR-ready models for a product page.
• Product interactives: viewers, configurators and touch-screen experiences.
• 3D product configurator service: image-layer and real-time configurators built from CAD.
• Configurator feasibility guide: whether your range suits a configurator.
• Shopify AR: what AR on a Shopify store actually changes.