Build vs buy a 3D product configurator

Building a 3D product configurator in-house can be the right call. It can also burn six figures and 18 months before you have anything your customers can use. This guide breaks down both paths honestly so you can make the right choice for your situation.

What "building in-house" actually means

A 3D configurator is not a form with a dropdown. It combines a 3D rendering engine (WebGL or a library like Three.js or Babylon.js), a rules engine that enforces what combinations are valid, a 3D content pipeline to get your CAD files into a web-ready format (GLB/glTF), and a quote or order output step. You need people who can do all four, or you will build two of them well and patch the rest.

Typical build effort for a mid-complexity configurator (10 to 50 product options, one product family):

• 3D modelling and optimization: 4 to 12 weeks depending on model count and format quality
• WebGL/Three.js development: 3 to 6 months for a senior developer
• Rules engine and pricing logic: 4 to 8 weeks
• Mobile and cross-browser QA: ongoing, 1 to 2 weeks per major browser/device cycle
• Accessibility (WCAG): easily skipped, costly to retrofit
• Hosting, CDN, and uptime: ongoing infrastructure cost and responsibility

Total first-year cost for an in-house build typically lands between €80,000 and €200,000, depending on team rates and complexity. That is before ongoing maintenance.

When building in-house is the right call

• You have a 3D or WebGL developer already on the team and they are not at capacity.
• Your configuration logic is genuinely unusual (e.g. structural engineering constraints, real-time pricing from an ERP, custom AR integrations) and no off-the-shelf platform supports it.
• You need deep integration with proprietary backend systems that platforms do not expose API access to.
• You are building configurator capability as a core product feature that competitors cannot replicate, not as a sales or quoting tool.
• You have a 3D content pipeline in place and the models are already web-ready.

When buying a ready-made platform makes more sense

• You have standard configurable products (dimensions, materials, options) without unusual logic.
• Speed to market matters. Most platforms can go live in days to weeks, not months.
• You do not want to own browser QA, mobile testing, WebGL updates, or hosting.
• Your development team is focused on the core product and cannot take on a 6-month 3D project.
• You want to validate whether a configurator actually improves your quote conversion before committing to a build.

Ongoing maintenance: the cost that surprises people

A configurator is not a one-time build. WebGL APIs shift. Mobile browsers update. New devices introduce layout bugs. A product range change means model updates. Plan for 10 to 20 percent of the initial build cost per year in maintenance if you build in-house. Platforms absorb most of this by default.

The parts most in-house builds never reach

Most internal projects stop at "the customer can rotate the 3D model and change a few options." That is the easy 60 percent. The features that actually move quote conversion are the ones that take the longest to build and keep working, and they are usually where an in-house build runs out of budget:

• Augmented reality. Letting a customer place the product at true scale in their own driveway or garden through their phone camera. Getting this right across iOS and Android, with correct scale and lighting, is a project in itself.
• In-context visualisation (photo matching). Dropping the customer's exact configuration into a photo of their own space, matched in 3D so the perspective and proportions look real. This is what turns "nice model" into "that is my house," and it is hard to build and harder to maintain.
• The back office. A configurator only earns its keep if the quote requests land somewhere useful. That means a dashboard to manage incoming leads, see each customer's exact configuration, and track conversion stats over time. Building the front-end 3D tool is the visible half; the lead and analytics back-end is the half that takes as long again.

If you build in-house, budget for these honestly or accept that you are shipping a demo, not a sales tool. A mature platform includes them as standard, which is much of what you are paying for.

A fair summary

	Build in-house	Ready-made platform
Time to first demo	3 to 6 months	Days to weeks
Year-one cost	€80K to €200K+	Varies; often per-use or SaaS
Customisation ceiling	Unlimited	Platform limits apply
Ongoing maintenance	Your team's problem	Mostly the platform's problem
Mobile/browser QA	Your team's problem	Mostly the platform's problem
3D content pipeline	Build or hire	Bring your own models

If you want to validate demand quickly and your product logic is not unusual, a pay-per-use platform like CPQ3D lets you embed a configurator in about two minutes and pay per quote request rather than committing to a build budget upfront.