Custom configurator or SaaS?
A SaaS configurator goes live in days on a subscription and handles a simple catalogue. Custom development costs more upfront, but it accepts any manufacturing rule, connects to your ERP and leaves you the code and the models. The decision depends on how many exceptions your catalogue carries.
What a SaaS configurator is
A SaaS configurator is a platform another company runs and rents to you. You upload models and options, pick a template and publish the configurator on your site for a recurring fee. Well-known examples are Zakeke, Threekit and VividWorks. The code is not yours and the scope is whatever the vendor sells.
The advantage is obvious: no programming. Within days you have a working configurator, with a 3D viewer, options and a connection to a store through the connectors the platform already has. For a simple catalogue it is a reasonable way to start and to learn what buyers ask.
The limit appears when your catalogue leaves the template. Platforms handle the common case very well — one product, a set of materials, one price — and the odd case less well: non-standard sizes, modules that condition each other, surcharges per finish, lead times that change with the combination. Each vendor has its own scope and terms, and they are worth checking before deciding.
What a custom configurator is
A custom configurator is written for your catalogue. It starts from your CAD, models your factory’s rules and builds the interface on top of your site. You pay for the project once, the code and models stay in your hands, and there is no platform fee.
It costs more upfront because it includes work a SaaS has already done: 3D engine, interface, options system, integration. In exchange, there is no ceiling on rules. If a combination is forbidden, it is forbidden; if price is calculated per linear metre, it is calculated; if the order has to land in an old ERP, it connects.
It also changes who decides. With a SaaS, the vendor sets the roadmap and you adapt your process. With custom work, the process leads and the software follows. For a manufacturer with rules of its own, that difference matters more than it seems in the first meeting.
Where they differ
They differ in four things: time to launch, upfront cost, rule ceiling and ownership. SaaS wins on speed and initial investment; custom work wins on flexibility, integration and control of the asset. Neither wins on everything.
| Criterion | SaaS (Zakeke, Threekit, VividWorks) | Custom |
|---|---|---|
| Time to launch | Days or a few weeks, from a template | Weeks or months, depending on scope |
| Upfront cost | Low: subscription and setup | Closed project quote based on scope |
| Recurring cost | Fee per store, product or usage | No platform fee; hosting and maintenance separate |
| Manufacturing rules | Whatever the vendor’s model supports | Whatever your factory needs |
| Integration | Standard connectors and vendor API | ERP, CPQ or CRM built to fit |
| Design | Widget adapted to your brand | Your site, your type, your flow |
| Ownership | The code belongs to the vendor | 100% of the code and the models |
When a SaaS fits
A SaaS fits when the catalogue is simple, the team has no development capacity and the priority is to launch soon. It also fits when the configurator complements a store rather than carrying the sale. In those cases the platform fee is cheaper than building, integrating and maintaining a configurator of your own.
- Few options per product: materials, colours, engraving or size.
- Price does not depend on the configuration, or moves with a simple table.
- The store already runs on Shopify, WooCommerce or similar, with a connector.
- There is no technical team, and no plan to build one.
- You want to test demand before investing in custom work.
- The deadline rules: a campaign or a trade fair is around the corner.
When custom development fits
Custom development fits when manufacturing rules are the product, when the configurator has to talk to the ERP or CPQ, and when the channel is B2B with distributors. It also fits when the experience has to be your website rather than a widget inside it.
- Free sizes, combinable modules or structural configuration.
- Price is calculated with rules and has to match the factory.
- The order enters the ERP or CPQ without manual transcription.
- Distributors configure on behalf of their own customers.
- You need logins, channel pricing or different catalogues per country.
- Mobile performance is critical, with concrete targets.
- The company wants to own the code and the models.
- You have hit a platform’s ceiling and every exception becomes a vendor ticket.
Cost over time
Comparing only the first invoice is misleading. A SaaS has a low entry point and a fee you pay for as long as the configurator runs; a custom build has a high entry point and maintenance you control. The crossover depends on the catalogue, the traffic and how many years you plan to use it.
SaaS fees usually scale with the number of products, stores or generated configurations. If the catalogue grows, the bill grows. With custom work, adding a product is modelling and rules work, not a plan change. Past a certain volume, that difference repays the initial investment.
When comparing, ask for both complete figures: three years of subscription at your expected growth, and the project plus its estimated maintenance. Then add the exit cost: what it takes to take the models with you and rebuild the configurator elsewhere. That number rarely appears in a SaaS proposal, and it is real.
Who owns the code and the models
With a SaaS, the code belongs to the vendor and you use the service while you pay. With custom development, the code and models are yours: you can change supplier, host where you want and keep selling if the relationship ends. Ownership is the least visible difference and the most expensive to recover.
The 3D models deserve their own paragraph. Many SaaS platforms take the models and keep them there, sometimes in a format you cannot reuse elsewhere. If you later want the same model for a printed catalogue, a showroom or a dataset, you need the original file. Always ask for an open format.
With custom work, delivery includes glTF for the web and, when needed, USD or FBX. The models can be reused in rendering, in a showroom or at a trade fair without paying for them again.
How to migrate from SaaS to custom
Migration is easier than it looks if the models are in an open format. You recover the models and rules, rebuild the logic and connect the order to the ERP. What you cannot recover is time: a proper migration takes weeks, not days.
The first step is an inventory of what lives inside the platform: models, materials, options, rules and usage data. The second is deciding what to keep and what to rebuild, because many rules were patches and should not be carried over. The third is building in phases: one product first, then the catalogue.
If the platform does not export the models in a usable format, they have to be rebuilt. That is the expensive scenario, and the reason to agree on export terms before signing, not after.
How to decide in one meeting
Four questions settle it: how many exceptions the catalogue has, where the price comes from, where the order goes and who wants to own the asset. If the answers are simple, a SaaS is enough. If there are rules of your own and an ERP behind, custom work pays for itself.
- How many real combinations does the product have, and how many exceptions?
- Is the price calculated, or read from a table?
- What has to happen when the buyer finishes: email, PDF or order line?
- Who maintains the rules when a new fabric arrives?
- What does the subscription cost over three years at the expected catalogue size?
- What happens to the models if you change supplier?
Frequently asked questions
Which is better, a custom configurator or a SaaS?
It depends on the rules. A SaaS solves simple catalogues with little investment and speed. Custom work pays off when there are many variants, manufacturing rules, configuration-based pricing or ERP integration, and when the company wants to own the code.
Are Zakeke, Threekit and VividWorks the same thing?
They are SaaS platforms for product configuration and visualisation, with different scope and terms depending on the plan. We cite them as examples of that model, not as a closed comparison: each vendor defines what is included and what is not.
Can we start with a SaaS and move later?
Yes, and it is a common route. The important part is agreeing from the start that the models can be exported in an open format. If the platform does not return them, migration means rebuilding the geometry and the cost changes.
Does custom development carry a monthly fee?
There is no platform fee. You pay for the project and, if you want, separate maintenance for catalogue updates. Hosting is on your side, as with any website.
How long does a custom configurator take?
A product with material variants usually runs four to eight weeks. A wide catalogue with manufacturing rules and ERP integration runs ten to fourteen. A SaaS goes live in days because it starts from a template that already exists.
What if my catalogue is not that complex?
Then you probably do not need custom development. If you have few variants and a fixed price, a SaaS — or even a well-made product page — solves it. Saying that early is part of the job too.
