Every sign is cut to order.
So we built a configurator
that cuts too.
A Belgian signage manufacturer sells across 22 product categories, each customisable to any dimension a customer asks for. We built a 2D and 3D configurator that assembles the product from cut profiles at runtime, produces the bill of materials, the quote, the order and the packing list.
- In business, 100 systems delivered
In business,
100 systems delivered
- Engineers on staff
Engineers
on staff- Quoting time, signage manufacturer
Quoting time,
signage manufacturer
- Sales growth in 12 months, LED manufacturer
Sales growth in 12 months,
LED manufacturer
A catalogue where nothing is quite standard.
The company sells standard signage across 22 categories and offers customisation in every one of them. In practice that means a customer's requirement is rarely a catalogue item. It's a catalogue item at a different length, with a different accessory set, in a different finish. Quoting one took around two days. Someone worked out which profiles were needed and at what lengths, which accessories attached, what the material came to, and what it would cost to ship because signage profiles are long, awkward, and expensive to send badly packed.
The variant count isn't large. It's unbounded.
With 22 categories, six to ten profile sizes in each, accessories on top, and any dimension the customer asks for, there is no finite list of products to model. Every attempt to pre-build the catalogue runs into the same wall.
Customers couldn't picture it.
A sign is bought on how it will look on a specific building. Drawings and renders had to be produced by hand for anything non-standard, which added days to a decision the customer wanted to make quickly.
Shipping was priced by judgement.
An experienced person decided how the order would be boxed and what it would weigh. That judgement was good, but it was an estimate and it went onto the quote as a cost.
We didn't model the products.
We modelled the parts.
The obvious way to build a 3D configurator is to model each finished product and let the customer pick one. With unlimited dimensions across 22 categories, that approach never finishes.
So we built a small model of every profile and every accessory instead the components the factory actually stocks. When a customer specifies a sign, the software cuts the profiles to length at runtime and assembles the accessories onto them, building the finished 3D model on the spot.The consequence matters commercially: there is no variant that hasn't been anticipated, because nothing is anticipated.The model is generated from the same components and the same rules the production floor uses.
The configurator does the same thing the factory does. It cuts to size and
assembles which is why it can produce a product nobody has ever ordered
before.
One specification.
Everything downstream.
Most configurators stop at the quote. But for this manufacturer, shipping was a real line on that quote and it was being estimated by eye.
- Configure
- 3D model
- AR preview
- BOM
- Quote
- Order
- Packing list & labels
2D and 3D, generated live.
The customer sees the actual product they've specified, at their dimensions, with their accessories not a representative image of something similar.
Augmented reality on the customer's own phone.
The sign can be placed against the real building, at real scale, before anything is manufactured. Customers use it themselves; the sales team uses it to close. For a product bought on appearance, seeing it in place removes the last hesitation and prevents the expensive kind of disappointment that arrives after production.
Bill of materials from the same model.
Profiles at their cut lengths, accessories, fixings. The BOM isn't derived separately from the visualisation it's the same assembly, described differently.
Quote and order, without an engineer.
Two days of specification work became roughly ten minutes, done by whoever is in front of the customer.
We solved the packing problem too.
Most configurators stop at the quote. But for this manufacturer the shipping cost was a real line on that quote, and it was being estimated by eye. Fitting a set of long profiles and irregular accessories into the fewest possible cartons is a bin packing problem formally NP-hard, meaning there's no exact solution at any useful size, only heuristics that get very close very quickly. That's what we built.
Optimised for both weight and volume.
Carriers charge on whichever is worse. The packing solution accounts for both, so the shipping cost on the quote reflects how the order will genuinely be sent.
Priced accurately, not generously.
An estimate has to be padded. A calculated figure doesn't which means quotes stopped carrying a hidden margin of caution on freight.
A manifest the customer can verify against.
The client is emailed a complete list and a box-by-box breakdown with the weight of each carton. When the delivery arrives, they can check every box against its stated weight confirming nothing was lost or tampered with in transit, before they sign for it.
Labels printed straight from the order.
The warehouse receives a packing plan rather than a puzzle. Labels print, boxes get filled in a defined order, and the time an experienced packer used to spend working it out disappears.
That last detail wasn't in any specification. It came from listening to what actually worries someone receiving forty boxes of aluminium profile.
"We're too custom for software" is usually an architecture problem.
Manufacturers tell us constantly that their products can't be configured, because every order is different. Very often what they mean is that every order is different in dimensions while being assembled from the same finite set of components, by the same rules, every time. That is a solvable problem. It just can't be solved by a system built around picking options from a list, which is what most configuration platforms are. Build from the components instead of the catalogue, and unbounded variety stops being an obstacle.
Is your product cut to size?
If your customers specify dimensions rather than pick from a list, and someone in your business turns that into a parts list by hand, this is the same problem. Thirty minutes. Show us one quote and we'll tell you honestly what could be automated.