HAITU AI
Back to the knowledge base
Systems & digitalization9 min readUpdated 2026-08-27

Why window manufacturers need industry-specific software

Plenty of window factories have tried a generic ERP and ended up using it only for stock and accounts, or abandoning it. The failure gets blamed on staff resistance. The deeper reason sits in the data model: generic systems have no notion of a window type, and almost all of a window factory's complexity comes from exactly that.

1. The data model doesn't fit

A generic ERP assumes goods are enumerable SKUs with a fixed bill of materials. In this industry that assumption does not hold.

Layout, profile combination, glazing and hardware can all differ unit to unit, and a change in dimensions cascades into cut lengths and quantities. The same "casement window" can carry completely different bills of materials on two different orders. Force it into a SKU model and the predictable result is a parallel spreadsheet maintained outside the system.

There's a direct test of whether software really understands windows: can it expand a bill of materials from the window type and dimensions, or does it expect you to pick materials by hand?

2. Material calculation is the hard part

Whether costs are accurate and whether material is wasted comes down to the material calculation. And the rules for it usually live in the head of one experienced person — the same window, calculated by two people, can come out differently.

More variables are involved than outsiders expect:

  • Profiles priced by perimeter or by length, with cut lengths adjusted for how sections meet
  • Glass priced by area, usually with a minimum chargeable area and a safety allowance
  • Hardware driven by the number of opening sashes and the opening type — different types need entirely different parts
  • Seals and consumables by length or by set, and easy to leave out

Turning those rules from personal experience into maintainable formulas inside a system is the most fundamental difference between industry software and a generic one. Once the rules are fixed, a change of staff no longer changes the answer.

3. Slow quoting loses orders outright

This bites hardest in export. Buyers typically ask several suppliers at once, and whoever returns a clear, professional quote first holds the advantage. If your quote waits on one person's availability and a round of configuration checks, the buyer has moved on by the time it lands.

Slow quoting is rarely a matter of effort. It's a broken chain: the drawing sits with design, the calculation rules sit with one experienced colleague, the prices sit in another sheet. Bringing all three onto the same data is what actually brings the time down.

4. Where the data breaks

Five breakpoints are common in window factories, and each one creates re-entry and a chance to get it wrong:

  1. 1Drawing to material calculation: the drawing is finished, then the materials are worked out again by hand
  2. 2Calculation to quote: the resulting quantities are copied into a pricing sheet
  3. 3Quote to order: once confirmed, the order details are keyed in afresh
  4. 4Order to production: work orders are transcribed again, and a mistyped dimension becomes scrap
  5. 5Production to purchasing: shortages are worked out by manually checking stock

Every breakpoint is a manual re-entry. The transcription costs time; the errors cost material and delivery dates. The practical test of a system's value is how many of those breakpoints it removes.

5. When not to put a system in

Earlier isn't always better. In some situations a system creates more problems than it solves:

  • The products and calculation rules aren't settled. If the rules change weekly, configured formulas can't keep up — sort the rules out first.
  • Nobody owns the rollout. Someone has to organize data, configure rules and push adoption; a vendor cannot do that from outside.
  • You only need to fix one thing. If what you want is a quoting tool, a full system isn't required.

Conversely, these signals usually mean it's time: quoting starts queuing as volume rises, scrap and re-orders from calculation errors increase, quoting stops when one colleague takes leave, and departmental numbers stop reconciling.

6. Questions worth asking during selection

  1. 1Can it expand a bill of materials from the window type and dimensions? Can the calculation rules be customized?
  2. 2Are drawing, calculation, quoting, orders and production running on the same data, or are they separate modules stitched together?
  3. 3Can it export drawing files usable in fabrication, and do the dimensions and scale survive the export?
  4. 4Can quotations carry drawings, and can it produce a drawing approval sheet?
  5. 5How long does rollout take, what do you need from us, and who organizes the product and calculation data?

The last question is the one most often skipped, and it decides whether the system gets used at all. Organizing data and settling rules is work no vendor can do on your behalf.

Key points

  • The SKU model behind generic ERP doesn't fit made-to-order windows — that's the root difference
  • Material calculation is the hard part; the value is in turning experience into maintainable formulas
  • Slow quoting comes from a broken chain, not from lack of effort
  • Judge a system by how many manual re-entry points it removes
  • Hold off if rules aren't settled, nobody owns the rollout, or you only need one tool

Now, what should you actually do next?

Six questions, about a minute — see which stage you're at and where to focus. No contact details needed to view your result.

Take the free assessment