Lumox kept its WebGL site and added 41 crawlable pages
A Canberra software firm decided its animated website was not the problem. The missing second layer was.
Lumox Technologies had a website built around WebGL scenes, controlled scroll behaviour and motion. It also had a discovery problem.
The company's diagnosis, in its own account of the work, was that the animation was not at fault. Search engines and answer systems could not see the full commercial story around it.
The conventional fix is to replace the experience with a template that crawls cleanly. Lumox rejected that on the grounds that the experience was part of the product.
Instead it built two layers. The first kept the interactive site intact. The second was a set of pages written to be understood by machines and by buyers arriving with a specific question.
That second layer now runs to 41 canonical pages listed in the company's public XML sitemap. The figure is checkable there rather than taken on trust.
The pages cover the company, its services, industries, buyer guides, frequently asked questions and contact routes. A Canberra hub page connects local search intent to the relevant capabilities. The company says it publishes its location at city level only and no street address.
Lumox is blunt that page count was not the lesson. Each page was given a distinct job: a service page explains fit and scope, an industry page describes operating context without claiming experience the company cannot substantiate, and a guide helps a buyer decide before contacting a provider.
Underneath, the company says every page carries a canonical URL, a descriptive title, a meta description, social metadata and structured data, with service and organisation entities using stable identifiers so the pages resolve to the same business.
Internal links were treated as navigation rather than keyword placement. Anchor text describes the destination, on the reasoning that a person and a crawler should both know where a link goes before following it.
The checks are the part most teams skip. Lumox reports that automated tests fail the build when a canonical URL, a sitemap entry, a structured-data relationship or a required internal link goes missing, and that browser regression tests guard the original scroll boundaries and interactions.
The company is also candid about the ceiling. Technical readiness does not produce authority; a search engine can crawl a page and decline to index it.
So the next phase is measured on indexed pages, independent referring domains and qualified search visibility rather than page count. Whether those numbers move has not been reported.
The takeaway for a founder sitting on an expensively designed site is a sequencing one. Before stripping out the thing that makes the site distinctive, check whether the business also has a stable, crawlable description of itself somewhere on the same domain. Adding that layer is cheaper than rebuilding the first one.