The rebuild did not fix the lead problem, and was never going to
The question that started the rebuild was not “should we move off WordPress”. It was “why does the website produce no enquiries”. Those are different questions, and answering the second one honestly is what made the first one easy.
It is worth writing down because the sequence is usually reversed. A site gets rebuilt, it is faster and better-looking, the enquiries do not change, and nobody can say why — because nobody measured what the old site was actually doing before replacing it.
What the numbers said
Two figures settled it.
Of 9,174 form submissions all-time, 8,420 — ninety-two per cent — were the careers form. The site was, in volume terms, a job board. The contact form was a rounding error next to it.
Of search clicks over ninety days, eighty-seven per cent were brand queries — people typing the company name, who already knew it existed — and eleven per cent were job-seeker queries. Commercial, category-level searches, the kind where somebody is looking for software or a team and does not yet know who you are, produced approximately one click in the quarter.
Read together, those say something specific: the site was working perfectly well for people who already knew the company, and was invisible to everybody else. There was no demand-capture problem to fix in the design, because there was no demand arriving to capture.
What a rebuild can and cannot change
A framework change moves some things and not others. Being clear about which is the whole discipline.
It cannot create demand. Nothing about static generation makes a person search for out-of-home advertising software. If nobody in your category ever arrives, a faster site serves zero people faster.
It can remove reasons to be absent. That is the honest version of the SEO argument. The old site had no pages for the categories the business actually sells into — no page about quoting for made-to-order fabrication, none about compliance evidence, none about billboard inventory. You cannot rank for a question you have not answered. That is not a WordPress problem; it was a content problem that the rebuild was an opportunity to fix.
It can remove reasons to be distrusted. A site that has been serving spam URLs for two years, or that renders its contact address as a JavaScript-decoded link that a crawler cannot read, has technical reasons for being discounted that have nothing to do with what it says.
It can make the difference measurable. This is underrated. With content in typed collections and a build that fails on a missing description, the question “which of our pages are too thin to compete” became a script rather than an opinion. It returned ninety-five of two hundred and forty-five, and — the genuinely useful part — the thin ones were not the blog. The articles averaged over eleven hundred words. It was the commercial pages, the ones that should win buying intent, sitting at two and three hundred.
The uncomfortable finding
The best-performing content on the old site was the part nobody was trying to monetise. The careers pages and the technical articles did the work. The service pages, which existed to sell, were the thinnest things on the domain.
That is a common shape and it is worth checking for, because it inverts the usual instinct. The reflex is to add more marketing pages. The evidence said the marketing pages were the problem — not their existence, their emptiness.
What we would tell someone in the same position
Measure before you rebuild, and measure the right thing. Not traffic. Traffic that could ever become a customer.
Split your search clicks into brand and non-brand. If the ratio looks like ours did, a redesign is not your problem and a redesign will not be your solution; you have a category-presence problem, and the work is writing pages that answer the questions your buyers actually type.
Then check what your forms are actually receiving. Ours had been telling us for years, in a table nobody had aggregated.
So was the migration worth it
Yes, for reasons that are mostly not about leads.
The site cannot be reinfected, because there is nothing in the document root that executes. The publishing pipeline refuses to ship a page with a missing description. Deployment is a directory swap and rollback is the same swap reversed. The content has a schema, so the number of products described as being in production is derived from the products rather than typed into a sentence that goes stale.
Those are real and they compound. But none of them is a lead. The leads, if they come, will come from the twenty thousand words of category content the rebuild created the space to write — and that work would have been just as effective on WordPress. The framework made it easier to do well and impossible to do sloppily. It did not do it.