Modernising a business network from WordPress to Lovable

Over eighteen months, PC Consulting Asia migrated an entire network of businesses off WordPress and onto a modern stack built with TanStack Start, deployed to Cloudflare Workers, with images owned via a permanent CDN and every entity modelled in a network schema graph. This article documents the methodology.

The starting point was familiar. Every site in the network was on WordPress. Every site carried plugin sprawl accumulated over a decade. Images lived in WP media libraries with no responsive variants and no owned CDN. Schema was, generously, partial. Historical URLs numbered in the thousands per site — most of them technical residue from an earlier content-syndication era.

The temptation, and the industry default, is to redesign. Rebuild the pages, redirect the old URLs to the homepage, ship a new theme, count it done. This approach destroys entity signals more effectively than doing nothing at all. It was rejected on the first site and never revisited.

Audit-first

Every migration in the programme began the same way: no code, only documents. Discovery pulled the union of live crawl, Wayback Machine, Google Search Console and Semrush backlink targets into a single URL inventory. Every URL received a classification — keep, 301, or 410 — against a decision framework that respected both real search authority and real historical clutter.

Redirect philosophy

A 301 is a promise that a URL has moved. A 410 is an admission that a URL was never worth preserving in the first place. The industry uses 301 far too liberally and 410 far too rarely, and the cost is diluted entity signals across every site that inherits the aggregator debt.

The rule the programme used: if a URL has any backlinks from non-syndication sources, or any meaningful GSC impression history, or a human-recognisable slug, it earns a 301 to the nearest live equivalent. Otherwise, 410. Applied per URL, never per pattern.

Entity-first schema

Every entity in the network — each brand, each person, each service — is defined exactly once with a stable @id. Sibling brands reference each other by @id; they do not redefine each other. The network schema graph is a first-class artefact of the programme.

Image ownership

Every image in every site was migrated into an owned CDN. Original filenames preserved. Responsive AVIF and WebP variants generated. ImageObject schema on every editorial image. Image sitemaps for every site.

What comes next

Case studies for each site in the network document the specifics — what was audited, what was implemented, what validated, what the outcome was. Read them from the case studies index.