WordPress migration and custom calculator for home care services

Challenge

The existing WordPress site used Elementor Pro and had accumulated overlapping plugins, integrations, and workarounds that made it difficult to maintain. The previous builder generated deeply nested containers and unnecessary div elements, while the number of active plugins added weight and increased the surface area for technical issues.

The project started from a documented GTmetrix baseline of grade D and 45% Performance. The migration had to address that technical debt without losing published articles, media, existing URLs, SEO metadata, or operational features.

The site already had established organic visibility, so protecting its SEO equity was a critical constraint. The rebuild had to preserve the relevant URL and path structure, slugs, content, media, metadata, and internal relationships that search engines and visitors already relied on. This added a separate migration and validation effort beyond rebuilding the visual layer.

Moving to a new server also created an architectural decision: copying the installation unchanged would have moved the same problems to new infrastructure. The useful content had to be separated from the visual structure and dependencies that needed to be replaced.

The organisation also needed a tool that could help a family caregiver estimate the possible financial compensation available for caring for an older adult. This required custom business logic that could not be handled by the page builder alone.

Solution

My work covered the technical audit, migration planning, frontend reconstruction, custom plugin development, and final validation.

I prepared a new WordPress environment and rebuilt the main pages and templates with Oxygen Builder rather than transferring the existing Elementor layouts. This provided more direct control over the HTML, styles, and components while preserving the articles, categories, tags, featured images, slugs, public URLs, and required SEO settings.

To migrate the editorial archive without bringing the Elementor structure into the new installation, I built a custom migration tool that connected to the WordPress REST API. It extracted the required blog posts and media from the source site so they could be transferred while retaining their public paths, content, and asset references. This made the content migration repeatable and reduced the risk of manual omissions.

I audited the plugin inventory to identify duplicated responsibilities, small features implemented through oversized packages, and components that were no longer required. This reduced the number of installed plugins by approximately 80%.

The caregiver calculator was implemented as a custom WordPress plugin. It collects information about the care situation and produces an initial estimate of the possible compensation a caregiver may be able to request. The service criteria and business rules were defined with the project team; my responsibility was the technical implementation and integration into the new site.

Other critical features were reviewed individually before their previous dependencies were removed. This included an embedded React application, editorial workflows, SEO configuration, forms, and external services. The server migration was supported by backups, validation checks, and a rollback strategy before publication.

Outcome

The site was published from the new server environment with its main templates rebuilt in Oxygen Builder and without Elementor as a presentation dependency for those pages.

The existing public URL and content structure was carried into the rebuilt site as part of the SEO-preservation strategy. The migration retained the paths, editorial archive, and required media rather than replacing them with a new information architecture. This case does not claim a measurable ranking change because no post-migration ranking comparison is available.

The plugin inventory was reduced by approximately 80%, leaving a smaller installation with fewer third-party dependencies and clearer responsibility for the features that remained. The custom calculator also introduced a new self-service flow for obtaining an initial compensation estimate.

A GTmetrix test generated on March 31, 2026 from London recorded grade B, 83% Performance, 85% Structure, a 1.9-second Largest Contentful Paint, 0 ms Total Blocking Time, and a Cumulative Layout Shift of 0.01. This improved on the documented baseline of grade D and 45% Performance.

The report is a point-in-time test rather than continuous monitoring. It also identified a 4.94 MB page payload and initial server response time as remaining optimisation opportunities, giving the project a concrete basis for future performance work.