A client’s website is not always the best place to test a new approach.
When we wanted to explore what Lovable could do with an established WordPress website, we chose Made Grand as the test case – gulp! That gave us room to experiment, measure the results and solve the difficult parts without passing the risk to somebody else (you’re welcome).
The original Made Grand website was a traditional WordPress build. Its SEO foundations were reasonably solid, Google had indexed its pages and WordPress remained a familiar and dependable way to manage the content.
We wanted to find out whether we could keep WordPress as the content management system and replace the public-facing website with a faster, more flexible frontend built using Lovable, without throwing away years of search authority.
Keeping WordPress while changing the frontend
The first stage was a headless build. WordPress continued to hold the articles, projects and other content. The new frontend retrieved that content and presented it through a modern React/Vite based website.
This meant the content management process didn’t need to change. Anyone already comfortable with WordPress could continue working in WordPress, while the public website could evolve independently. This approach is perfect for legacy CMS data and arguably the key reason we wanted to test it..
The build proved that the approach worked. It also exposed the areas that needed more attention.
What server-side rendering improved
The first React/Vite frontend relied more heavily on JavaScript in the browser. That is workable for visitors, but it can make it harder for search engines and other tools to see a complete page immediately. To see this, open a React website and view the page source. It’s often quite different from the page content you see on screen.
When server-side rendering became available in Lovable, we adopted it as the next practical improvement. The page content, titles, descriptions and structured information could then be delivered as part of the initial HTML response rather than depending entirely on the browser to assemble them.
That was an important step, but it wasn’t the end of the SEO work.
A modern rebuild still inherits its history
Google already knew the old WordPress website, which we could verify through Google Search Console and Google Analytics.
We had changed the URLs of some established pages. We don’t recommend doing this, but felt it was necessary to test our theories. Some had been replaced and others were no longer needed. Search Console continued to report legacy URLs while the new pages were being discovered and assessed. This caused a lot of extra work and was a bit of a headache as there are no instant updates, just an invisible send-and-wait process with no timeline.
This is a normal migration problem, but it needs active management. We worked through the old and new page structures, improved the redirects, aligned the canonical URLs and sitemap and made sure missing pages returned the correct response.
We also refactored the structured data so that Made Grand, its services, articles and author information formed a clearer connected structure instead of a collection of isolated pieces. Unfortunately, Lovable’s first pass didn’t meet our requirements, although its output has improved since.
Fixing measurement as well as visibility
SEO improvements are difficult to assess when the analytics are incomplete.
During the process, we found that the cookie-consent implementation we added by prompting Lovable was preventing Google Analytics from loading correctly. We repaired the consent flow, tested it in Tag Assistant and added meaningful events for page views, Calendly starts, email clicks and phone number clicks.
This gave us a better basis for judging what visitors actually do after reaching the website, rather than relying on page views alone.
The data also kept the project honest. Between 21 April 2025 and 12 August 2026, the connected search report recorded more than 76,000 impressions across Made Grand landing pages. It also showed that visibility didn’t automatically translate into enough clicks. That was an opportunity to identify where we could improve.
Because analytics consent wasn’t working correctly for part of the period, we will not pretend that every before-and-after comparison is perfect. The important point is that we found the issue, implemented a secure fix and now have stronger measurement for assessing future changes.
Why we experiment on Made Grand
There is always some risk when a long-established website changes platform, structure or rendering model. Rankings can move, old URLs can remain visible for longer than expected and analytics can reveal problems that were previously hidden.
We believe the responsible way to understand how AI-prompted websites are built and how they work is to test these changes on our own website first.
That gives us direct experience of:
- Modernising legacy WordPress websites
- Using WordPress as a headless CMS
- Building and refining Lovable frontends
- Moving from browser-rendered pages to server-side rendering
- Mapping and redirecting legacy URLs
- Repairing metadata, canonical URLs and structured data
- Diagnosing Search Console indexing issues
- Implementing consent-aware GA4 tracking
- Measuring the result over time
The experiment has not been about proving that one platform solves everything. It has been about understanding where automated tools help, where they fall short and where experienced technical review still matters.
Related reading
If you’d like to explore the technical side in more detail, these articles cover important parts of the journey:
- Testing Lovable’s new TanStack SSR setup: a real-world SEO experiment
A closer look at moving a Lovable build to TanStack SSR and what that meant for real-world SEO testing. - The hidden DevOps problem with headless WordPress and SSR
Why a WordPress API can work perfectly in a browser but fail when accessed by a server-rendered production website. - From Traditional WordPress to a Modern Headless Site Built with AI
The original overview of the Made Grand rebuild, including the headless WordPress architecture, AI-assisted development and the decision to preserve existing content, URLs and SEO authority.
Turning the lessons into a repeatable service
The work we have done on Made Grand is now becoming a reusable process (a ‘skill’ may be the preferred terminology).
We can apply the same structured approach to an established WordPress website, a headless CMS build or a modern frontend that has inherited SEO and analytics problems. That includes reviewing the existing evidence, protecting useful URLs, correcting the technical structure and putting reliable measurement in place.
You don’t always need to begin with a large SEO project or immediately bring in a separate SEO team. A capable technical review can often identify the important problems, establish the right foundations and show where specialist support would genuinely add value.
The best experiments are the ones you can measure, learn from and repeat. That is why we ran this one on Made Grand and accepted the risks involved.
If your website has grown through several platforms, migrations or partial rebuilds and the technical picture no longer feels clear, start a conversation with Made Grand. We’d be happy to share our experience.
If you're planning an MVP or early-stage product and want to make sure the foundations are right before you build, we're happy to talk it through.
Start a conversation