Preventive SEO means auditing and fixing technical issues before a site goes live, not after Google penalises it. This case study shows how that approach delivered 100%+ traffic growth in one month for a high-value eCommerce client migrating from a legacy CMS to Next.js.
Good technical SEO is the single fastest lever for traffic recovery, and this post shows exactly why. When most people search for preventive SEO they land on articles about negative SEO, but the scenario that deserves that label happens far more often: a website redesign handled without an SEO safety net. That is what this case study is about.
The client has given permission to publish these screen captures provided the business cannot be identified. That is standard practice for every case study we publish.
The Website
The site in question is a well-established eCommerce business and a category leader in its niche. The market is not enormous, but the economics are significant: average order value sits around $4,000, with products ranging from $500 to $25,000.
We originally engaged with this client to resolve a critical Google Merchant Center (GMC) display problem. The fix, as in roughly 90% of cases, came down to technical SEO: removing incorrect structured data and replacing it with the correct Schema markup. One principle worth internalising here: it is better to have no Schema than to have the wrong Schema. Incorrect markup actively misleads Google’s parsers and can suppress product eligibility entirely.
With the correct Schema deployed and the right link signals in place, the site went from zero GMC impressions to 4,000 overnight.
The Redesign Challenge and the Case for Preventive SEO
The client’s site was ageing. Its custom CMS had grown unwieldy and its design was outdated. The decision was made to migrate to Next.js with a completely new front end. The new pages looked excellent and the CMS architecture was a genuine improvement. The concern was traffic: a poorly executed migration can erase years of ranking equity in days.
Rather than launching and reacting to whatever Google did next, the client contacted us to oversee the transition before going live. That proactive posture is the definition of Preventive SEO: identify and resolve issues on a staging environment so Google never sees them on the production site.
| Metric | Value | Change | What it means |
|---|---|---|---|
| Reactive SEO | Fix issues after ranking drops | Weeks to months of lost revenue | Google demotes pages before you know there is a problem |
| Preventive SEO | Audit staging before launch | Traffic maintained or improved at launch | Issues resolved before Google indexes the live site |
The audit process combined a custom Python-based crawl system with manual code review of key templates. Because the site contained over 45,000 pages, the strategy was template-focused: auditing product.html once covers every product page. Individual page review at that scale is neither feasible nor necessary.
Technical SEO Findings
The initial audit report ran to 32 pages and covered a large number of issues. After filtering by severity and impact, two categories dominated: incorrect Schema and Client-Side Rendering (CSR).
CSR is a recurring problem for Next.js and similar JavaScript-heavy frameworks when server-side rendering is not configured correctly. Google’s crawler fetches the raw HTML and may never execute the JavaScript, meaning it sees an empty page. This was one of the primary findings in our research across 113 sites affected by Google core updates. Full details are in Google Core Update: New Research Explains the Losses.
| Issue | What it means | Severity |
|---|---|---|
| Incorrect Schema | Structured data that misrepresents content type or uses wrong properties, causing Google to ignore or penalise the markup | Critical |
| Client-Side Rendering (CSR) | Page content rendered by JavaScript after the initial HTML load; Google may index empty pages if SSR is not configured | Critical |
| Generic meta titles | All 45,000+ product pages shared the same title tag, removing any keyword differentiation | High |
| Missing meta descriptions | No unique descriptions across product pages, reducing click-through rate potential | High |
| Missing alt text | Images lacked descriptive alt attributes, affecting both accessibility and image search visibility | Medium |
| Identical OG tags | Open Graph descriptions for social sharing were the same across all pages | Medium |
The development team fixed issues iteratively as we reported them. New deployments occasionally introduced new problems, and the CSR issue in particular required several rounds of correction before it was fully resolved with proper server-side rendering.
The GMC Manual Action
A miscommunication during the staging phase led the team to submit staging product feeds to Google Merchant Center. This triggered a manual action on the live domain. The action was resolved relatively quickly, but it introduced a delay and illustrated an important GMC mechanic worth understanding.
Scaling with Python and AI: Fixing 45,000 Pages
Once the rendering and Schema issues were resolved, one major problem remained from the legacy CMS: all product pages shared a single generic title, carried no unique descriptions, had identical Open Graph tags, and contained minimal body content. Fixing this manually across 45,000 pages was not an option.
We built a Python pipeline to handle it at scale. The system worked as follows:
- Title tags generated from each page’s H1, ensuring keyword relevance without duplication.
- OpenAI’s API used to research each product and generate a unique body description.
- A shorter version of each description used as the meta description field.
- Alt text added to all product images, serving both SEO and WCAG accessibility requirements.
- Open Graph tags updated with unique descriptions for social sharing.
This is a practical example of Programmatic SEO applied to a large catalogue. The alternative, leaving 45,000 pages with identical titles and no descriptions, would have meant thin-content signals across the majority of the site’s index.
u003cpu003eThe original estimate for the full project was 15 to 20 days. It took two months. The additional time came from UX issues uncovered during the audit: users were abandoning the purchase flow before completing checkout. Since the agency’s roots are in UX research, we resolved those friction points as part of the engagement rather than leaving them for a separate workstream.u003c/pu003e
Results of Preventive SEO
The 90-day window below covers roughly the full project period, including the pre-launch dip. The thinner line shows the equivalent period from the previous year for context.
Traffic fell approximately 16% during the transition period. Given that Google’s data processing creates a lag in Search Console reporting, the real figure is likely slightly better than the graph shows. The following graph covers the 28 days after the new site went live:
What Actually Drove the Results?
The honest answer is that no single factor can be isolated. The traffic uplift is almost certainly the product of several improvements working together: correct Schema restoring product eligibility, server-side rendering ensuring Google could index content, unique titles and descriptions differentiating 45,000 previously identical pages, and a cleaner UX reducing bounce signals. Research into which factor contributed most would require controlled testing over a longer period.
What is clear is that the preventive approach prevented the catastrophic ranking losses that typically follow a poorly managed migration. For context on what those losses look like when preventive work is skipped, see our analysis of three sites affected by Google core updates.
Key Takeaways: Preventive SEO in Practice
If you are planning a website redesign or CMS migration and want to protect your existing traffic, the time to act is before the new site goes live. Preventive SEO is far less expensive than recovering from a ranking collapse. For further reading on what recovery looks like when preventive work is skipped, see our post on recovering after Google quality updates. For an overview of the technical factors that influence ranking shifts, see 33 key factors behind SEO ranking shifts.