From WordPress to Next.js: migrating without losing SEO (real case with numbers)

Legal-tech client with 6 years on WordPress, 340 indexed URLs, decent domain authority (DA 38) and top-10 positions for 28 commercial keywords. Their problem: the site loaded in 4.2s on mobile, the WordPress dashboard was a battlefield of conflicting plugins, and every WooCommerce update broke something. They wanted to migrate to Next.js but were terrified of "losing all the SEO I've built for 6 years".

Result at 8 weeks post-go-live: stable organic traffic (-2% first month, +14% second, +27% third), 0 keywords lost from top 10, LCP from 4.2s to 0.9s on mobile, contact form conversion +35%, hosting bill from €89/mo (shared server + Cloudflare) to €18/mo (Vercel hobby + Supabase). The secret wasn't speed: it was that we DIDN'T break a single URL, schema or canonical tag.

Decision matrix: should you migrate? Migrate IF: (1) your mobile LCP is over 3s and you've exhausted cache plugins. (2) your dev team suffers with every WP/plugin update. (3) you need features WordPress does badly: real-time dashboards, AI/Claude integration, multi-tenant, serious i18n. (4) mobile traffic is growing and Core Web Vitals slow you down. DON'T migrate if: (a) your CMS is operated by marketing without a technical team — you lose visual editorial freedom. (b) you use WooCommerce with 500+ products and very specific plugins. (c) you have editorial content (large blog with non-technical writers) — first look at headless options like Sanity/Payload before going fully static.

The 301 redirect plan (most underrated step): download full sitemap.xml from WordPress + Search Console URLs (the ones receiving clicks). Cross both in a CSV. For each legacy URL write the new one. Four cases: (1) same URL → no redirect needed. (2) URL changes slightly (e.g. /sobre-nosotros/ → /about/) → direct 301. (3) URL disappears without a successor → 301 to parent section, NEVER to home (Google flags it as soft 404). (4) URL disappears without context → 410 Gone, not 404. Implement redirects in next.config.js or, if >100, in a Cloudflare Worker rule that runs before Next.js.

Replicating existing schema.org — what 80% forget: WordPress with Yoast or Rank Math automatically emits Organization, BreadcrumbList, BlogPosting, FAQPage, Person. Your Next.js must emit EXACTLY the same schemas, with the same @id, same URL and same structure. Run Schema.org Validator on 20 representative URLs BEFORE launch and THE DAY AFTER. If a migration loses BlogPosting.dateModified or Organization.@id, Google starts re-evaluating the entire site and you lose 30% traffic for 4-6 weeks.

Performance: WordPress vs Next.js — the real benchmark from this project: WordPress home with 12 active plugins and WP Rocket cache → LCP 4.2s, CLS 0.18, TTI 6.1s, weight 2.4MB. Next.js home with SSG + ISR + next/image → LCP 0.9s, CLS 0.02, TTI 1.4s, weight 380KB. Those are real numbers measured with PageSpeed in production. The improvement comes from 3 places: (1) WebP/AVIF images generated at build vs WordPress serving 1MB PNGs. (2) JavaScript hydrating per component vs WordPress loading jQuery + 8 plugins on every page. (3) Static HTML cached at CDN vs WordPress generating HTML on-demand per visit.

The 5 things that ALWAYS break (saving you 3 weeks of panic): (1) Images with WP Smush's custom .webp extension break next/image — fix: build-time pipeline with sharp that regenerates EVERYTHING from originals. (2) Shortcodes embedded in posts (Elementor blocks, CF7 forms) — fix: identify the 8-10 most common patterns and create equivalent React components. (3) WooCommerce sessions: if you migrate the catalog to Shopify/Stripe, user sessions are invalidated — communicate 7 days ahead to customers and prepare cart recovery. (4) Schema markup with legacy @id — replicate IDENTICAL for the first 90 days. (5) Hreflang in multi-language sites — Next.js has quirks with i18n in App Router; use explicit <link rel='alternate' hreflang> in each page's metadata.

Real project timeline (6 weeks): W1 — complete audit: sitemap, schema, performance, plugins, external integrations, critical dependencies. Migration CSV approved by client. W2 — Next.js skeleton development: layout, header, footer, typography system, design tokens, first 5 templates. W3 — content migration: systematic scraping of existing WP, conversion to Next.js structure (MDX or JSON), image regeneration in sharp pipeline. W4 — replicate schemas, redirects, integrations (Google Analytics, Tag Manager, forms, CRM). W5 — full staging with temporary domain, exhaustive QA (clients test 100 random URLs), patch final bugs. W6 — DNS swap during low-traffic window (typically Sunday early morning), 72h monitoring, fine-tuning. Legacy WordPress remains 30 days in pre-prod as emergency rollback.

Indicative cost of a real 2026 migration (SprintMarkt pricing): simple corporate site (15-30 URLs, no e-commerce, no multi-language): €4,500-6,500. Site with large blog (100-300 URLs) + multi-language: €7,000-11,000. Corporate site + WooCommerce with 200+ products to Stripe/Lemon Squeezy: €12,000-18,000. Site with client dashboard / custom management panel: €16,000-25,000. All quotes include redirect plan, schema replication, image regeneration, staging testing with 5 client rounds and 60-day post-go-live monitoring.