Auto site lost 1M traffic after Next.js migration - any fixes?

conversionninja

We run a big automotive content site - think news, reviews, specs - and it used to pull in over a million organic visits a month. Then we decided to revamp the whole thing and migrate to Next.js through an external agency. What a disaster. Traffic absolutely plummeted.

The migration was riddled with mistakes: improper implementation, broken internal links, content that took forever to get discovered. Crawling and indexing turned into a nightmare, and we're still patching things up months later.

One maddening example: a new article shows up on the homepage instantly, but it doesn't appear on the relevant category page for nearly an hour. That delays internal linking and basically tells search engines 'don't bother looking here'.

We've been working with devs to fix issues one by one, but recovery is glacial. current stack is Next.js + Cloudflare. For anyone who's crawled back from a disastrous migration: what actually moved the needle? Which technical SEO gotchas are most overlooked in Next.js? On a huge content-driven auto site, what would you audit first? Internal linking, crawl budget, Core Web Vitals, rendering - I'll take any battle scars you're willing to share.

adcraft

Sure, standard advice - but if your hydration is busted or you're fetching data client-side on dynamic routes, SSR won't save you. Check your getServerSideProps is actually running per request, not caching stale content.

sleepermode

are you sure this isn't tied to the Google algorithm changes over the past three years rather than the Next.js migration? I've seen plenty of sites lose traffic purely from core updates, especially in the automobile vertical. Worth checking the timing against Google's update timeline before blaming the tech stack.

weekendhustle

Oh, that's a tough one - I feel for you! 😊 I get that migrations can be delayed for all sorts of reasons, but speed really does make a difference in these situations. Out of curiosity, when did you actually roll out your fixes?

Assuming the basics are sorted (sitemap, technical on-page, no staging being indexed, proper robots.txt, click depth under five, etc.), the biggest win tends to be fixing internal links and overall site structure. A solid migration plan with relevant 301s is key, and don't forget to check incoming backlinks too - they can shift unexpectedly after a move.

One thing worth keeping in mind, the search landscape has shifted a lot in the last couple of years. Most sites I know lost somewhere between a third and half of their traffic thanks to Google AI Overviews and other features shrinking the pie for everyone, especially content-heavy ones like yours. A 70% drop isn't unheard of. So even if you do everything right and climb back up, I wouldn't expect to hit that previous million mark again. Still, every recovery counts 👍

ctrjunkie

that one-hour delay on category pages screams misconfigured caching between Next.js and Cloudflare. When an agency botches a migration on a site that size, digging up the missing 301 redirects usually recovers around 60 to 80 per cent of what you lost. pull your server logs to spot where Googlebot is actually dropping off or failing to render the JavaScript. Getting back to a million visits is going to be tough - search has shifted loads since 2023 - but fixing those technical crawl traps will stabilise your baseline.

chatbox

honestly, that's the exact move I've seen work in cases like this. When you've got a site that size and the 301 chains start looking like a tangled mess from Screaming Frog, you're better off serving a 410 on the old URLs. Search engines interpret that as "gone for good" and drop the old index faster - less confusion than a 301 that might point to a page that doesn't match intent.

If you're on Cloudflare, setting those 410s at the edge via a bulk redirect rule is a no-brainer. Takes minutes, not hours. Pair that with a fresh crawl of the new architecture in GSC and Ahrefs, and you'll see the recovery timeline shrink. just make sure none of the old URLs have external backlinks you still want to preserve - those you 301 individually. otherwise, 410 every last one.

copycat

Let me break this down into three buckets for you - think of it like triaging a patient who's been in a coma since 2023.

1. Technical basics under the hood (Next.js & caching)

Take your homepage plus 10-20 article pages and run full speed diagnostics. You want performance at 80 minimum, and accessibility, best practices and SEO all above 95 - nothing less. Test every page type you have, and focus on mobile because that's where most of your visitors live. Now, how's your cache setup? If a newly published article takes an hour to show up on a category page, something's broken. You need to either re-cache collections on publication or fix your Next.js / Cloudflare cache options. Someone mentioned making the most of SSR - absolutely do that. Also, serve resized images for each breakpoint. One extra trick: create LLM-readable versions of your content, either as markdown files or served directly to AI crawlers.

2. The map and signposts - sitemaps, robots, redirects

Your sitemap must update automatically the second an article goes live - no delays. Run Google's diagnostics to see what's happening with internal links, backlinks, anything broken. Do you have a full list of redirects from the old site? Every old URL needs to point to the right new one. No 404s, no errors, and make sure canonical tags are in place and working.

3. Bring in the smart brain - AI-powered audit

Hook up your backend to something like Claude Code or Codex via API or MCP, depending on your setup. Pull down every page's content, SEO titles, descriptions, URLs, plus all image metadata. Run a scripted diagnosis to catch any mismatches or incorrect redirects. Doesn't matter if you have 100 pages or 10,000 - batch analyse, batch fix. Keep a local copy of your whole site so the AI can read, understand, and update everything.

After that, look at your topical coverage and internal linking structure. Fix titles, categories, the works.

Honestly, the fact this migration has been dragging on since 2023 means there's been a massive screw-up somewhere. If your agency isn't fixing it, let them go and bring in people who can. All of this technical work can and should be done in under a week.