Relaunching a Website Without Losing What You've Built

Posted by Sean on March 14, 2017

We launched a new website this month. New platform, new design, the lot. For most of the business it’s an exciting day; for the person responsible for search and analytics, it’s the day you can undo three years of work in an afternoon if you’re not careful.

A replatform is about the most dangerous thing you can do to organic traffic, because so much of what you’ve built is tied to the specifics of the old site: the URLs, the redirects, the internal links, the tracking that’s been patched and extended over years. Change the foundations carelessly and it all comes loose at once. The worst of it is that most of the damage is invisible on launch day and only surfaces weeks later in your rankings.

So most of the work happened before anyone saw the new site. The things that mattered, roughly in order:

  • Map every old URL to its new home and 301 it. Not just the important pages, every page; the long tail is where the losses hide.
  • Keep the URL structure wherever there’s no real reason to change it. “Tidier” URLs are rarely worth the risk.
  • Crawl the old site and the new one and compare, so nothing drops off the map unnoticed. Screaming Frog earned its licence fee this month.
  • Rebuild the tracking, then check it again after launch, because analytics breaks without a sound; a dead tag reports that nothing is happening rather than that it has stopped recording.
  • Watch rankings and traffic daily for the first few weeks, so that if something has slipped you catch it in days rather than at the next quarterly review.

I’d love to say it all went perfectly. It didn’t; something always breaks, and a couple of things did. The point of the preparation isn’t to be flawless, it’s to keep the breakages small and to spot them fast, before they become a month of lost traffic nobody noticed.

Whatever it feels like on the day, go-live is the start of the risky bit rather than the end of it. The fortnight after, watching the numbers and fixing whatever the plan missed, is when you actually find out how the migration went. We’re in that fortnight now, coffee in hand, watching the graphs.