A Redesign Is a Migration in Disguise

Everyone plans a redesign as a design project and then discovers halfway through that it is a data migration with a new skin on top. The design is the visible part and rarely the risky part. The risk is that the site currently earns something — rankings, traffic, a page that quietly produces half your enquiries — and a redesign can throw that away in a single afternoon without anyone noticing for six weeks. The sequence below exists to stop that.

Step 1 — Write Down What You Have Now

Before anyone opens a design tool, record the baseline. Without it you will not be able to tell whether the redesign helped, and every later argument becomes a matter of taste.

• Monthly visitors, and the same figure for your top twenty pages
• Enquiries per month, and which pages they came from
• Which search queries you currently rank for, and on which page
• Speed scores on mobile for the homepage and one service page
• A full crawl of every existing URL — export it, you will need it in step 4

Step 2 — Inventory the Content: Keep, Merge, Kill

Put every page in a spreadsheet with its traffic and its enquiries, then mark each one. Keep what earns. Merge near-duplicates into one stronger page. Kill what has produced nothing in a year and never will. Most outdated sites shrink by a third at this step, and that is a good outcome — thin pages dilute the ones you want found.

Step 3 — Fix the Structure Before the Design

• One page per service, named the way customers describe the problem, not the way you describe your offer
• Real case studies: the situation, what changed, what it produced
• Navigation that fits in one line and does not need a dropdown to explain itself
• A decided page-type list — home, service, case, article, contact — because that list is what actually gets designed
• Then, and only then, the visual design. Designing before the structure is settled produces beautiful pages for content that does not exist

Step 4 — The URL Map. This Is the Step People Skip.

Take the crawl from step 1 and give every old URL a destination on the new site. Every single one. Pages you deleted point to the closest relevant page, not to the homepage — a homepage redirect tells a search engine the content is gone and hands back whatever that page had earned. This becomes your redirect list, and it needs to be live from the first minute the new site is. Not the next day.

Step 5 — Build on a Staging Site, Closed to Search

• Build somewhere private, blocked from indexing — and write yourself a note to unblock it at launch, because forgetting is one of the most common launch failures there is
• Move the content properly, checking formatting and images rather than trusting an automatic import
• Reconnect analytics, goals, ad pixels and CRM before launch, not after
• Test every form by actually submitting it and confirming the email arrives
• Check on a real phone, not a resized browser window

Step 6 — Launch Day, in Order

• Take a full backup of the old site and keep it. You may need to compare something in three weeks
• Deploy with the redirects already in place
• Remove the indexing block. Verify it is actually removed by checking the live site
• Submit the new sitemap and check that key pages return the right status codes
• Submit a test enquiry from a phone on mobile data and confirm delivery
• Launch early in the week and early in the day. Never on a Friday afternoon

Step 7 — The First Thirty Days

Expect a dip. Search engines need to recrawl and reassess, and two to four weeks of softer numbers is normal rather than a sign of failure. What is not normal is a dip that keeps deepening past a month. Watch the 404 report weekly and add redirects for anything real that appears. Compare against your step 1 baseline at thirty days, then again at ninety. If traffic has not recovered by then, the cause is almost always a redirect that was missed or a page whose content was cut during the rewrite.

Conclusion

A redesign done in this order is unremarkable: the site looks current, the numbers hold, and nothing dramatic happens. That is the goal. Drama in a redesign is nearly always something going wrong. If you are choosing who does the work, ask them to walk you through steps 4 and 6 before you talk about visuals. Someone who has done this properly will have opinions about redirects. Someone who has not will want to show you moodboards.