Why traffic drops after a redesign, and how to avoid it

Almost every drop comes from four things

Changed URLs with no redirects. Content trimmed in the name of a cleaner look. A no-index setting left on from the staging site. Tracking that never got reinstalled, which does not cost you traffic but does mean you cannot tell what happened.

Design is almost never the cause. We have seen sites get much better looking and lose most of their search visits inside two weeks, and the reason has always been on that list. Handle those four and a redesign is a low-risk project.

Build the redirect map before you build the site

Not at the end. Before. Crawl the existing site to get every URL, then pull your top pages from Search Console for the last twelve months and mark the ones producing clicks. That is your protected list.

Every old URL needs a permanent redirect, a 301, to the closest equivalent page on the new site. Closest equivalent is the operative phrase. Sending forty retired pages to the homepage is treated as a soft error rather than a real match, and the value evaporates. If there is no equivalent page, decide whether that page should have survived at all.

Check the redirects again after launch, not before. Staging redirects and live redirects fail in different ways, and a chain of three hops loses more than one clean hop does.

Do not throw away the words

The most common self-inflicted wound in a redesign is a service page going from nine hundred words that answered real questions to a hundred and twenty words under a nice photograph. It looks better. It ranks for nothing.

Before the rewrite, note which pages bring in search visits and what they say. Those pages can be shortened and made clearer. They cannot be emptied. If the new design has no room for the text, the design needs an accordion, a second section, or a rethink.

Internal links matter here as well. A redesign that reduces the navigation from twenty items to six often orphans the pages that were dropped from the menu, and a page nothing links to gets crawled less and ranks worse even when the URL still works. Keep a route to every page you kept, from the footer, a service index, or the body copy of a related page.

Watch the headings too. An H1 that gets replaced with an image of the same words disappears as far as search is concerned, and so does a page title that gets rewritten to lead with your company name instead of the service.

The launch-day switches people forget

In WordPress, the checkbox under Settings then Reading that discourages search engines is on for most staging sites and is forgotten on a meaningful share of launches. In other platforms it is a site-wide no-index toggle with a different name. The result is the same: your new site quietly asks Google to leave.

Also check robots.txt for a blanket disallow line, confirm the canonical tags point at the live domain rather than staging, and load the site in a private window on a phone to make sure nothing is sitting behind the password you set months ago.

Tracking does not come along by itself

A new build usually means a new template, and analytics tags live in templates. Reinstall the measurement tag, rebuild your conversion events for form submissions and phone taps, and test both by actually submitting the form and tapping the number.

Two things go wrong here quietly. Form notifications get sent to an address nobody has read since 2019, and tracking phone numbers get replaced with the main line during the rebuild, which erases your ability to tell which channel produced a call. Check both in the first week, because a month of missing leads is worse than a month of missing rankings.

What a normal dip looks like

Even a clean migration usually wobbles. Rankings move around for a couple of weeks while the new pages get crawled and reassessed, and it is normal to see a modest dip that recovers within a month or so.

What is not normal is a steep drop that keeps going into week four, or a Search Console pages report showing hundreds of URLs newly excluded. That is a redirect or an indexing problem, not patience, and the earlier you look the cheaper it is to fix. Keep a screenshot of your traffic and top pages from the month before launch so you have something to compare against.

What is worth losing

Not all traffic is worth protecting. Old blog posts about topics you no longer sell, pages built for a service line you dropped, thin town pages written by a previous agency: consolidating those into fewer, better pages is a real improvement even though the visit count goes down.

The distinction is whether the traffic was ever going to call you. Decide that deliberately, page by page, before launch. Losing visits on purpose is a strategy. Losing them by accident is the thing this whole checklist exists to prevent, and it is why we usually spend the first week of a rebuild on the old site rather than the new one.