The redesign checklist, and how businesses lose their traffic overnight

A redesign is the riskiest thing a small business does to its website, and almost nobody treats it that way. The old site is dated, the new one looks better, everybody is pleased for about six weeks, and then somebody notices the phone is quieter.

It is not a mystery when it happens, and it is preventable. Here is the order to do it in.

Before you change anything

This is the part that gets skipped, and skipping it is what makes the damage invisible until it is too late to attribute.

Write down where you are now. Traffic for the last twelve months, enquiries per month, and which pages produce them. If you have Search Console, export the queries and pages. If you have nothing, install something and wait a month before starting. A redesign with no baseline cannot be judged, only argued about.

List every URL on the current site. All of them, including the ones you forgot. Your sitemap is the starting point, and a free crawler will find the rest. This list is the single most important document in the project.

Find your best pages. Sort by traffic and by enquiries. There will be two or three pages doing most of the work, and they are frequently not the ones anyone would have guessed. A blog post from four years ago is often carrying more than the homepage.

Note what already ranks. Which terms bring people in, and which pages they land on. These are the pages to be careful with. You are not redesigning them, you are relocating something that is already working.

The four ways redesigns lose traffic

URLs change and nothing redirects. The single biggest cause. Every old address that has no redirect to its new equivalent is a page that was ranking and now returns an error, plus every link anyone ever made to it. This is entirely preventable and it is preventable only before launch.

Content gets cut for looking cluttered. A designer sees a page with nine hundred words of dense text and tidies it into three hundred. Those six hundred words were why the page ranked. Length is not automatically quality, but removing the substance and keeping the layout reliably loses the position.

The new site is slower. It happens more often than you would expect, particularly when the redesign adds video, large photography or a page builder. We measured what that costs on our own previous site in what a heavy homepage costs you, and the numbers were worse than the design brief implied.

The staging site gets indexed, or the live one gets blocked. Both happen. A staging copy that search engines find competes with your real site. A live site that ships with the staging block still switched on disappears entirely, and the symptom is silence.

The checklist for launch week

In order. Most of it is dull, and dull is the point.

Map every old URL to a new one. A spreadsheet with two columns. Where a page is genuinely gone, map it to the closest equivalent rather than to the homepage, because a redirect to the homepage tells a search engine the page is gone and tells a visitor nothing.

Set up the redirects before launch, not after. They should be live the moment the new site is. A week of errors is a week of losing what those pages earned.

Check the block is off. Whatever kept search engines out of the staging site must not travel to production. Load the live site and check the robots file and the page source yourself rather than asking whether somebody did.

Confirm every important page still exists. Take the list of best pages from before and open each one on the new site. Not the equivalent page, the actual URL.

Compare the content on those pages. If the new version is dramatically shorter, decide deliberately rather than by accident whether that is what you want.

Test the forms. Submit every one and confirm the message arrives. A form that silently fails looks identical to a quiet month, and businesses have run for months in that state without knowing.

Measure the speed on a phone. Before and after, on the same page. If the new site is slower, that is a finding to act on rather than a detail.

Submit the new sitemap. Then watch coverage for errors over the following fortnight.

What to expect afterwards

A small dip in the first two to four weeks is normal, even when everything was done properly. Search engines have to recrawl and re-evaluate. Recovery to the previous level usually happens inside six to eight weeks, and a well-executed redesign then goes past it.

What is not normal is a drop that keeps going, or an immediate fall of half or more. That is a redirect problem or an indexing problem, it is fixable, and it is most fixable in the first fortnight. Which is exactly why the baseline from the first section matters: without it, you are guessing whether anything is wrong at all.

When not to redesign

Sometimes the honest answer is that the site does not need replacing.

If the site is reasonably quick, works on a phone, and produces enquiries, and your objection is that it looks dated to you, consider that your customers are not comparing it to this year’s design trends. They are deciding whether you seem competent and whether they can find your phone number.

A refresh of the copy, the photographs and the calls to action often does more for enquiries than a rebuild, at a tenth of the cost and none of the risk. We will say so if that is what we find, which is one of the reasons the first stage of any project is looking at what you already have rather than proposing what to replace it with. The web design service page sets out how that runs, and what a website costs covers the difference in price between the two.

Four questions to ask whoever is doing it

Ask these before you sign. The answers tell you more than a portfolio does.

“Who is writing the redirect map, and when do I see it?” Somebody has to own this by name. If the answer is vague, or if it is something that happens “at launch”, that is where the traffic goes. A good answer is a person and a date before launch week.

“What will you measure before we start?” If nothing is measured beforehand, nobody can tell afterwards whether the redesign worked, and every conversation about it becomes a matter of opinion. A supplier who wants a baseline is a supplier who expects to be judged on results.

“Which of my current pages are you keeping as they are?” The right answer names specific pages and a reason. An answer of “we’re rebuilding everything” means the pages currently earning your enquiries are being rewritten by somebody who has not yet checked which ones those are.

“What happens if traffic drops after launch?” You want to hear a diagnosis process and a window, not reassurance. Drops are normal for a fortnight and recoverable when caught early. A supplier who says it will not happen either has not done many of these or is not planning to be around.

None of these are gotcha questions. Anyone who does redesigns properly will have ready answers and will be glad you asked, because these are the same four things that go wrong on projects where the client never asked.

The short version

Measure before you touch anything, list every URL, know which pages are actually carrying the business, and have the redirects live the moment the new site is. Redesigns do not lose traffic because the design changed. They lose it because the addresses changed and nobody wrote down where everything used to be.