A new website is a chance to fix a lot of things at once. It is also a chance to break things that were quietly working: old links that customers bookmarked, pages that search engines trusted, and basic accessibility that nobody thought to check again.

This is the checklist we use when rebuilding a business website.

We rebuilt idlefusion.com in 2026, and we did not get everything right the first time. Where it helps, we note the real mistakes we found and fixed on our own site.

Before you build: know what you have

List every URL on the current site

Start with an inventory. Export the old sitemap, crawl the current site, and check your analytics and search console for pages that receive visits or links. Include old blog posts, PDFs, and campaign landing pages; other sites often still link to them.

Decide what happens to each one

Every old URL should end up in one of three states:

  • Kept at the same address.
  • Moved to a new address, with a redirect.
  • Retired on purpose, returning a real “not found” or “gone” response.

Write this down as a simple spreadsheet: old URL, new URL, and status. It becomes your redirect map and your test plan.

URLs and redirects

Use permanent redirects

When a page moves, redirect the old URL to the new one with a permanent redirect: a 301 or a 308. Permanent redirects tell search engines the move is final, so they can transfer what they know about the old page to the new one.

Temporary redirects (302 and 307) say the opposite: the old page will be back. Some frameworks and hosting platforms use temporary redirects by default. On our own rebuild, several redirects were going out as 307s. We changed them to 308s.

Redirect straight to the final page

Each old URL should go directly to its final destination in one step. Redirect chains slow visitors down and add room for errors.

Pick one host and redirect the other

Your site should live at one address, either with “www” or without it. The other should redirect permanently to it. If both serve the full site, search engines see two copies of every page.

Our www address was serving a duplicate of the whole site. It now returns a 301 redirect to the apex domain, idlefusion.com.

Choose a trailing slash rule and keep it

Decide whether page URLs end with a slash (/services/) or without one (/services). Then use that form everywhere: in links, in canonical tags, in the sitemap, and in redirects. The other form should redirect to the chosen one.

Canonicals and indexing

One canonical URL per page

Each page should declare one canonical URL that points to itself, in its final form. The canonical should never point at an address that redirects.

This was one of our own mistakes. Our pages used trailing-slash URLs, but the canonical tags dropped the slash. That meant every canonical pointed at an address that redirected somewhere else. The fix was to make each canonical match the real, final URL exactly.

The sitemap lists only indexable pages

Your XML sitemap should list the pages you want in search results, and nothing else. Every URL in it should return a 200 status, be the canonical version, and be free of “noindex.” Leave out redirects, error pages, and duplicates.

Use noindex for unfinished pages

If a page is live but not ready, such as a draft, a placeholder, or a staging copy, add a “noindex” robots meta tag. Remember that search engines must be able to crawl a page to see that tag, so do not also block the same page in robots.txt. Remove the noindex when the page is ready, and add it to the sitemap then.

On each page

A unique title, about 60 characters or fewer

Every page needs its own title that describes what is on it. Search results often cut off long titles, so keep them to roughly 60 characters, including your brand name if you add it.

A unique description, around 150 to 160 characters

The meta description does not directly decide rankings, but it often becomes the snippet people read in search results. Write one per page, in plain language, and keep it near 150 to 160 characters so it is less likely to be cut off.

One H1 per page

Give each page a single main heading that matches its purpose. Use H2 and H3 headings below it in order. This helps screen reader users navigate, and it gives search engines a clear outline.

Structured data

Add structured data in JSON-LD to describe your site plainly:

  • Organization for your business name, logo, and contact details.
  • WebSite for the site itself.
  • BreadcrumbList on pages below the home page, matching the visible breadcrumbs.

Validate it with a structured data testing tool after launch. Only describe what is actually on the page.

Accessibility

Color contrast that meets WCAG AA

Normal text needs a contrast ratio of at least 4.5:1 against its background. Large text needs at least 3:1. Check every color pair in the design, including links, buttons, form placeholders, and text over images.

On our own site, a code-heavy page used syntax colors that fell below 4.5:1. Low contrast is easy to miss when you are the one who chose the colors. We adjusted the colors until every pair passed.

Full keyboard access

Unplug the mouse and try the site. You should be able to reach every link, button, menu, and form field with the Tab key, see clearly where focus is, and operate everything with Enter or Space. Add a “skip to content” link so keyboard users can bypass the navigation.

The basics that are easy to drop

Check alt text on meaningful images, labels on every form field, and a page language set in the HTML. A rebuild often replaces components, and these details can disappear in the process.

Performance

Size images for how they are shown

Serve images at roughly the size they appear, in a modern format, with width and height set so the page does not jump while loading. Load images below the fold lazily.

Cache sensibly

Files whose names change when their contents change, such as fingerprinted CSS and JavaScript, can be cached for a long time. HTML pages should be cached briefly or revalidated, so visitors see updates promptly.

Errors

A real 404 status

When a page does not exist, the server must return a 404 status code, not just a page that says “not found.” If a missing page returns 200, search engines may treat every broken link as a real, empty page.

Our custom 404 page looked correct but returned a 200 status. We fixed it to return a true 404. The page still helps visitors find their way; it just tells browsers and crawlers the truth.

After launch: check it all again

Launch is when most of these problems become visible. Plan time for a full review in the first days:

  • Test every row of your redirect map and confirm each returns one permanent redirect to a page that returns 200.
  • Check the www and non-www addresses, and both trailing slash forms.
  • Crawl the new site and look for broken links, redirect chains, missing titles, and duplicate descriptions.
  • Confirm canonicals match final URLs.
  • Submit the new sitemap in your search console and watch its coverage reports for errors over the following weeks.
  • Request a made-up URL and confirm it returns a 404.
  • Run an accessibility check and a keyboard pass on your main templates.

Then check again a few weeks later, once search engines have recrawled the site.

You can see how we approach website projects on our website development services page, and an example of web and mobile work for a mission-driven organization in the Everybody Everywhere case study.

Talk it through

If you are planning a rebuild and want it done without losing what already works, tell us about your website project. Matthew reads every message himself and usually replies within one business day.

Plan your website More on website development