Can you redesign your website in phases? Usually, yes, but a phased website redesign only works when the right part is phased. The parts that hold the site together, meaning the navigation, the page templates, the design and the URLs, should change across the whole site on one launch day. What can follow is depth: rewritten copy on the long tail of pages, project write-ups, real photography, spec sheets, new industry pages. A site that launches with the new frame everywhere and thinner content on some pages is a normal first phase. A site where half the pages still sit in the old design behind the old menu isn't a phase. It's two websites sharing a domain.
Why does the question come up so often?
When I hear it on calls, it usually isn't impatience. It comes from one of a few real constraints. The budget covers part of the scope this year and not all of it. The content isn't ready: the projects were never written up, the good photos don't exist yet, the spec sheets live in a spreadsheet. There's a hard date, like a rename, that the full site can't make. Or the catalog runs to hundreds of products, and the thought of redesigning every one of them makes the project feel endless.
All four are good reasons to phase something. None is a reason to phase everything.
What has to change on the same day?
Anything a visitor crosses on the way from one page to another.
The navigation, header and footer. If the new homepage menu points at sections that still open in the old design, the buyer clicks once and seems to land on a different company, and starts wondering which version is current.
The templates. Designing the homepage now and the inner pages "later" usually means designing the homepage twice, because the components it's built from are the same ones the product and service templates need. The templates are the design.
The URL structure, and the redirect map that goes with it. Google's own guidance on site moves with URL changes is direct about it. For small and medium sites, it says "We recommend moving all URLs on your site simultaneously instead of moving one section at a time," because that helps Google detect the move and update its index faster. Google doesn't put a page count on that, but most company sites without a very large catalog sit at that end.
Changing URLs in two rounds also means doing the redirects twice. Say phase one moves an old product page to a new address, and phase two reorganizes the catalog again. The original address now has to be re-pointed, or it chains through the phase-one address to the final one. The same Google page advises redirecting "to the final destination directly" and, where that isn't possible, keeping a chain to "ideally no more than 3" hops. That's a second migration, with a second round of the ranking fluctuation Google says can happen while it recrawls. The website migration checklist covers what a migration done properly involves. The point here is to do it once.
What can sensibly come later?
Anything that deepens a page that already exists at its final address, in the final design.
- Copy on the long tail. Rewrite the pages that carry the business first: the homepage, the main service or product families, the quote path, the about page. Move the rest onto the new templates with their current content intact, then rewrite them in order of what earns traffic or inquiries.
- Proof. If the photos of past work are thin, launch with the best you have and replace them as the portfolio builds over the first year. The same goes for case studies that haven't been written yet. A project page template that's ready on day one turns each new write-up into a content task, not a design task.
- Product depth. Spec sheets, drawings, application notes. These add to a product page without moving it.
- New pages. An industry or service page added after launch is new content, not a move, so it doesn't touch the redirect map.
That's also how a large site gets redesigned without designing every page. A hundred interior pages are rarely a hundred designs. They're a handful of templates, and once the product and service templates exist, most of the site moves onto them together.
What about launching with only our top products?
That depends on what happens to the other product pages on launch day. If the plan is to leave them off and add them later, push back. A product page that disappears takes its search traffic with it, and part numbers and model names can keep getting searched for years. Google's documentation on HTTP status codes says URLs that are already indexed and then return a 4xx error "are removed from the index." A page that errors for the three months before phase two is a page you'll be rebuilding search visibility for.
The better version: every existing product page moves onto the new product template at its final URL on launch day, with the content it has now. The top products get the full rewrite. The rest get enriched later, a product family at a time. How big that move is depends on where the product data lives. If the product pages already come from a database or a structured export, moving all of them onto one template is a small job. If each page was built by hand, getting the product data into a structured form is the real work of phase one, so scope and budget it first. Products you're actually retiring get decided now, using the same cases as discontinued product pages, rather than dropped by omission.
Isn't that just a full redesign with weaker content?
That's the strongest objection, and it's half right. Phase one costs most of what the structural build costs. What phasing saves on is the writing, the photography and the case studies, not the build. Yes, the frame ships whole, and on day one some pages read no better than they did last month. But they read the same as they do today. Nothing gets worse, the URLs are final, and every later phase is writing rather than rebuilding.
The alternative looks cheaper and usually isn't. A new homepage and twenty new pages sitting beside two hundred old ones makes the old ones look worse by contrast, leaves two designs to maintain, and keeps the hardest part, the structural move, still ahead of you. If the budget can't cover the frame at all, the honest smaller project is usually a refresh, or a rewrite of the key pages on the current site at their current URLs, with the redesign planned whole for later. Website redesign vs rebuild walks through those sizes of project.
How do you tell if a website redesign in phases will work?
Three questions settle most plans.
- Will any URL change in phase two? If yes, it isn't a phased redesign. It's two migrations. Settle the structure first so every later phase lands on addresses that already exist.
- After phase one, can a visitor click from any page to any other without crossing between two designs or two menus? If not, the phase line is in the wrong place.
- Is the thing being put off depth or structure? Copy, photos, case studies and specs are depth, and depth phases well. Navigation, templates, URLs and the platform are structure, and structure goes together.
The one real exception is size. Google says you "can choose to move larger sites one section at a time," which makes problems easier to spot and fix. That fits a catalog of many thousands of items in its own section of the site. It doesn't fit a forty-page company site nearly as well. Even then, the header, footer and menus are best switched site-wide on the first day.
A plan that passes all three is sound, and its first phase is often shorter than people expect. For how long that phase runs on the calendar, see the website redesign timeline.
