Website Redesign Cost, Timing, and Process [2026]

website redesign — what it means

A website redesign is the process of rebuilding an existing site by overhauling its design, structure and technical platform. Success or failure hinges on a migration design that does not lose the assets you already hold — your search rankings and your accumulated content.

Hi, I’m Masato, a Webflow specialist. I run a web production company called Webharu in Shiogama, Miyagi, Japan.

“We should redesign our website soon, but I have no idea what it costs or when we ought to do it.”
“If we rebuild it badly and lose the search rankings and traffic we have now, that’s a real problem.”

In redesign consultations, these two worries always arrive as a pair. The longer a site has been running, the more assets it holds — accumulated articles and traffic — and the heavier the decision to rebuild becomes. We often hear from people who left their site alone for years simply because they did not know where to start.

This article is for owners and web managers with exactly that problem. It covers everything from market rates to a migration process that does not fail.

Let me start with the conclusion.

Treat the cost of a website redesign as roughly the same market rate as building a new site, plus the work of migrating your existing content. Now that AI-assisted production has reached a practical stage, the barrier to rebuilding itself is far lower than it used to be.

But what matters more than cost or schedule is the process that carries over the assets your current site holds — search rankings, traffic and content — without losing them. Treat that lightly and you can end up watching traffic fall from the day the new site goes live. This article covers that part too, in concrete terms.

The six signs that it is time to redesign

Here are six signs that it is “about time.” One on its own may be worth watching, but if several apply, it is time to consider a redesign seriously.

• It is hard to read on a phone: text too small, buttons hard to tap, layouts breaking. Most traffic today is mobile, so this alone is lost opportunity. Even if it looks fine on a desktop, be careful if you have never viewed your own site on an actual phone. If the first few seconds feel hard to read, people leave before they read a word of your copy.
• It takes more than three seconds to load: Loading speed affects both usability and search evaluation. The cause is usually heavy images, an old theme, or a build-up of unused plugins. If it feels slow, it is worth actually measuring it. Even just checking your score with a free measurement tool tells you where you stand.
• You cannot update it yourself and need the production company for every fix: If every small wording change or news post requires a quote and a lead time, the bar for updating is too high. The result is that you stop updating at all, which leads straight into the next sign.
• The information has been frozen for years: Prices, project examples and staff profiles left exactly as they were several years ago make visitors wonder whether the company is still active. Outdated pricing or project examples in particular push people toward competitors while they are comparing options before an inquiry. The longer you put off updating, the more the list of things to fix snowballs.
• Almost no inquiries come in: If you have traffic but no response, suspect the path to contact — where to find what, and how to reach you — before you suspect dated design. Simply having your phone number or a price guide in a hard-to-find place costs you opportunities. Refreshing the design alone will not move the numbers if the path is still broken.
• You are paying a maintenance fee and nothing else: Paying a monthly maintenance fee when no work is being done is a sign to review. If you ask what you are paying for and get only a vague answer, the redesign is a good moment to review the contract itself. The technical platform can change the structure of maintenance costs entirely, so we have also laid out how to choose on our Webflow vs. WordPress comparison page.

You do not need all six to apply. Two in particular — “you cannot update it yourself” and “no inquiries come in” — should be treated as higher priority than the rest.

Cost and timing are usually the next questions, so let me take them in order.

The best and worst times to redesign

The conclusion on timing first: avoid launching right before the event you most want to use the site for, and start moving at least two to three months ahead of it.

• Avoid your busy season and the run-up to important events: Launching right before a period when you want to use the site heavily — a hiring push, a trade show, a campaign — is best avoided. Right after launch you are most likely to spot layout breaks and missed paths to contact, and you need time to fix them.
• Working backwards, start two to three months before you need it: Count the production period itself plus a calm window after launch to check and adjust, and work backwards from there. That keeps you safe.
• Fiscal year-end and the start of a new business year work well as a trigger: Many companies move in step with the timing that makes budgeting easiest. It is not a requirement, though — if several of the signs in the previous section apply, there is little reason to wait.

What a website redesign costs

Aspect	New build	Redesign
Cost	The market rate	The market rate, plus the migration work
Schedule	The market rate	The market rate, easily extended by the volume of migration
Biggest risk	Delays in progress, such as waiting on copy	Gaps in carrying over search rankings and data
Most important work	Designing the structure from zero	Sorting what to keep and what to change

First, the baseline. Checking several sources, including the production-company matching service Web Kanji, the market rate for website production in Japan runs roughly ¥50,000–150,000 for a template-based site, ¥200,000–1,000,000 for a standard small-business site, ¥1,000,000–1,000,000 with acquisition and SEO design included, and from ¥1,000,000 for e-commerce or booking integrations. A redesign starts from that same baseline, with the work of “how much of the current site gets migrated” added on top.

People often assume it should be cheaper because you are not starting from zero, but in practice it frequently goes the other way. A redesign adds work that a new build never has:

• Taking inventory of how many pages and articles the current site has
• Sorting what to migrate and what to cut for good
• Building the old-URL to new-URL mapping table and setting up the redirects
• Cross-checking after migration that no information is missing

That is why the same page count takes more effort than a clean new build. In particular, the more a site holds — dozens of articles and news posts, or well over a hundred — the more the inventory and cross-checking hours pile up. It is not something a simple page count can measure.

Our own prices, as of August 2026, are as follows (excl. tax, lower bound shown).

• AI Website Production (standard configuration, in as little as 1–2 weeks): from ¥1,000,000
• Full-Scratch Website Build (original design, designed from zero): from ¥1,000,000
• Web Application Development (membership, booking, payments, admin screens): from ¥2,500,000

The figures shown are lower-bound guides. What moves them is the page count, multilingual support, the volume of existing content to migrate, the number of departments and approval steps involved, and security requirements. Once a quote is confirmed, we do not change it after work begins.

What goes into a redesign from ¥1,000,000 | where the hours actually go

The main arena for redesigns is the Full-Scratch Website Build (from ¥1,000,000). Whether ¥1,000,000 feels expensive or reasonable depends entirely on whether you can see what is inside, so let me open all of it up first.

• Strategy and requirements definition session (3 hours) — The place a redesign goes wrong most often is not implementation; it is skipping this step. We start with the inventory of what to keep and what to drop
• Original design in Figma — No templates. The hours of a person designing from zero are almost the entire price gap with AI Website Production
• Implementation in structured CSS (BEM), responsive layout, loading-speed optimization
• SEO technical design — Structured data, metadata, and the design for carrying the sitemap across. This is the implementation side of the “migration without losing rankings” covered later in this article
• Blog CMS design, contact form, launch, DNS setup
• Style guide plus an AI update manual — We hand the site over in a state where either you or another company can take over running it after delivery. The full code is delivered, so the site keeps working even if you part ways with us
• One month of post-launch support

What moves the figure from there is the page count, multilingual support, the volume of existing content to migrate, the number of departments and approval steps involved, and security requirements (NDAs, security checklists). Put the other way: a standard corporate site redesign that does not hit those variables lands around ¥1,000,000. Quotes come itemized so you can see what each variable costs, and once a quote is confirmed we never change it after work begins.

The ¥1,000,000 AI Website Production tier includes one 60-minute discovery session, the full site build, high-speed loading tuning, the groundwork for the AI era (structured data, llms.txt), basic SEO (metadata, sitemap.xml, robots.txt), GA4 and Search Console measurement setup, one contact form, DNS setup at launch, and one week of post-launch bug fixing (up to two minor revisions).

What it does not include is a CMS, multilingual support, or fine-tuning of an original design. Those are discussed as options if you need them. Some production companies charge separately for parts of the list above, so at the quote stage make sure you ask what is included and what is billed separately.

On top of that sits a migration option for the existing site at +¥80,000. As a sense of scale, that is close to the 62 articles and 24 project entries we actually migrated in our own site redesign. Other guide prices: adding a CMS is +¥50,000, multilingual support is +¥150,000, on-site photography is +¥70,000, and API integration with external systems is +¥200,000. The longer a site has been running, the more articles and project entries it has and the more the migration volume changes, so get the overall picture on our pricing page and then ask us in a free consultation for an accurate estimate.

If you want an original design built from zero in Figma rather than fast AI production, the Full-Scratch Website Build (from ¥1,000,000) is a separate option. Delivery can be the full code (Astro / Next.js) or Webflow, and the schedule is correspondingly longer: about 1.5 months for around 10 pages, about 2 months for 20 pages. If your redesign is about building the brand carefully, it belongs on your shortlist (see the details of our Webflow production here).

Migrating without losing rankings or assets | five things to ask your production company

The scariest thing about a redesign is search rankings and traffic dropping the moment you go live.

To a search engine, a page whose URL has changed is basically a different page. Without a design that carries the previous evaluation across, rankings you built up over years can be treated as close to a reset. When 301 redirects are set up correctly, it is generally said to take anywhere from a few days to a few weeks for search engines to recognize the new URLs.

This part gets technical, so we have put the items to ask your production company into a table. You do not need to understand all of it yourself. Being able to ask “how is this handled?” is enough.

What to check	What to ask	Why it matters
URL design and 301 redirects	Whether permanent page-by-page redirects (301) are set from the old URLs to the new ones	Without them, both search engines and visitors get stranded on the old URLs, and the evaluation is not carried over
Migration of content assets	Whether every article, project entry and news post you have accumulated is migrated to the new site without omission	The more items there are, the easier it is to miss some by hand, so a mechanical check is needed
Carrying over meta information	Whether each page’s title, description and so on are designed with how they currently appear in search results in mind	Change this and your click-through rate in search results moves with it
Carrying over measurement tags	Whether Google Search Console, GA4 and the rest of your existing measurement setup are carried over to the new site	If measurement breaks, you can no longer compare before and after the redesign
Post-launch index checks	Whether there is a routine for confirming after launch that the new pages are correctly recognized by search engines	Left alone, some pages can stay missing from search results for a long time

The first and fourth rows matter most. If 301 redirects are missed, it is not only search engines: people arriving from bookmarks or from links on other sites end up stranded on the old pages. And Search Console and GA4 should be carried over from the existing properties rather than rebuilt. Create them fresh and you lose the ability to compare against pre-redesign data, which means you can never verify afterwards what the redesign actually changed. One more thing about URLs: if you change the domain itself, the carry-over effect drops even with 301 redirects in place. Unless you have a specific reason, keep the domain as it is.

Simply confirming these five items before you sign prevents most of the “we had it rebuilt and the rankings dropped” trouble that surfaces afterwards.

What to have ready before work starts

Having the following on hand before you approach a production company makes both the quote and the migration go smoothly.

• Login details for your current site’s admin panel (CMS)
• Administrator details for your domain (where the contract is held)
• Access permissions for GA4 and Search Console
• A list of the articles, project entries and news posts you want to migrate

Domain administrator details in particular tend to be forgotten at redesign time. We have had people who could not work out which company held the contract and lost days just searching for it. GA4 and Search Console access is the same: if it cannot be carried over, you lose the ability to line up the numbers before and after the redesign at all. Migration itself can be done in several ways — re-entering by hand, bulk import via CSV, or using the export function of the old CMS — and which one suits you changes with the volume of articles and data.

A full record of the redesign we ran on our own site

From here it is my own experience. In July 2026 we fully redesigned our own site with AI (Claude Code). It took seven days: four days of planning plus three days of building. We built 26 new pages, migrated 62 articles and 24 project entries from the old Webflow site, connected four forms including the contact form, and set up GA4, Search Console, structured data, the sitemap and llms.txt.

I worked alone, and the AI did the hands-on work. That said, we produced around 50 design variations and narrowed them down. To be honest, it took about a month of mass-producing AI garbage before the design came out right. Most of those four planning days went into that design trial and error.

My role was to decide the concept and structure, judge the design aesthetically, fact-check the copy and make the commercial calls, and review and give the go-ahead at each step. The AI does the hands-on work; the human makes the calls. The AI subscription cost about ¥30,000 a month, and the models I used were Opus and Fable in Claude Code.

We chose Astro as the technical platform, for its compatibility with Cloudflare, its loading speed and its SEO performance. As a result, running costs fell from just under ¥100,000 a year for the old Webflow subscription to ¥1,500 a year for the domain alone. Measured Lighthouse mobile scores are 98 for the top page and 97 for article pages (the first measurement was 73, and we improved it by cutting fonts). Current Search Console figures show 190,000 impressions and over 5,700 clicks, and we appear on the first page of search results even for the single word “webflow.” Being able to rebuild while keeping 62 articles of accumulated content is, I believe, the foundation under those numbers.

That experience convinced me that on client redesigns too, the precision of the inventory and the checks before migration feeds directly into the numbers after launch. The checklist in the previous section is that conviction turned into items. I wrote up the whole redesign in the full record of our seven-day redesign.

Being honest: we caused two incidents in this redesign

I have written nothing but good numbers so far, but the same redesign also produced two incidents. Nobody noticed either of them for more than two weeks after launch. For anyone considering a redesign, I think these are more useful than the success story, so here they are.

Let me be clear about where I stand first. As I wrote above, this site was built with the AI doing the hands-on work and me making the calls. Both of the incidents below were, as work, carried out by the AI. But failing to catch them was my responsibility. My role is to give the instructions, review, and approve, so the moment I let them through, the fault is mine. That holds whether or not you use AI — the same is true when you outsource to people.

Incident 1: 57 Korean pages vanished from search results

The old site had a Korean version. Some articles were read in Korea before they were read in Japan: one article that sat on the first page of Korean search results for the single word “webflow” (average position 5.6) had 11,663 impressions in seven weeks.

When we built the new site, the instruction I gave the AI was “we will not rebuild the Korean version for now.” In response, the AI wrote a single line of configuration saying “redirect everything under /ko/ to the Japanese version.” Faithful to the instruction. But it did not think as far as what that one line meant: throwing away the search rankings we had built up in Korea. And I let that line pass in review. That is where the incident starts.

We launched on July 16. The pages stayed in search results for a few days, but from July 20, impressions for the Korean pages were effectively zero. Google read the redirect, concluded the pages had been merged into the Japanese version, and dropped them from the index. On July 26 we rebuilt the Korean version and removed the redirect, and still they did not come back. When we looked into it, Google had simply not come back to crawl. Once a URL is judged to be a redirect, its re-crawl priority drops.

I noticed something was wrong on August 1, sixteen days after launch. What triggered it was looking at the numbers and catching that Korean traffic had gone to zero. From there I had the AI investigate, and the cause and the date it started came out immediately. Put the other way: if I had not caught it, the AI would never have reported it. The site displayed perfectly normally, so there was no reason anywhere to suspect it.

Incident 2: for two weeks from launch day, analytics was not running

This one was worse. From launch day on July 16 until July 30, GA4 showed zero traffic. To be precise, a handful of records existed, but for two weeks not a single visitor was counted.

Two causes overlapped. One was that the Google-side file that loads the measurement tag was returning a 404: the measurement ID itself was installed correctly, but the tag itself was not being served. The other was that a security setting (CSP) the AI had configured was blocking the destination the measurement data is sent to. Either one alone might have been caught, but with both stacked it took time to separate them. Here too, it was the AI that tracked down the cause and fixed it. All I did was say that the traffic looked strange.

Here again, the site displays completely normally. No errors appear. Until you open the dashboard and see “0 visits today,” nothing looks wrong at all. The result is that the two weeks of data you most want — the two weeks right after a redesign — do not exist anywhere in the world. We restored it between July 31 and August 1, and we now measure around 20 visitors a day. But when we try to verify how the numbers changed after the redesign, there is nothing to compare against. It was our own site, so we can still laugh about it. On a client project, it would have been unrecoverable.

What the two incidents have in common: the site looks fine

In both cases, open the page and it displays normally. The server returns 200 (OK). Nothing appears in the error logs. What was broken was only “how it looks to Google” and “the back end of measurement” — nothing surfaced within the range a human eye can inspect. That is why two weeks went by unnoticed.

The reason we put “post-launch index checks” on the checklist earlier is this experience. If you check by eye on launch day and call it done, this kind of incident sails straight past.

So we had the AI build the act of noticing itself. A mechanism that watches the live site automatically. Whether the main URLs are returning real content (rather than a redirect or a 404), whether pages we want in search have an “exclude from search” directive attached, whether the measurement tag itself is alive, and whether data is actually arriving in analytics — it checks these four things twice a day, automatically. If anything is abnormal, a notification comes to me. The last one, whether data is arriving, is the one I rely on most, because whatever the cause, if measurement is dead it always trips. After building it, we fed in the dates of the July incidents to test it, and it correctly flagged them as abnormal.

This whole sequence sums up how I work with AI right now. The AI is faster at the work. The AI is faster at finding causes. But noticing that something is off is, for now, still a human job. And human eyes have limits — as here, where nothing on screen was broken at all. So the noticing had to be built into the system too.

The scariest thing in a redesign is not breaking loudly. It is time passing while you fail to notice that something is broken. When you choose a production company, ask them what they check after launch and how. A company with a checking procedure is safer than one that looks at the screen and calls it done. More companies are building with AI now, but what you should be looking at is not that — it is who watches the site afterwards, and how.

The patterns that make redesigns fail

Here are three failure patterns we see often in consultations.

• Changing only the design and leaving the content as it was: The look is new, but the copy, the structure and the path to contact are all as before. Most of the reason inquiries do not increase lies in the content rather than the design, so this does not fix the underlying problem. Use the redesign as an opportunity to review the copy and the paths as well.
• Going live with no redirects: This is item one on the checklist above. The urge to launch quickly is understandable, but skipping this risks rankings dropping not temporarily but over the long term. Before launch, always confirm that a list of the old URLs and a mapping table to the new ones exists.
• Starting without deciding the goal: Begin because “it just looks old” and you will spend the time arguing about design taste while the real goal — more inquiries, or a site that works for hiring — gets left behind. Before you start, narrow it down to the one thing you most want this redesign to achieve. With that fixed, it also becomes your criterion whenever you get stuck on design preferences.

Put the other way: avoid those three and a redesign is unlikely to go badly wrong.

And honestly, there are cases where a redesign is not needed at all. Your business and prices have not changed recently, inquiries through the site are enough as they are, and the only complaint is that it “looks a bit dated” — none of the six signs at the top apply. In that state, partial improvements such as swapping photos or adjusting the paths to contact are often enough, rather than a full rebuild. If we look into it and conclude you are better off not rebuilding your current site, we tell you so.

▶ Related reading: The website commissioning handbook | seven things to decide before requesting a quote, and four clauses to check in the contract | There are only five numbers a website manager needs to check each month | the minimum set in GA4 and Search Console

Working with us

We take on redesign projects too. Every project is published with the client’s name on our work page — for example the full corporate site redesign for JTB Americas, the site redesign for Ibuki Clinic, and the design redesign for Taisho, a Wakasa Bay wild-caught seafood restaurant.

The process is the same four steps as a new build. In a discovery session (60 minutes, once) we organize the goal of the redesign and what is being migrated; we build every page with AI while a human finishes it; we take a check-and-revise round (up to two revisions) before launch; and we go live with the DNS switch and the measurement setup done.

Because there is an existing site to migrate from, we spend more time at the discovery stage sorting what to keep and what to change. We do not treat launch as the end: for the first week after launch we watch the search index status and the traffic trend together with you. We consider all of that part of the redesign.

Looked at purely as a figure, a redesign is a rebuild — but what is actually happening is the work of carrying everything you have built up into its next form without breaking it. Before you worry about cost and schedule, we can start by sorting out together what to keep and what to change.

As above, pricing is from ¥1,000,000 for AI Website Production, with the migration option for an existing site from +¥80,000. The figure moves with page count and migration volume, so start by asking us for a rough range in a free consultation. Even if you are still undecided about the timing or the approach, feel free to get in touch through our contact form.

FAQ

How much does a website redesign cost?

The baseline is roughly the same market rate as a new build, with the work of migrating existing content added on top. It varies with page count and migration scale, so we recommend checking a rough figure with our simulator.

Will a redesign hurt our search rankings?

There is a risk if the process is wrong. But if you handle URL design and 301 redirects, content migration, carrying over meta information and GA4 / Search Console, and post-launch index checks properly, you can aim to migrate while holding your rankings.

Can we change the domain during a redesign?

Unless you have a specific reason, we recommend keeping the domain as it is. Changing it makes it much harder to carry over the evaluation you have built up with search engines.

What should we check once the new site goes live?

Checking the display by eye is not enough. Confirm four things immediately after launch and again a few days later: that the old URLs redirect correctly to the new ones, that no search-exclusion directive has been left on by mistake, that the measurement tag is actually running and data is arriving, and that the number of pages registered in the sitemap is what you expect.