Website Redesign or Targeted Fix? How to Tell Which One You Need

A business owner looks at the company website, decides it looks tired, and asks three agencies to quote for a rebuild from scratch. More often than not, that owner has just agreed to spend five times what the problem was worth, when a week of focused work would have done the job. A full website redesign earns its cost when the problems are structural: technology that has you locked in, an architecture nobody can navigate, a mobile experience that barely works. When the problem is contained (one page that won’t convert, a slow site, content that stopped in 2021), a targeted fix costs a fraction and gets you most of the way there. This article gives you a scored checklist so you can work out in ten minutes which of the two situations you’re in.

Design research has been saying this for decades. Gobeli and Brown estimate, as cited in an Aarhus University paper on incremental innovation in web design, that around 80% of new products are incremental innovations: improvements to something that already exists rather than radical rewrites. Most of a designer’s day-to-day work is making existing things better. ¹

Restyling, redesign and rebuild are three different jobs (with three different budgets)

People use the three terms interchangeably, and they shouldn’t. The difference can move a quote by an order of magnitude, so get them straight before you talk to any supplier.

A restyling changes how the site looks and feels while keeping the existing structure. New visuals, typography and images, maybe a handful of sections reshuffled, but the pages and user journeys stay the same. A redesign rethinks those journeys. You rework the menu, the route a visitor takes to get in touch, the information architecture. A rebuild starts on a new platform, often a different CMS, and reconstructs everything.

Job	What changes	When it makes sense	Relative effort
Restyling	Visuals, UX, visible content; structure untouched	The site works but looks dated	Low (days)
Redesign	Journeys, menu, information architecture	Users get lost and can’t find what they need	Medium (weeks)
Rebuild	Platform, code, content from scratch	The technology holds you back or the architecture can’t be saved	High (months)

The most common reason for a redesign is content people can’t find. In a user research project run by the Teacher Retirement System of Texas with a sample of 494 people, the concern raised most often across every study was information overload combined with poor findability and discoverability on the site. ² You fix that by reorganising how the pages are structured. That’s redesign territory: too much for a simple restyle, and far less than a full rebuild.

The measurable signs your site needs rebuilding (numbers first, gut feeling second)

“The site feels old” is a feeling. Decisions come from numbers, so measure before you touch anything. Every signal below can be checked for free in under ten minutes with tools you already have.

The site is slow for real users

Open PageSpeed Insights, paste in three URLs (home, a service page, a content page) and look at the Core Web Vitals measured in the field. Google’s published thresholds are clear: LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1. If your LCP sits at 5 seconds, half your visitors have left before they see the offer. On its own, this signal doesn’t call for a rebuild. Slowness usually comes down to uncompressed images, cheap hosting or no caching, and all three can be fixed without touching the structure.

Mobile is an afterthought

Go into Google Analytics, segment traffic by device and compare two numbers: the share of visits from mobile and the mobile conversion rate. For Italian SMEs, mobile traffic almost always tops 60%. If mobile brings in 65% of visits but converts at a third of the desktop rate, the site was built for a big screen and squeezed badly onto a phone. A site that makes people pinch to read or zoom in to hit a button has a structural problem, and that’s where a rebuild becomes likely.

The site doesn’t bring in leads

Traffic arrives, the forms stay empty. Before you decide the whole site has to go, isolate the problem. If a service page gets 400 visits a month and produces two enquiries, you’re looking at a problem with that one page, and the rest of the platform may be perfectly healthy. A conversion rate below 1% on a commercial page with qualified traffic points to copy, social proof or the call to action, and all of those can be fixed on a single page.

The technology is holding you hostage

Try editing the text on a page yourself. If you can do it in ten minutes, your CMS is healthy. If you have to raise a ticket, wait for a developer who replies every other day, and pay for every change, the technology has become a hidden recurring cost. A CMS that no longer gets updates, abandoned plugins, total dependence on a supplier who doesn’t pick up the phone: these push you towards a rebuild far more than any question of looks.

The brand has moved on and the site hasn’t

You’ve redone the logo, repositioned, added two services, and the site still describes the company you were three years ago. The gap between who you are and what the site says chips away at the trust of anyone weighing you up. Someone who clicks through from a recent post and lands on a site talking about a different offer starts wondering which version is true, and when in doubt, they close the tab.

For the first audit, a language model can speed things up once you’ve gathered the data by hand. Here’s a prompt that works:

“Act as a conversion consultant for small and mid-sized businesses. I’m pasting data for three pages of my site: URL, monthly visits, conversion rate, LCP, share of mobile traffic and mobile conversion rate. For each page, tell me whether the problem is contained (a single page) or structural (the whole site), and give it a priority from 1 to 5. Justify every call with the numbers I’ve given you, and don’t invent any others.”

The prompt speeds up the diagnosis without replacing it, as long as the numbers come from your own measurements.

When a targeted fix is enough (the 4 cases where rebuilding is money down the drain)

No web agency enjoys writing this section, because rebuilding a site pays better than fixing one. Yet good design work is almost always incremental. The Aarhus University study is blunt about it: treating incremental innovation as less important, or even trivial, is a misunderstanding, because most design practice consists of small, targeted improvements, a long way from the glamour of technological revolution. ¹ Here are the four SME scenarios where a full rebuild is a waste.

The site holds up but one key page doesn’t convert

The structure is sound, the CMS works, the traffic comes in, but the page meant to generate leads stays silent. You work on that one landing page: headline, social proof and call to action. This is a job measured in days.

The site is slow but well built

The pages are well organised and the journeys are clear, but everything loads slowly. Nine times out of ten, the culprits are heavy images, underpowered hosting or no caching. Compress the images, move the site to proper hosting, switch on a cache. Your Core Web Vitals jump within days, and the structure stays exactly as it was.

The content is stale but the structure works

The architecture does its job; the copy needs work. It talks about services you no longer offer, shows case studies from three years ago, or uses a tone that doesn’t sound like you. What you need is an editorial refresh and a visual update of your materials on the platform you already have. A language model can help you rewrite in bulk, provided you give it the right tone:

“Rewrite these three service pages without changing any of the facts. Use a direct, practical tone aimed at owners of Italian SMEs. Mix short sentences with longer ones. No emphatic adjectives without data to back them up. For each page, add a concrete call to action tied to a contact request. The current copy is pasted below.”

For the visual side, our design team redoes the layout of your materials without rebuilding the site.

The design is dated but the CMS is healthy

You can edit pages yourself and the technology is up to date, but the look is stuck a few years back. This is a pure visual restyle: new typography, a new palette and fresh images on the existing structure. Weeks of work instead of the months a rebuild takes, and visitors will perceive almost the same result.

The decision checklist: 10 questions, one score, one decision

Answer yes or no to each question and count the “no”s. The total tells you what to do, free of hearsay and of the bias of anyone who stands to gain from selling you the most expensive option. Keep this part handy when you’re deciding whether to rebuild your website.

1. Can people use the site on mobile without zooming in to read or to tap buttons?
2. Is the LCP of your main pages under 2.5 seconds?
3. Can you edit the text of a page yourself in under an hour?
4. Do the CMS and its components still receive updates?
5. Does the site reflect the company’s current offer and positioning?
6. Can a visitor reach the contact page in two clicks or fewer?
7. Do your commercial pages convert at least 1% of qualified traffic?
8. Is the menu structure clear to someone who doesn’t know the company?
9. Do you have full access to the code and content without depending on a single supplier?
10. Has your main content been updated in the last twelve months?

How to read the score:

• 0 to 3 “no”s: targeted fixes on the weak points are enough.
• 4 to 6 “no”s: you need a partial redesign focused on journeys and architecture.
• 7 or more “no”s: a full rebuild is the economically sensible choice.

What drives the cost (and why ten patches cost more than one rebuild done properly)

No serious quote comes from a fixed formula, because the factors that move the price change from project to project. The ones that really count are the number of pages and templates to produce, migrating existing content, integrations with your CRM or management software, copywriting, migration SEO, and producing photos and visual assets. An eight-page site with the copy already written is on a different planet from a fifty-page site migrating off an old CMS and hooking into your management system.

There’s also a cost nobody puts on the quote: the cost of not deciding. Every stopgap on a rotten technical base adds to the last one. You fix the speed, then mobile, then a plugin breaks, then another. A year later you’ve spent as much on patches as a proper rebuild would have cost, and you’re still standing on the same shaky foundations. If the checklist gave you seven “no”s, every euro spent on targeted fixes is a euro you’ll end up spending again. Our self-service tools run on a pay-as-you-go model.

How to rebuild your site without losing your Google rankings

The riskiest moment of a rebuild is launch day. That’s when sites throw away months of organic traffic, almost always through avoidable migration mistakes. Follow these steps in order.

1. Take inventory before you touch anything. Export from Search Console the list of URLs that get traffic and the queries they rank for. Without that starting snapshot, you’ll never know what you’ve lost.
2. Map every old URL to its new URL, one to one. Every address that changes needs a 301 redirect to its exact counterpart. Redirecting everything to the homepage is the fastest way to wipe out the rankings of your inner pages.
3. Keep what already works. Pages that rank well keep their title, H1 and the substance of their content. If a page sits at number one for a query that brings you customers, touch it gently and leave its substance alone.
4. Don’t launch without a verified Search Console property and an updated sitemap. Submit the new sitemap on launch day, and verify the property beforehand.
5. Monitor for 4-8 weeks. Check coverage, crawl errors and your main queries every week. A dip in the first two weeks is normal; a dip that hasn’t recovered after a month points to a missed redirect or a lost page.

Here too, a language model can take on the tedious part. Once you’ve exported both URL lists, old and new, you can ask:

“I’m pasting two columns: URLs from the old site and URLs from the new site. Match each old URL to the most relevant new one by content, produce a table ready for 301 redirects, and list separately any old URLs without an obvious match so I can check them by hand.”

You still need to check the matches page by page, because a wrong redirect sends users and search engines to the wrong place. But starting from a reasoned draft saves you hours on a site with dozens of pages.

A site rebuilt in 2026 has one more job than a site built five years ago: being read and cited by answer engines such as ChatGPT, Gemini and Perplexity. The things that help a person find information (well-structured content, headings that answer real questions, direct answers at the top of each section) are the same things that make your site findable by AI. The information overload and poor findability that user research flags as the number one problem for people ² get in the way of a machine trying to pull an answer from your pages just as much. Rebuild without thinking about this, and the new site is out of date on launch day.

Fonti

Footnotes

1. The same, but better: Understanding the practice of designing for incremental innovation in web design. pure.au.dk ↩ ↩²

2. [PDF] UX Research - Teacher Retirement System of Texas. trs.texas.gov ↩ ↩²