When a Website Deserves a Rebuild

The request almost always arrives in the same form: the site is old, we want a new one. In more than half of those cases our recommendation is not to rebuild, because the problem that prompted the request will not be solved by a new site.

Rebuilding is an expensive decision: money, months of management attention, and a genuine risk of losing search rankings built over years. This sets out when it is worth it and when it is not.

The wrong reasons, which are the common ones

"It looks dated." An outdated design is a real problem if it damages trust, but the fix is usually a visual refresh rather than replacing the underlying structure. Appearance is a layer; a website is not only appearance.

"A competitor launched a new site." Their site is not a benchmark because you cannot see their results. It may be more attractive and sell less.

"I cannot edit anything myself." That is a permissions, training or admin-interface problem, and it may be solved in days. Do not build a new website because nobody gave you the right password.

"We are refreshing the brand." If the identity is the problem, start there — otherwise you will build a new site and discover three months later that it does not represent the new identity. The correct order is covered in brand identity is not a logo.

The right reasons

The technical foundation cannot be repaired. A site assembled from accumulated plugins on a haphazardly modified theme, where every small change breaks something else and security updates cannot be applied without collapse. Here the cost of patching exceeds the cost of building within a year.

Performance cannot be rescued. If a page needs more than four seconds to become usable on a mid-range phone on an ordinary network, and the cause is proven to be the theme's own structure rather than images and plugins, partial repair will not be enough. Performance is not a luxury: each second of delay removes a measurable slice of conversion, and you are paying for that visit through advertising.

The business model changed. A brochure site became a store, a single-product store became a marketplace, a local service now sells in two countries with two currencies. The old structure was not designed for it, and bending it costs more than building.

The site measures nothing and cannot. If the purchase path does not permit correct conversion tracking, you are spending on advertising blind. That alone can justify a rebuild.

The purchase path itself is the problem. Too many checkout steps, forced registration before purchase, missing payment methods common in your market. Sometimes repairable, sometimes written deep into the platform.

The test before deciding

Do three things before you decide. They take a week and save months.

Measure the current state precisely. Load speed on mobile, conversion rate, cart abandonment rate, and the most-visited pages. Without these figures you will not know after launching whether you improved — which happens more often than anyone admits.

Watch five real users. Ask five people from outside your company to complete a purchase or an enquiry while you watch in silence. You will discover problems in half an hour that nobody reported in a year.

Ask for both quotes. A serious supplier can tell you: fixing these five points costs this much and returns that; a full rebuild costs this much. Anyone offering a full rebuild without diagnosis is selling a project rather than a solution — which is the commitment behind Noprmg.

If you rebuild: how not to lose your search rankings

The most dangerous part of a rebuild is not the cost but the loss of organic traffic. We have seen sites lose more than half their visits after launching a "better" site because nobody handled the old links.

Capture the complete old URL map before anything else. Every published link gets an equivalent or a permanent redirect to the closest page by subject. A link with no equivalent is redirected to a section page rather than the home page, and what has genuinely ended is declared gone explicitly rather than left to fail silently.

Move the content, not only the design. Pages that bring traffic must survive with their text and headings; optimise them weeks after launch, not during it.

Watch the first thirty days daily. Crawl errors and drops in impressions appear early. Correcting them in the first week costs hours; correcting them after three months costs months.

The cheaper alternatives people skip

Before a full rebuild, try these in order.

Performance work alone. Compress assets, cut plugins, enable caching, and fix what blocks rendering. This often produces a marked improvement for a fraction of the cost of building.

Rebuild specific pages only. The product page, the checkout, and the single most important landing page. These pages create most of the result and can be renewed inside the existing site.

Simplify the purchase path. Removing one unnecessary step or field can raise checkout completion by an amount equivalent to a large advertising increase, at no extra media cost — an arithmetic we set out in the budgeting piece.

What actually costs money in a website project

Misunderstanding cost is a leading cause of failure. The most expensive line item is neither design nor development: it is decision-making.

A project that runs three months late is usually late because nobody approved the copy, or three people gave contradictory opinions, or the content was never ready. A site whose text, assets and structure are settled before implementation begins is delivered in half the time with half the friction.

Before you start, prepare four things: the list of required pages, the copy for each page even in draft, the name of the single person with final approval, and the metric by which success will be judged after launch.

Realistic timelines

A brochure site of five to ten pages: four to six weeks with content ready, three months without.

A store on an existing platform: six to ten weeks, with most of the time spent on product entry, shipping and payment rather than on design.

A custom build for genuinely specific logic: three months and upward, requiring a written analysis phase before any code.

Anyone promising a complete store in a week means a ready template with products entered in haste — a legitimate choice for testing a market quickly, provided you know that is what you bought.

After launch: the first month is the real project

Many treat launch as the end; it is the beginning of measurement.

In week one, watch crawl errors, broken links and real page speed on actual devices. In week two, walk the purchase path yourself from an ordinary phone on an ordinary network, and complete a real paid purchase.

Then compare against the figures you recorded before the build. Did conversion improve? Did load speed improve? Did organic traffic rise or fall? If it fell, the cause is most likely technical — redirects or link structure — and it is fixable when caught early.

The short version

Ask one question: which number will improve after the rebuild? If you cannot name a specific number and measure it today, you do not need a new website. You need a diagnosis.

For that diagnosis, describe your situation, or read about our sectors.