Should You Rebuild Your Website or Fix the One You Have?

There is a conversation that happens in nearly every marketing meeting eventually. Someone says the website looks tired. Someone else agrees. Within twenty minutes the group has talked itself into a full rebuild, and nobody in the room has asked the only question that matters: what is actually broken?

We should be upfront about our own position here, because it colours the advice. Rebuilds are more profitable for agencies than fixes. A rebuild is a defined project with a large number attached to it. A fix is a series of smaller invoices that are harder to sell and harder to celebrate. When you ask an agency whether you need a new website, you are asking a question they have a financial reason to answer one way.

So here is the version we give people on the phone, including when it costs us the bigger project: most sites we are asked to rebuild do not need rebuilding. They need three or four specific things fixed, and the owner has been told the whole thing is beyond saving because that is an easier sentence to sell.

What a rebuild actually costs you, beyond the invoice

The price of a new site is the part everyone focuses on, and it is the part that matters least. The real cost is what happens to your organic traffic on launch day, and the published research on this is not encouraging reading for anyone about to sign a rebuild contract.

Even a migration done properly — every redirect mapped, every URL accounted for, nothing rushed — typically sees organic traffic fall somewhere in the region of 10 to 25% for the first month, with full recovery taking anywhere from two to eight months depending on how much structure changed. That dip is not a mistake. It is Google reprocessing a site it thought it understood.

Done badly, the published figures are considerably worse: drops in the 30 to 80% range, with recovery periods measured in quarters rather than weeks, and in the worst documented cases twelve to eighteen months to get back to where the site started.

What a migration does to organic traffic

Three outcomes, plotted against the launch date. Even the well-executed one costs you a quarter.

Ranges drawn from published migration research: a typical well-run migration dips roughly 10–25% for the first month and recovers over two to eight months, while botched migrations are widely reported at 30–80% with recovery measured in years. Even a flawless migration sees four to eight weeks of fluctuation while Google reprocesses the site — that part is unavoidable and documented by Google itself.

Look at that green line for a moment, because it is the one nobody draws in the sales meeting. A site that is improved in place does not have a launch day. There is no cliff, because there is nothing to fall off. Improvements land one at a time, each one measurable on its own, and the traffic you already earned keeps arriving throughout.

If your business depends on the phone ringing, that window has a price, and it is rarely included in the quote. A contractor doing $40,000 a month with half of it originating in organic search is risking real money for a quarter to get a site that looks better. Sometimes that is worth it. Often it is not, and nobody did the arithmetic out loud.

The four ways rebuilds actually lose traffic

Here is the part that should change how you think about the whole decision. When a redesigned site loses its rankings, the design is almost never the cause. The published post-mortems converge on a small number of technical failures, and one of them dwarfs the rest.

Why rebuilt sites lose traffic

It is almost never the design. It is the plumbing underneath it.

Published post-mortems consistently put redirect and URL-mapping failures at roughly three quarters of all post-redesign traffic loss. The practical implication is uncomfortable: the single most expensive part of a rebuild is a spreadsheet nobody wants to own.

What fixing actually achieves

The case for fixing is not "it is cheaper", although it usually is. The case is that the specific things making your site underperform are mostly not design problems, and a rebuild therefore does not reliably solve them. It just re-houses them in nicer packaging.

Here is the pattern we see most often when a business tells us their website is not working. The traffic is fine. The rankings are fine. The site simply does not convert the people who already arrive. And the reasons are depressingly consistent.

What the cheap fixes are actually worth

Published conversion effects for changes that require no rebuild at all.

Form-field effect as published by HubSpot; the mobile figure is Google's own finding that 53% of mobile visits are abandoned past three seconds — those visits are not lost to design, they are lost before the design loads. The lower two vary too much by industry to put a single number on honestly, so we have not invented one.

Speed is a conversion problem before it is an SEO problem

Google's own research found that 53% of mobile visits are abandoned if a page takes longer than three seconds to load. Read that as a business number rather than a technical one: for every hundred people your marketing successfully persuaded to tap through, more than half never saw anything you wrote.

Those visits did not fail because of the layout. They failed before the layout existed. And a rebuild does not automatically fix this — plenty of beautiful new sites ship slower than the thing they replaced, because the page builder that made the design easy also made the page heavy.

You can check your own in about a minute, free, using PageSpeed Insights, and compare the result against Google's published Core Web Vitals thresholds — largest contentful paint under 2.5 seconds, interaction to next paint under 200 milliseconds. We wrote up what those numbers mean in practice in our guide to Core Web Vitals for local businesses, and the broader question of whether speed genuinely affects rankings separately.

Forms ask for more than they need

HubSpot's published research found that reducing a form from four fields to three increased completions by around 50%. Not a redesign. Not new branding. One fewer box.

Almost every business site we look at asks for something it does not need at the first contact. Company name. Address line two. A dropdown of services that the person has to interpret before they can proceed. Each one is a small tax on someone who was, a moment ago, willing to contact you.

The page does not say what the business does

This is the most common single problem and the least technical. A visitor lands, reads the headline, and cannot tell within a few seconds what you do, where you do it, or what happens if they get in touch. They leave, and it registers in your analytics as a bounce rather than as the failure of a sentence.

Rewriting a headline costs nothing and can be live this afternoon. It is also, in our experience, the change most likely to produce a visible difference in enquiries — which is awkward for anyone who just quoted you for a rebuild.

None of these three are design work. They are conversion work, and they apply equally well to a site built last month or eight years ago. If you want the fuller version of this argument, we covered the specific case of traffic arriving but the phone staying quiet in more depth elsewhere.

The three cases where rebuilding genuinely wins

We would be doing exactly what we accused everyone else of doing if we pretended the answer is always "fix it". It is not. There are three situations where a rebuild is the correct, cheaper, less risky decision, and they are worth naming precisely.

1. The platform is the constraint

Some sites cannot do what the business needs, and no amount of patching changes that. A booking system that cannot take deposits. A page builder that cannot produce a page under four seconds no matter what you strip out. A CMS the owner cannot log into without a support ticket.

The test here is simple: list the five things you want the site to do in the next two years, and ask whether the current platform can do them at all. If two or more are flat impossible, you are not choosing between fixing and rebuilding. You are choosing between rebuilding now and rebuilding later with more pages to migrate.

2. Speed cannot be recovered

Weight can usually be reduced. Sometimes it cannot. A site assembled from a heavy theme plus fifteen plugins, where half the load is scripts that arrived with features nobody uses, can reach a point where removing the bloat means removing the site.

The honest test: get a competent developer to spend two hours on it. If two hours of focused work moves the mobile score meaningfully, the site is fixable and you have found your cheaper path. If two hours barely moves it, the foundation is the problem. That two hours costs a fraction of a rebuild and it answers the question properly rather than philosophically.

3. The business it describes no longer exists

This one is not technical at all. Sometimes the site is a perfectly good description of a business you stopped running three years ago. Different services, different customers, different geography, different price point.

When the gap is that wide, fixing becomes an exercise in rewriting every page anyway, and you end up paying rebuild prices for a patchwork. At that point, rebuild deliberately — and treat the migration plan as the main deliverable rather than an afterthought. That is how we scope a custom build when one is genuinely warranted, and we publish what this sort of work costs so the number is not a surprise at the end of a sales process.

If you do rebuild, protect the thing you already earned

Say you have run the tests above and a rebuild genuinely is the right call. The difference between the blue line and the red line on that first chart is entirely within your control, and it comes down to a handful of decisions made before anyone opens a design tool.

1. Take a baseline first. Export twelve months of organic sessions by page, your top organic terms by page, conversion rates by page type, and current Core Web Vitals. Without this you cannot tell afterwards whether the rebuild worked. Most businesses discover they never took it and now cannot answer the question.
2. Build the URL inventory from more than one source. Your sitemap is incomplete. Crawl the live site, pull Search Console's indexed pages, check your analytics for pages receiving traffic, and reconcile all three. The pages missing from your sitemap are frequently the ones quietly ranking.
3. Map every old URL to exactly one new URL. Not to the homepage. A redirect to the homepage tells Google the old page no longer exists and passes nearly nothing. If there is genuinely no equivalent page, that is a decision to make deliberately, not a default to fall into.
4. Protect your ranking pages from the rewrite. Any page in your top fifty by organic entrances keeps its substance. Redesign the wrapper freely; leave the content that earns the ranking intact.
5. Never change the domain at the same time. If a domain move is genuinely necessary, do it separately, with Google's Change of Address tool, and leave months between the two events. Stacking a domain change on top of a redesign is how twelve-month recoveries happen.
6. Crawl staging against live before cutover. Every URL, every redirect, every canonical tag, every robots directive. This is the step that separates the blue line from the red one and it is the step that gets cut when the launch date slips.
7. Watch Search Console weekly for ninety days. Assign it to a person by name. Decide in advance what drop triggers an investigation. Detection lag is the expensive failure, and it is entirely preventable.

If your developer or agency cannot walk you through those seven points without hesitating, that is useful information about how the project will go. It is a fair question to ask before you sign anything, and how it is answered tells you more than a portfolio does.

How to decide in the next ten minutes

You do not need a consultant to get most of the way to an answer. Run these four checks yourself, in this order, and the decision usually makes itself.

1. Run PageSpeed Insights on your busiest page, on mobile. Under 2.5 seconds for largest contentful paint is healthy. Four or more, with a heavy theme underneath, is a genuine rebuild signal.
2. Open your site on your own phone and try to contact yourself. Time it. If it takes more than about twenty seconds from landing to a submitted form, the problem is friction, and friction is a fix, not a rebuild.
3. Open Search Console and look at your top twenty pages. If those pages get real traffic, you have something worth protecting and a rebuild is a genuine risk. If the site gets almost no organic traffic at all, you have far less to lose and the calculation changes entirely.
4. Try to make a small edit yourself. Change a phone number or a paragraph. If you cannot without emailing someone, that is a platform constraint, and platform constraints compound every single month you keep them.

Three or four of those pointing the same way is a clear answer. A split result usually means what it looks like: fix the specific things that are broken, and revisit the rebuild question in a year with better information and a site that has kept earning in the meantime.

That is the advice we give whether or not it wins us the work, and it is the same approach behind how we structure SEO engagements generally — find the constraint, fix that, measure, repeat. It is less exciting than a relaunch. It is considerably better arithmetic.

The decision tool

Answer six questions about the site you actually have. The scoring is the same logic we use on a discovery call, written down honestly — it is weighted toward fixing, because the evidence is weighted toward fixing, and it will tell you to rebuild when the answers warrant it.

If it lands you somewhere in the middle, that is a real result rather than a failure of the tool. A middling score usually means the site is structurally sound but neglected — which is the most common situation we see, and the one where a rebuild wastes the most money.

A note on doing this in a competitive local market

Everything above applies anywhere. There is one thing that makes the decision sharper in a market with real competition, and it is worth saying plainly.

In a thin market, losing rankings for a quarter is survivable, because the businesses above you are not actively working on it. Your position is largely still there when you come back. In a market where several competitors are running genuine campaigns, a three month absence is not a pause — it is a handover. Somebody else occupies the position, earns the clicks, collects the reviews those clicks generate, and consolidates. Getting it back costs considerably more than holding it would have.

We have watched this play out repeatedly doing Denver SEO for local service businesses, where a handful of categories — roofing, dental, legal, home services — are contested hard enough that the top three positions change hands within weeks when one participant goes quiet. If that describes your category, weight the risk side of this decision more heavily than a generic guide would tell you to.

The practical version: in a competitive market, if you are going to rebuild, do it in your slow season, not before your busy one. We wrote about how sharply demand swings by month in our piece on seasonal search demand, and the same logic applies to any market with a seasonal shape. Launching a migration four weeks before your peak is an unforced error, and it is one we see every year.

If you want an outside read on where your site actually sits before you commit to anything, that is what a free audit is for — and if you would rather just check a few things yourself first, several of the free tools we published cover the basics without anyone needing your email address.

Questions people ask before rebuilding