When to Rebuild Your Website

Almost every guide who asks me whether to rebuild has a repair problem, not a replacement problem. They have four or five specific complaints, each of which is a half-day fix, and the accumulated irritation of living with all five has become a feeling that the whole thing needs throwing out. That feeling is understandable and it is expensive to act on, because a rebuild costs a season of momentum even when it goes perfectly. So this is a triage rather than an argument: write down what is actually wrong, sort each item into repair or structural, and rebuild only if the structural column has something real in it.

What is the actual test?

Whether the thing you want is reachable from where you are. If every complaint on your list can be resolved by editing content, swapping images or deleting form fields, you have a maintenance backlog wearing the costume of a rebuild.

That distinction is worth being strict about because the two have wildly different costs. A backlog is evenings. A rebuild is a project, a supplier, a migration and a period where things are worse before they are better.

The test is also easy to apply. Take each complaint in turn and ask what specifically prevents fixing it in place. If the answer is "nothing, I just have not", that is not a reason to rebuild.

Run the measurement first so the list is real rather than remembered. Which numbers to collect, and in what order, is set out in the diagnostic piece.

Fishing with Manny, TX

What genuinely counts as structural?

Four things. You cannot edit it. You cannot extend it. It excludes people who use assistive technology. Or somebody else controls it and will not hand it over.

The cannot-edit case is the most common real one. A site built by somebody who has moved on, on a system nobody documented, with no login you hold, is a site that will keep aging regardless of intent.

The cannot-extend case shows up when the business changes. You added a second boat, a new water, a different trip type, and the site has no way to represent it without breaking its own layout.

The control case is not a design problem at all, and it belongs first in the sequence, because it changes what a rebuild would even mean. The ownership piece walks each asset in turn.

Why does accessibility sit in the structural column?

Because the failures usually live in the markup rather than the content, which means they are not editable from the outside. A form control with no label, a control that cannot be reached with a keyboard, a page whose structure is invisible to a screen reader.

W3C's forms guidance sets out what correct looks like: label every control, group related fields, provide instructions explaining how to complete each one, validate input while offering a way to undo, and notify users clearly about errors and successful completion.

One detail there has a payoff for every visitor, not only for people using assistive technology. A properly associated label gives a larger clickable area that includes the label text, which matters on a phone for small controls like radio buttons.

Test yours before assuming. Try to complete your booking form using only the keyboard, tabbing through it. If you cannot, neither can a real group of people, and that is usually a build-level fault rather than a content one.

What about speed? Is that structural?

Almost never. Slow guide sites are slow because of photographs, and photographs are content. Replacing the whole site to fix image weight is replacing a truck to change a tyre.

Measure before concluding. Three published thresholds apply: two and a half seconds to render the main content, a fifth of a second to answer a tap, and a layout-shift score no worse than 0.1. Each is scored on three quarters of genuine visits, phones counted apart from computers.

If you fail those and your pages carry a dozen full-resolution images, resize them and measure again. That is an evening, it is free, and it resolves most cases outright, as argued in the speed piece.

The genuinely structural version exists but is rarer: a site whose theme loads enormous amounts of code before showing anything, where no amount of image work moves the number. You will know because the measurement barely improves after the images are fixed.

Is a rebrand a reason?

Only if the rebrand is already real. A new name, a new logo and a new boat is a genuine trigger. Wanting a fresher look is not, and the two get conflated constantly.

The distinction is whether anything outside the website changed. If your business is the same business and you are simply tired of looking at the site, that is a colour and typography problem, and on most platforms it is a settings change rather than a rebuild.

Where a real rebrand does justify it, do the rebrand first and completely. A half-executed rebrand baked into a new site is the most expensive way to discover you did not like the new name.

Be honest about which of the two you actually have. Guides diagnose themselves with an out-of-date design more often than with anything else, and it is the diagnosis worth trusting least, precisely because you stare at that site a hundred times more than any client will. The mistakes piece covers why.

What does a rebuild cost that nobody quotes?

Momentum. Every page address that changes is a link somebody made that now leads nowhere, and the traffic loss shows up weeks later, after the launch, when the cause is buried.

The fix is not complicated and it is not optional. Every existing page must be mapped to its replacement and redirected, and that mapping has to be produced before anything is switched off.

Export the list of your current page addresses now, before you speak to anybody. After a rebuild nobody can reconstruct what used to exist, and the list is the only artifact that lets you check the job was done.

Ask every prospective supplier one question about it: how will you handle the old URLs. A competent answer contains the word redirect and an offer to map every page. A vague answer is the warning, and the wider handover checklist is in the switching piece.

When is the right time of year?

The week your season ends. Not spring, and certainly not June, because a rebuild consumes exactly the resource you do not have during a season: your attention.

No build gets finished without input nobody else can give. Images. Detail about the water you run. Sign-off on wording. Calls on which trips to list. A supplier stalled on you in July is a supplier idling, and the meter runs anyway.

Starting at the end of a season also gives the new site months to settle before it matters. Search behaviour after a migration takes time to stabilize, and you want that happening in February rather than in the week your calendar should be filling.

The rest of the offseason list, in priority order, is in the offseason piece.

Should you hire before or after deciding the scope?

Before, assuming you are hiring anyone. The platform's own documentation flags a redesign as one of the better moments to bring help in, and urges doing it early, so search considerations are designed in at the start instead of patched on at the end.

That page earns a read before any sales conversation for several other reasons. Most of this is within reach of a small local operator without help. An unpaid listing carries no fee. And no amount of ad spend moves a site's standing in the unpaid results.

It also carries one safeguard almost nobody bothers with. If a prospective shop proposes an audit, keep their permission on your search reporting at view-only for now.

And it tells you the question to ask, which is what results they expect and in what timeframe, and how they will measure it. That single question separates people quickly, as does the longer interview in the twelve questions piece.

What is the cheaper alternative to a rebuild?

Move onto a mainstream platform and redo the content yourself instead of commissioning anything. Billed yearly, the middle plan works out at $29 each month, so $348 across twelve.

Those are published rates read on 25 July 2026, on a ladder of $19, $29, $49 and $99 a month, with the page noting monthly billing costs up to 36 percent more than the annual figures shown.

That middle plan is the one a working business wants, because it is where a proper mailbox, extra editors and custom code become available. The entry plan will serve a site perfectly well and simply carries less of what an operating business ends up needing.

This route trades money for your hours, which is a real trade rather than a free lunch, and the three-year ledger with both columns priced is in the ownership cost piece.

Go Fish Charters, FL

What should you do instead, most of the time?

Spend four evenings on the repair column and then re-measure. If the numbers move and the irritation goes, you have saved a project. If nothing moves, you now have evidence rather than a feeling.

Evening one is mechanical: test the form from your own phone on cellular, call your own listed number, check the domain is in your account, and fix whatever fails.

Evening two is images: resize everything and measure the difference. Evening three is rates and trips: make the price findable and the trips distinguishable. Evening four is the writing, cutting rather than adding.

Then take a fresh measurement and compare. Four evenings against a rebuild is an obviously good bet, and the sequence of what a visitor actually needs, in order, is in the anatomy piece.

What if the site was built by somebody who has disappeared?

Establish what you can reach before deciding anything, because that determines whether this is a rebuild or a recovery. Three logins matter: the domain registrar, the hosting or platform account, and whatever holds the site's content.

Work backwards from the domain, because that one is recoverable through the registrar even when nobody answers email, and it is the asset you cannot afford to lose. Everything else can be rebuilt.

If you can reach the platform but not the person, you may not have a rebuild at all. Plenty of abandoned sites are perfectly editable once somebody logs in, and the abandonment was of the relationship rather than of the software.

If you can reach nothing, treat the existing site as a reference document rather than an asset. Screenshot every page, save the text, note every address, and then build fresh with that as your source material.

How do you brief a rebuild so it does not drift?

Put the four jobs in writing before the first call and hold the line until each is met. Somebody can find the price. Somebody can book from a handset. Pages arrive quickly. And it is obvious who runs the operation.

Briefs drift because everything sounds reasonable in isolation. A gallery, a blog, a conditions feed, a newsletter signup, a trip-comparison table. Each is defensible and collectively they are why builds run late.

Hold the four and treat everything else as phase two, explicitly, in writing. A supplier who is told there is a phase two behaves differently from one who thinks this is the only chance to include something.

Then check the finished thing against the same four rather than against how it looks. A build that is beautiful and fails the second job has not delivered, and that judgment is easier when the criteria were written before anybody showed you a design.

What do experienced guides do differently?

They keep a running list all season and decide in October. And they never decide during a bad month.

The running list is what turns a vague dissatisfaction into a triage. Every time something annoys you during the season, write the specific thing down. By October you have eight items, and eight specific items sort cleanly into two columns.

Deciding in October rather than in July is the other half. A rebuild decision made in a thin August is a decision made about a season, not about a website, and the two get confused because they arrive together.

Experienced operators also rebuild less often than newer ones expect. A site that is maintained does not need replacing on a schedule, and the guides with the best sites are usually the ones who have edited the same site for six years rather than commissioned three.

What does the repair column actually look like in practice?

Short, specific and boring, which is why it gets skipped in favour of the exciting option. Most guide sites have between four and seven items in it and none of them takes longer than an evening.

Write each one as an action rather than a complaint. Not "the booking page is bad" but "delete the experience-level dropdown and the how-did-you-hear field". The rewrite from complaint to action is where half the perceived difficulty disappears.

Then order them by how long each takes rather than by how much they annoy you. Working shortest-first produces visible progress in the first hour, and visible progress is what makes anybody finish a list like this.

Cross items off as you go and keep the list afterwards. Next October you will have forgotten what you fixed, and a completed list is the evidence that the repair route worked, which is what stops the rebuild question returning every year.

What are the common mistakes?

Rebuilding to fix content problems. Launching without redirects. Starting in spring. And treating a rebuild as a way to avoid deciding what the site should say.

The last one is the subtle failure and the most expensive. A rebuild does not generate copy, choose trips or decide what your rates are, and a project that starts before those are settled turns into months of waiting on you.

The redirect mistake is the one that costs measurable traffic and takes weeks to become visible. Get the URL map as a named deliverable rather than a promise.

The spring mistake compounds everything else. A build starting in March is a build competing with your season for your attention, and it is why so many rebuilds launch late, incomplete, or in the middle of the months they were meant to help.

What surprises people?

That most rebuild requests are maintenance backlogs. That speed is almost never a reason. And that the strongest reason to rebuild has nothing to do with how the site looks.

The control case is the one worth sitting with. A site you cannot edit, on a platform you cannot leave, held in an account that is not yours, will not improve no matter how much you dislike it, and that is the one situation where replacing it is unambiguously correct.

Everything else earns the four evenings first. Nearly all of what sits on the guide websites hub is repair work rather than replacement, which is itself the argument.