Rebuild the website or improve the existing one: a guide to the decision

Few decisions get made with less information than this one. On one side a quote to fix things up, on the other a quote to start over, and in between the feeling that the current website isn't good enough. But the question isn't “rebuild or not”: it's understanding which problems you're actually facing, because three different problems have three different answers, and only one of the three calls for starting over.

The fastest way to get it wrong is to start from the quote. The cheapest way to decide is to start from an inventory: what you have, what's holding you back, what can be carried over.

An inventory before any judgment

You don't need a long analysis. You need three lists, actually written down, because from memory a website always looks smaller or bigger than it really is.

• The pages that do something: the ones that get visits, the ones you email to customers, the ones a salesperson opens in front of a client. There are almost always far fewer of these than the total, and they almost never match the menu items.
• The technical constraints: who can edit the website and with what tool, what requires outside help every time, what's connected to something else — the business management system, the catalog, the application form, a private area.
• The content worth keeping: copy, photos, spec sheets, videos, translations. The value of a website rarely lies in the website itself; it lies in the material inside it, which took time to produce.

Three problems that look like the same problem

“The website doesn't work” is a sentence that hides three different ones. Telling them apart is almost the whole job of this decision.

• The message. The pages exist but say the wrong thing, or say it to the wrong person: they talk about the company instead of the customer's problem, the services are listed with internal names, the projects have no context. This is fixed by writing.
• The navigation. The information exists but can't be found: sections that look alike, pages reachable only from the menu, paths that end in a dead end. This is fixed by reorganizing, and sometimes by moving a few pages.
• The platform. The limit isn't what's written, it's what can be done: there's no page type suited to what you need, every change goes through a vendor, the website can't handle a catalog or an extra language, the mobile version can't be fixed without rebuilding the page template.

The practical rule is this: if fixing a message problem requires calling in technical work, the real problem isn't the message.

Three levels of intervention

Between fixing and rebuilding there's a third level, often the best fit and almost always the least considered.

Three situations, and the choice that follows

The scenarios below are reconstructed as examples; the features, though, are the ones that come up most often.

A services company, a website from a few years back, few pages

The pages can be updated by people working at the company, the website reads well on a phone, but the services are described using words only internal technical staff use, and there's not a single project told with context. Here the platform has nothing to do with it: the cost of a new website would go toward laying out the same text again. The fix is on the services pages, a case-studies section is added, and the decision to rebuild is picked back up later, with the data that will have been gathered in the meantime.

An industrial company that wants to add a private area for resellers

The website tells the company's story well and the main pages bring in leads, but there's no way to manage access, tiered price lists, and documents. The question isn't whether the website looks good: it's whether the platform can host that function without becoming fragile. If it can, it's a partial rebuild. If every attempt requires non-standard additions, a full rebuild becomes the cheaper path over time, even when today it looks like the opposite.

A company that has changed its line of business, a website inherited from a vendor that no longer exists

No one at the company can change a single line, the credentials are incomplete, the website describes a business that isn't the current one anymore. Here there's no decision to make: the message, the control, and the ability to make changes are all missing at once. The real work, though, isn't technical: it's understanding what the company does today, and for whom.

What needs to change, and how you'll know

Before choosing, write down what should happen if the fix works. Not a generic goal, but the sentence you'd like to be able to say in a few months: more leads from a specific service page, fewer calls asking things that are already written down, the ability to publish a page without calling anyone.

You also need the opposite: the condition that makes you stop. A partial rebuild that after two sections has already needed three fixes outside the quote is telling you something about the platform. A targeted fix that can't be completed because every change breaks another one is telling you the same thing. Deciding this in advance keeps you from deciding under pressure halfway through, when stopping costs more than continuing.

What this guide doesn't cover

This guide takes you as far as the decision, before the project starts. The move from old to new — page URLs, redirects, content that has to stay reachable — is a job of its own and has a dedicated guide. And if your starting point is “I have visits but no leads,” the diagnosis has to come first: the problem might not be the website at all.