What Should You Keep From Your Old Website, and What Should Go?

Keep what the site has earned and let go of what merely accumulated. The addresses that already rank, the pages that bring in enquiries, the photographs of your actual premises, the proof you have done this work before — those carry over. A website content audit before redesign is the exercise that tells the two apart, and it belongs at the start of the project, before anyone designs a screen.

It has to be done on paper because the instinct runs backwards. Pages get judged on how they look the morning somebody opens them, and the ones that look worst are often doing the most work. The tidy pages nobody visits survive because nobody ever argued with them.

What a website content audit before redesign actually tells you

There are two jobs here, usually run together. Nielsen Norman Group draws the line between them: a content inventory is a list of every piece of digital content you currently have, while the audit is the part that examines and evaluates its quality. One is a census. The other is a verdict.

Start with the census. Export every URL — from the sitemap, or a crawl if the sitemap has drifted — and give each page one row. Then attach the few things that decide its fate: organic clicks over the last twelve months from Search Console, entrances from your analytics, any enquiry or WhatsApp message you can trace to it, when it was last touched, and whether anything outside your site links to it.

What comes back is almost always lopsided. A small group of pages does nearly all the work — the homepage, two or three service pages, and one article somebody wrote years ago that happens to answer a question people keep asking. Underneath sits a long tail nobody has opened in a year, including pages the owner had forgotten existed.

At this stage you are not judging the writing. You are finding out which pages the business would actually miss.

Doing this first stops the redesign deciding by default. A page nobody lists gets rebuilt out of habit or dropped without discussion, and both cost money. Whether the site needs the work at all is a separate question worth settling first.

Four decisions, not two

Keep or go is too blunt a choice, and it is why so many audits stall. Every row should end with one of four verdicts.

• Keep. The page earns its place and what it says is still true. Keep is mostly a decision about the address rather than the wording: the same URL can be rewritten from top to bottom and still be the same page to a search engine, to anyone who bookmarked it, and to every site that linked to it.
• Rewrite. The page earns, but it describes a process you have changed or takes six hundred words to say what needs two hundred. Same address, new content.
• Merge. Three thin pages about the same service, each ranking for nothing in particular and competing with each other. Fold them into one page that carries the subject properly.
• Retire. Nothing links to it, nobody reads it, and it describes something the business no longer does. Even then, the decision is where that address sends people afterwards, not simply a deletion.

Merging is where nervous owners back out, so the mechanics are worth knowing. Google treats a redirect as a strong signal that the target of the redirect should become canonical, and consolidating brings the signals held by those separate URLs, including the links pointing at them, into a single preferred one. Three weak pages folded into one strong page is a gain, not a sacrifice.

Retiring asks the same care in reverse. An address that vanishes with nothing in its place turns every old link and bookmark into a dead end, which is why holding on to rankings through a redesign is mostly about knowing where each retired URL points the morning after launch.

What owners give away too easily

These are the assets a revamp tends to lose — not because anyone decided to lose them, but because nobody wrote them down as worth keeping.

• Addresses that already rank. A URL that has spent four years accumulating links and search history is an asset you did not pay for. Swapping it for a tidier one with no redirect throws that away for neatness.
• Real photographs. Your premises, your team, your work half-finished on site. Replacing them with stock photography is the quickest way to make a credible local business look like a template.
• Testimonials and case studies with names on them. A named customer describing a specific job outweighs any amount of writing about commitment to quality. If they are buried, move them forward.
• The specifics. Coverage areas, lead times, model numbers, the sizes you stock, the jobs you do not take. Redesigns smooth these into something cleaner and less useful, and they are what the buyer came to check.
• Answers you have already written. An FAQ built out of real WhatsApp enquiries beats anything invented from scratch, because the questions are the ones people ask.
• The privacy notice. Malaysia’s Department of Personal Data Protection is direct about it: a privacy notice must be consistent with all provisions under Section 7 and must be served when a customer’s personal data is processed. Dropping it while adding a new enquiry form is a step backwards.

What survives that should not

The other half of the list is the material that stays put because nobody wants the argument about it.

• The news page that stopped three years ago. A visibly dated page is worse than no page: it suggests the business may not still be trading.
• Near-duplicate service pages. Built to chase separate search terms, now splitting one subject between them and each ranking worse for it.
• Vision, mission, and a paragraph about passion. Nobody has bought anything because of it, and it pushes the useful material down the page.
• Pages for work you stopped taking. They generate the enquiries you have to turn down, costing you time and the enquirer their goodwill.
• The PDF brochure. Check your own download figure before defending it. If it is genuinely used, keep it and bring it up to date.

Google publishes a self-assessment for exactly this judgement, and one question is the most useful thing to hold an old page against: is this the sort of page you’d want to bookmark, share with a friend, or recommend? Asked honestly of a decade-old site, it clears a surprising amount. Cutting a page is still not the same as deleting its address: send the old URL somewhere sensible rather than letting it resolve to nothing.

Keeping the audit from swallowing the project

Audits stall when they become a scoring exercise. Nielsen Norman Group points out that inventories and audits are ongoing tasks, so they’re never really finished. For a revamp you do not need finished. You need decided.

• One working session with the list open, not a fortnight of individual review.
• One verdict per row, no blank cells, and no category called “maybe”.
• One person signs it off — usually whoever answers the phone, because they know which pages customers mention.

Frequently asked questions

Where iRevamps fits

Every revamp we run starts with this list, and the client sees it before any design work begins — one row per page, one verdict each, the reason written beside it. Our website revamp and design service treats what your site has already earned as the starting material rather than something to clear away, which is usually why a revamp beats a rebuild. If you have put the work off for fear of losing what already works, that fear is answerable on a spreadsheet before anybody touches the site.