Most redesigns begin with an opinion and a mood board. The ones that succeed begin with an audit, because you cannot improve a site you have not measured, and because the audit is what stops a project turning into a taste argument between three stakeholders. Done properly it takes a few days and it produces something far more useful than a report: a scope.
Pull the data before you look at the design
Open analytics first, deliberately, before forming any aesthetic opinion. You are looking for four things.
Which pages get traffic, ranked. Where that traffic comes from (search, direct, referral, paid), because a site living on paid traffic needs different work to one living on organic. Where people leave. And what percentage of visits produce the outcome you care about, which is the number the whole project will be judged against.
Then look at search specifically: which queries you already rank for, and which pages earn them. This list is the thing a redesign most often destroys by accident, and knowing it in advance is half of protecting your rankings. Note it down, page by page, before anyone starts deleting.
Inventory every page and give it a verdict
Export a full list of URLs. On a small business site that is fifty rows; on an older content site it can be thousands, and the number itself is usually the first surprise.
Against each page put its traffic, its conversions, and one of four decisions: keep as is, improve (the page matters and underperforms), merge (three thin pages saying the same thing become one good one), or remove. Most sites can delete a third of their pages without losing anything except maintenance burden, and consolidating near duplicates usually improves rankings rather than harming them.
This inventory becomes the redesign scope. Every page marked improve is work; everything marked remove needs a redirect plan.
Five checks that find the real problems
Performance. Run the real measurement on a mid range phone rather than your laptop, and separate the structural problems from the easy ones. An unoptimised hero image is an afternoon; a build that ships a megabyte of JavaScript to render text is an architecture decision.
Mobile. Not “does it fit” but “can you complete the main task with one thumb on a train”. Most conversion loss on business sites is hiding here.
Accessibility. Run an automated scan for the obvious failures, then check the things scanners miss: keyboard navigation, focus states, and whether every form field has a real label. This is both a legal and a usability question.
Technical SEO. Indexing status, duplicate titles, missing metadata, broken links, redirect chains, and whether the site is even crawlable. Also check what structured data exists, since it is often absent entirely.
The conversion path. Walk through it as a stranger. Count the clicks from landing to enquiry, and note every point where you would need to already know something to continue.
Ask the people who are not you
Internally, ask the sales team which questions they answer on every call. Those answers belong on the site and usually are not there. Ask support what confuses customers.
Externally, watch five people attempt one task on the current site and stay silent while they do it. Half a day of that is worth more than any heuristic review, and the method is cheap enough for any budget. Add a five second test on the homepage if positioning is in question at all.
Look at the competition properly
Not to copy, and not to collect screenshots of things you like. Compare claims: what do three competitors promise, and what do you promise that is different? If the answers are interchangeable, no redesign will fix your conversion rate, because the problem is the words.
Note what they publish, what proof they show, and how they handle pricing. In most markets the site that says the price wins the shortlist.
Turn it into decisions, not a document
An audit that ends in a 40 page PDF has failed. The output should be three lists: what to fix now (cheap, high impact, no redesign needed), what the redesign must solve, and what to stop doing.
That third list is the one nobody writes and everybody needs. It is also where the honest answer sometimes appears: that the site does not need rebuilding at all, and a refresh plus better copy would do.
Once the scope is real, the design work has something to aim at, and a brief that specific gets you a better proposal from anyone you hire.
We run this before every rebuild we take on. If you want the same audit on your site, hello@beconfidency.agency, and the results are on the work.
If you would rather have the audit and the rebuild handled by one team, that is exactly what our website redesign service is for.