A redesign is a symptom, not a diagnosis
When a team decides it needs a new website, the trigger is almost never the website. It is a sales call that went badly, a competitor’s launch, a new CMO, or the slow accumulation of embarrassment about a page nobody has updated since 2022.
Those are legitimate prompts. They are not diagnoses. If you rebuild without establishing what is actually failing, you will get a site that looks current and performs identically — which is the most common outcome of the projects we are asked to review.
Six questions worth answering first
Answer these honestly before anyone opens a design tool. If you can answer all six, you are ready to brief a redesign. If you cannot, the first project is a decision, not a build.
- What specific decision should a visitor be able to make more confidently after the rebuild?
- Which three pages generate the most qualified pipeline today, and why?
- What do sales say prospects misunderstand about your offer?
- Which URLs currently earn search traffic you cannot afford to lose?
- What can your team publish today without a developer, and what cannot they?
- What will you measure eight weeks after launch to decide whether this worked?
The three projects hiding inside "redesign"
Nearly every redesign request is one of three different projects. The first is a positioning problem: the site says what you do accurately and unpersuasively. That is a message and structure project, and it can often be delivered on the platform you already have.
The second is a capability problem: the site is fine but your team cannot change it. That is a CMS and content-modelling project. The third is a genuine technical problem: performance, accessibility or a platform that has stopped being maintained. Only the third reliably requires a rebuild, and it is the least common of the three.
What a good brief looks like
A strong redesign brief is short and specific. It names the audience, the decision, the two or three metrics that define success, the URLs that must survive, and the constraints — brand, budget, dates, integrations — that are genuinely fixed.
It does not contain a list of competitor sites the founder admires. That list is useful later, in a design conversation, but as a brief it guarantees a project judged on taste rather than on outcome.
What to do in the two weeks before you brief anyone
Pull twelve months of analytics and Search Console data and write down the ten pages that produce the most qualified activity. Not the most traffic — the most activity that a salesperson would recognise as useful. That list is usually shorter and stranger than anyone expects, and it constrains the redesign productively.
Then sit in on three sales calls. Note every moment a prospect asks something the website should have answered. Those moments are your information architecture, handed to you for free.
Finally, ask the person who publishes content what takes them longest. If the honest answer is "anything", you have found a CMS project hiding inside a design request, and you have just saved several months of building the wrong thing beautifully.
Questions people ask about this
Structurally, every four to six years. But the useful answer is that a site owned by someone who improves it continuously rarely needs a redesign at all — it evolves. Sites that need dramatic rebuilds are usually sites nobody has been allowed to change.
For a B2B marketing site of twenty to fifty pages, budgets typically run from the mid five figures to low six figures depending on content migration, integrations and how much of the message needs rewriting. The variable is almost never the number of pages.
The decision a visitor should be able to make, the two or three metrics that define success with today’s baseline, the URLs that must keep their rankings, the constraints that are genuinely fixed, and the single person who approves direction.