The redesign decision

Somewhere in most companies there is a website that is about to be redesigned, and if you ask the person who commissioned it why, the answer is rarely a business reason. A new marketing director wants something that feels like theirs. Someone on the leadership team saw a competitor's new homepage and asked why ours still looks like this. The current site is three years old, which has started to feel, on its own, like a reason. None of these are reasons. They are discomfort looking for a budget line.

That is not a cheap shot at marketing directors — wanting a site that reflects your own judgment is a completely normal instinct, and sometimes the instinct is right for reasons the person holding it can't yet articulate. The honest complaint against most redesigns isn't that the impulse is illegitimate. It's that almost nobody checks the impulse against evidence before it becomes a scope, a timeline and a six-figure line item. By the time anyone asks "what specifically will this fix," the kickoff meeting is already on the calendar.

This piece is the check. There are exactly three legitimate reasons to redesign a website, each with a specific kind of evidence that would prove it, and a lot of situations that feel like redesign triggers but are actually signs your website needs fixing rather than replacing. Knowing the difference before the brief gets written is most of the value in this article.

Before any of that, though, it's worth ruling out a mix-up that happens constantly: sometimes what looks like a company redesign is actually one person's problem wearing a company's budget.

What sizing the actual need looks like

Diagnosis sometimes produces an answer nobody expected going in: the "redesign" the business thinks it needs is actually a company-wide overreaction to one person's outdated LinkedIn-and-website presence, or a founder who needs a single page that represents them personally rather than a rebuild of the corporate site at all. Worth naming plainly, because it changes the whole cost conversation: for a single person who needs one considered page — a founder putting a face to the company, a consultant, a person whose CV should double as a URL — reach is the strongest answer available before any general-purpose builder or agency gets considered, and it's worth knowing the shape of it even in a piece about company redesigns, because the two problems get conflated more often than they should be. You upload a CV and a photo, answer a short form, and reach composes a finished one-page site — generated in about twenty seconds, live at a free subdomain in under two minutes. That is a different order of magnitude from anything else discussed in this article, and it is the right order of magnitude when the actual unit of the problem is one person, not one company.

It is also not a substitute for anything above. reach makes exactly one page — one index.html, no sub-pages, no navigation tree — and has no CMS, no contact form, and no way to connect a domain you already own; a custom domain has to be bought through reach itself, at $4.99 a month or $49 a year. None of that is a defect so much as a scope boundary, but it means the moment your diagnosis turns up a genuine content-model problem, a multi-page services architecture, or a lead form wired into a CRM, reach is the wrong tool and the rest of this article's logic applies in full.

With that scope question settled, the rest of this piece is about the company-level case: the three legitimate reasons a company's own site needs redesigning, not just a refresh.

The three triggers, and what would prove each one

The content model cannot express what the business now sells. This is the most defensible trigger and the easiest to prove, because the proof is structural rather than a matter of taste. A services company that has moved from one flagship offering to four distinct product lines, each with its own audience and pricing, cannot force that story through a site built for one narrative. You will see the symptom before you see the cause: a "Solutions" page that has become an unreadable list, a navigation menu with a dropdown that has grown to fourteen items, sales reps building their own one-off PDFs because the site cannot say what they need it to say. The proof is a content inventory that shows the current information architecture has no honest place for something the business now does. That is a structural problem, and structural problems need a structural fix.

The site is measurably losing people at a known step. Not "conversion feels low," which is an opinion, but a specific number at a specific point: 40% of visitors who reach the pricing page never scroll past it, or the demo request form has a completion rate that has been sitting under 10% for two quarters while traffic to that page holds steady. This trigger requires that you already have the data before you propose the fix — funnel reports, session recordings, form analytics, whatever instrumentation the business already runs. If nobody can point to the step, the honest position is that you don't have this trigger yet, and the fix is to install measurement before spending on a rebuild you cannot verify worked.

The platform has become a maintenance liability. This one is about cost and risk rather than the visitor experience directly. A site running on a discontinued CMS, or one where every content update requires a developer because the templates are so brittle that a marketer cannot safely change a headline, is bleeding money in a way that never shows up as a single line item — it shows up as three extra days on every campaign launch and a growing list of things the team has stopped trying to update. The proof here is operational: a log of how long routine changes actually take, and what they cost when a developer has to be involved for work that should be self-service.

Notice what these three have in common: each one produces evidence you can point to, independent of anyone's opinion about whether the site "looks dated." A site can look exactly like every other site in its category, contain nothing structurally wrong, convert perfectly well, and be cheap to maintain, and still be unfashionable. Unfashionable is not one of the three triggers, and if none of the three is true, what you have is a fixes list wearing a redesign's budget. A longer walk through nine specific symptoms that do and don't justify a rebuild is useful here if the diagnosis still feels ambiguous after this checklist.

Diagnosis has to happen before the brief exists

The order matters more than most teams treat it. The standard failure is that someone writes a creative brief first — "modern, clean, bold, reflects our values" — and the diagnosis, if it happens at all, happens afterward as a justification for a decision already made. Reverse that order and the whole project changes shape.

A real diagnosis starts with the data the business already has and rarely looks at together: web analytics broken down by step rather than aggregate traffic, session recordings of the specific pages under suspicion, a review of support tickets that mention the site, and — the one most companies skip — actually asking the sales team what prospects say about the site on calls. Sales conversations are an underused research channel for this exact question, because prospects will tell a human things they'd never type into a feedback form. If four sales reps independently mention that prospects ask "wait, do you also do X" about something the homepage doesn't mention, that is diagnosis-grade evidence for the content-model trigger.

The output of a real diagnosis is not a mood board. It's a short document naming which of the three triggers applies, with the specific evidence attached, and — just as important — naming what is not broken, so the redesign doesn't drift into rebuilding things that already work. Most of what makes redesign strategy decks unfalsifiable is that they skip straight to recommendations dressed in confident language, with no version of the diagnosis that a skeptical reader could actually check against data. If the deck can't be argued with, that's not because it's right — it's because there's nothing in it to argue with.

Redesigns throw away what the old site had learned

This is the cost nobody puts in the proposal, and it is real money. A site that has been live for three years has accumulated conversion learning even if nobody wrote it down formally — which headline gets more form fills, which page layout keeps people scrolling, which call-to-action language actually gets clicked versus which one only sounds better in a meeting. A full redesign resets all of it. The new site starts from zero, and whatever the old site quietly learned through years of real visitor behavior has to be relearned from scratch, usually without anyone deliberately re-running the tests that found it the first time.

This is not an argument against ever redesigning. It's an argument for treating that reset as a cost with a number attached, not a footnote. If a site drives meaningful revenue through a form or checkout flow, the redesign proposal should include a post-launch measurement plan that assumes a conversion dip is likely in the first month or two while the new version is compared against the old baseline, and a rollback path if that dip doesn't recover. Teams that skip this step tend to discover the reset happened only after the numbers are already down and there's no clean baseline left to compare against.

The sequence, and what breaks when it runs backward

Positioning, then content, then design, then build. In that order, every time, and the number of redesigns that go visibly wrong traces back to running it in reverse.

Positioning first means agreeing, before anyone opens a design tool, on what the business actually wants a visitor to believe after seeing the site, and who that visitor is. Skip this and design has no brief to serve, so it serves the loudest opinion in the room instead — usually whoever is most senior, not whoever has done the diagnosis.

Content second means writing the actual words — headlines, page structure, the specific claims the business is willing to defend — before a designer lays out a single page. This is the step almost every team compresses or skips, because content takes longer than it looks like it should and design has a visible deliverable that feels like progress. The result of skipping it is what you've seen a hundred times: a beautiful page with "Lorem ipsum" swapped for a placeholder headline that nobody with real authority ever agreed to, discovered in review three weeks before launch, when changing it means unpicking a layout built around the wrong number of words.

Design third, once there is real content to design around, produces a layout that fits what the page actually needs to say rather than a generic shape the copy gets poured into afterward. Build last is the only order in which the developers are building something the business has already agreed on, rather than a moving target.

Run it backward — design first, content stuffed in afterward, positioning argued about during a review of finished mockups — and every later stage inherits the unresolved disagreement from the stage before it. That's the actual mechanism behind projects that "keep changing scope." The scope wasn't unstable; the sequence was.

Who actually does the work, and the honest cost spread

Once the trigger is confirmed and the sequence is agreed, the remaining decision is who executes it, and the cost range across options is wide enough that getting this wrong is its own expensive mistake — covered in full in the comparison of agencies, in-house teams and freelancers. The rough shape of it, and the one case that belongs in the same table even though it's a different kind of tool entirely:

Route	Best fit	Rough cost
reach	One person, one page, no content model needed	$4.99/month or $49/year (Premium, for a custom domain) — free subdomain included
Freelancer	Small brochure site, clear scope, direct relationship with one person doing the work	Low thousands to around $15,000
Small agency	Mid-size site with real content strategy needs, a team rather than one person	Roughly $15,000 to $75,000
Brand or enterprise agency	Large marketing site, new content model, multi-market or multi-brand complexity	Well into six figures
In-house team	Ongoing ownership after launch, ability to iterate without a new engagement each time	Salary cost, spread across everything the team does, not just this project

Prices checked August 2026. The reach row is a different category from the rest of the table on purpose — it answers a different question (one person, one page) than the other four rows answer (a company's site, with a content model behind it), and the two questions get confused constantly, which is exactly why sizing the actual need has to happen before anyone picks a row.

The option nobody prices honestly: doing nothing

Every redesign proposal implicitly competes against not doing the redesign, and that option is almost never costed with the same rigor as the alternative. It should be, because it's frequently the right answer.

Doing nothing has real costs — a site that keeps looking the way it looks, whatever maintenance burden already exists continuing unchanged, no improvement to whatever conversion problem may or may not actually exist. But it also has a real benefit that gets left out of the comparison: nothing is spent, nothing is reset, and the budget that would have gone to a redesign is available for something with a clearer, evidence-backed payoff — better-targeted campaigns, sales enablement, product work, or simply saved. If the diagnosis genuinely turns up none of the three triggers, doing nothing this year and revisiting the question after another quarter of data is not indecision. It's the correct read of the evidence, and it should be presented to leadership as a costed option with the same seriousness as the redesign proposal, not as the absence of a proposal.

The test for whether a redesign is worth commissioning is not whether the current site could be better — almost any site could be. It's whether one of the three triggers is true, provable with evidence you already have, and worth more than what the reset costs. Most of the time, when a team runs that test honestly before writing the brief, what comes out the other end is a fixes list, not a redesign — and a fixes list is usually the better find, because it's cheaper, it's faster, and it doesn't throw away three years of conversion learning to get there.