A redesign should be decided on signals, not on fatigue
Most redesigns are launched because the site “looks old”. That is a legitimate but insufficient reason—and, above all, impossible to quantify. A project decided on that basis produces a nicer-looking site that converts exactly the same, and the business wrongly concludes that redesigning is pointless.
A redesign is justified when it fixes a measured problem. The site does not convert, it does not rank, it costs too much to evolve, it does not display properly on mobile, or its technical foundation has reached end of life. Each of these reasons can be confirmed with numbers, and each calls for a different response.
This distinction drives the budget. A site that converts poorly because its message is unclear can be fixed with an editorial rewrite for a few hundred euros. A site whose architecture blocks any evolution must be rebuilt. Confusing the two means spending ten times too much—or ten times too little.
Before any decision, record three numbers: the number of monthly visitors from search, the number of enquiries received via the form, and mobile load time. These three values indicate whether the problem is visibility, conversion, or technical. Without them, the redesign is decided blind.
Signal one: the site generates no enquiries
The most common symptom—and the most misdiagnosed. A site that generates no contact can suffer from two opposite issues, and confusing them leads you to rebuild what was working.
First case: nobody comes. Organic traffic is under one hundred monthly visits, and the site does not appear for any commercial query. The problem is visibility, and an aesthetic redesign will change nothing. The response is editorial and architectural.
Second case: people come and leave. Traffic exists, pages are viewed, but the form stays empty. The problem is conversion: unclear message, lack of proof, invisible pricing, a buried call to action. These issues are often fixed without a redesign, by reworking the service pages.
The diagnosis comes down to a quick look at Search Console and a measurement tool. It takes an hour and prevents committing a budget to the wrong problem.
Signal two: mobile speed is outside the thresholds
A site whose main page takes more than four seconds to display its main content on mobile loses visitors before it even speaks to them. The benchmark threshold is the largest visible element displaying in under two and a half seconds, measured on real visits.
This signal does not, on its own, justify a redesign. Many slow sites can be fixed by changing hosting, compressing images, and cleaning up plugins, as detailed in our article on Core Web Vitals. These interventions take a few days and produce an immediate effect.
It becomes decisive when optimisation hits a ceiling. A site built on an overloaded generic theme, with twenty nested plugins and a stack of containers, resists any optimisation. You gain three tenths of a second, lose as much at the next update, and the threshold remains out of reach.
Signal three: every change requires a developer
A professional site must allow a non-technical person to add a page, edit text, or publish news. When every change requires a billed intervention, the site stops being a tool and becomes a constraint.
This situation has two causes. Either the site was built without a usable editing interface, which happens with poorly thought-out custom builds. Or it was built with an interface, but so heavily customised that any change breaks something.
The real cost of this rigidity is measured in missed opportunities. A business that gives up publishing because it is complicated also gives up ranking. Over three years, the lost revenue far exceeds the price of a redesign.
Signal four: the site no longer reflects the business
Services have changed, the target has shifted, the business has moved upmarket—and the site still talks about what it did five years ago. This gap produces two negative effects at the same time.
It attracts the wrong prospects. An outdated positioning captures off-target enquiries that the team processes and then declines, wasting time. It repels the right ones: a client looking for a premium service does not recognise themselves in a generalist presentation.
This signal is the hardest to see from the inside, because the team knows the real business and reads the site while projecting it onto the text. The test is to have someone external read the homepage and ask them what the business sells, to whom, and at roughly what price. The answers are often revealing.
Signal five: the architecture prevents you from ranking
A site where all services fit on a single page can only rank for one query. A business that sells four distinct services needs four distinct pages, each targeting its own search intent.
The same problem occurs in reverse with sites that have accumulated dozens of pages with no hierarchy. Several pages target the same query, compete with each other, and none progresses. This cannibalisation is fixed by reworking the architecture using the method explained in our article on the semantic silo.
When the URL structure itself is the issue, the rework becomes a redesign. Unreadable URLs, excessive depth, an organisation modelled on the internal org chart rather than customer intent: these issues are fixed by rebuilding, with a carefully planned redirect map.
Signal six: the technical foundation is end-of-life
A theme abandoned by its author, a PHP version that no longer receives security fixes, plugins whose development has stopped. These three situations produce a site that works today and will become impossible to maintain tomorrow.
The tipping point usually comes with a mandatory update. The host enforces a new PHP version, the theme is not compatible, nobody can fix it, and the site breaks. This is handled poorly in an emergency, even though it was predictable months in advance.
Monitoring these three elements is part of a serious WordPress maintenance contract, and that is precisely the point of the monitoring: turning an unplanned outage into a planned project.
Signal seven: the site does not inspire trust
In a market where the client compares three providers, perceived credibility is the tie-breaker. It depends on elements that have nothing to do with graphics: full contact details, up-to-date legal notices, documented work, identified authors, authentic reviews.
A site that shows stock images, no address, generic text, and no proof creates doubt that even the most polished design cannot dispel. These signals fall under the evaluation framework we detail in our article on E-E-A-T, and they matter as much to visitors as to search engines.
The good news is that, in most cases, this signal can be fixed without a redesign. Photographing your team, documenting three projects, and completing your legal pages takes a few days and produces an immediate effect.
Signal eight: the competition has pulled ahead
A simple check is enough: type your three main commercial queries and see who appears. If the competitors shown have sites that are clearly more complete, faster, and better structured, the gap will not be closed with a touch-up.
This is the only signal that imposes a timeline. SEO is a relative game, and a competitor who redesigned their site a year ago has a lead that grows, because a recent site accumulates faster. Waiting two more years makes catching up more expensive.
Signal nine: maintenance costs more than rebuilding
This is the most objective signal—and the least looked at. Add up, over twelve months, all billed interventions on the site: fixes, adjustments, troubleshooting, complicated updates, workarounds. Compare that total to the price of a new site.
On the projects we take over, this calculation is regularly unfavourable to the existing site. A business that spends three thousand euros a year patching a site built on a compromised base would have paid off a rebuild in eighteen months, with a result in a different league.
Redesigning without losing rankings: the method
A poorly prepared redesign can wipe out in three weeks what was built over three years. The risk is real, documented, and entirely avoidable with a method.
Before: the inventory
Export the full list of existing URLs, with each page’s organic traffic and rankings. This list is the reference for the entire project. It identifies the pages to keep at all costs, those to merge, and those to drop.
Also record the external links received, page by page. A page that receives links from other sites carries value that must be preserved, even if its content seems secondary.
During: the mapping
Every old URL must have a destination on the new site. A line-by-line mapping, validated before go-live, prevents the most costly mistakes. Pages with no direct equivalent should point to the closest semantic page—never to the homepage, which is a lazy redirect that is misinterpreted.
Keep URLs that work when possible. Changing a URL structure because it looks nicer, on pages that rank well, is a risk taken with no upside.
After: monitoring
The four weeks after the switch-over are decisive. Monitor crawl errors, index coverage, and the rankings of key pages daily. A forgotten redirect is fixed in ten minutes if spotted the next day, and costs months if spotted in January.
Also plan for a temporary dip. Even a perfectly executed redesign causes a wobble of two to six weeks, while Google recrawls and recalculates. Announcing this phase prevents panic and rushed decisions.
The five mistakes that ruin a redesign
All are common, and all can be prevented upfront.
- Going live without a redirect plan. The most costly and most common mistake. Every old URL that is not redirected is a lost page.
- Leaving the staging environment indexable. A preproduction site accessible to search engines creates duplicate content and can sometimes rank instead of the real site.
- Forgetting to remove the indexing block at go-live. The opposite of the previous one—and it makes the entire site disappear for weeks.
- Reducing content. A visually cleaner site often results in shorter pages, which lose the keywords that were ranking them.
- Redesigning during peak season. The post-switch-over wobble then hits at the worst possible time. A redesign should be planned during a lull in activity.
What a redesign will never fix
Some redesigns fail on the initial diagnosis rather than on execution, because the real problem was elsewhere than the site. Three situations recur regularly and should be named before committing a budget.
The first is a lack of demand in the target market. A perfectly built site for a service nobody searches for will generate no more than a mediocre site for the same service. The preliminary check is to verify that searches actually exist, and that they are not already captured by players whose position is out of reach in the short term.
The second is an offer problem. When the price, timeline, or scope does not match what the market accepts, the site simply amplifies the mismatch by exposing more widely an offer that does not convince. Enquiries come in, quotes go out, nothing gets signed. Redesigning the site does not change that equation.
The third is how enquiries are handled. A form that receives fifteen contacts per month and a business that calls back three days later loses most of what the site produced. On the projects we monitor, improving response time often generates more revenue than any technical optimisation—and costs zero euros.
Saying this is part of the job. An agency that accepts every redesign project without questioning these three points sells a service that will not produce the expected result, and the client will conclude that the web does not work for their business.
Content: the underestimated line item in any redesign
On projects that slip, the cause is editorial in the vast majority of cases. Development moves forward, mock-ups are approved, and the site waits six weeks for its copy.
The volume always surprises. A twenty-page site represents around fifteen thousand words to produce, proofread, and validate—excluding service descriptions, legal notices, and short navigation text. This work falls to people who already have a full-time job, and it only moves forward at the end of the day.
Reusing old content is a common false economy. Copying existing text into a new design reproduces exactly the flaws that prevented the site from converting. If the message was unclear, it will remain unclear. A redesign is precisely the opportunity to rework that message—and it is the part that produces the most measurable impact.
Photography deserves its own line. A premium site built on stock images immediately loses credibility, and a professional shoot represents a budget that appears in no development quote. This expense is decided at the start, not when the site is ready to go live.
Best practice is to appoint a content owner, set a delivery date per batch, and write down what happens in case of delay. These three lines in the specifications avoid the awkward conversation in month three.
Measuring the outcome of a redesign
A redesign without measurement is a gamble. Three indicators are enough, to record before the switch-over and then at three, six, and twelve months.
Organic traffic first, distinguishing the pages that brought it in before. An overall increase sometimes masks the collapse of a strategic page offset by gains elsewhere—and that information matters.
Then the number of enquiries received, relative to the number of visits. This is the only indicator that directly links the site to the business. A redesign that halves traffic but triples the enquiry rate is a success, even if the overall dashboard seems to decline.
Finally, the quality of enquiries, measured by the conversion rate into quotes and then signed deals. This data lives in sales follow-up, not in the analytics tool, which explains why it is rarely connected back to the site. Yet it is what tells you whether the new positioning attracts the right prospects.
Full redesign or partial rework
Between patching and a full rebuild, there is an intermediate path that is often more relevant—and rarely proposed.
A partial rework consists of keeping the technical foundation and redoing what causes problems: service pages, architecture, speed. It fits when the base is sound but poorly exploited, which is more often the case than one might think.
A full redesign is required when the foundation itself is the issue: abandoned theme, incompatible plugins, unreadable URL structure, inability to work in the code. In that case, every euro spent on the existing site is wasted.
The sorting criterion is technical and can be determined in a few hours of review. That is the purpose of a technical SEO audit, which quantifies both scenarios before a decision rather than assuming them.
What to prepare before consulting
A redesign is prepared like a new build, with one additional difficulty: the existing site. The elements to gather before contacting a provider fit into a short list.
The export of URLs with traffic and rankings. The list of access credentials and their holders, often incomplete and time-consuming to reconstruct. The inventory of what works today and must be kept—something redesigns handle poorly because the focus stays on what is wrong.
Finally, quantified objectives. What must the new site produce that the old one does not—and how will it be measured. This question, asked when drafting the website specifications, turns an aesthetic project into a measurable one.
The switch-over schedule
The go-live day is chosen. A Tuesday morning is better than a Friday evening, because an anomaly detected the next day can be fixed during the day rather than remaining in place all weekend.
Also plan a publication freeze of a few days around the switch-over. Adding content while redirects are being put in place blurs the diagnosis when a page does not get indexed: you no longer know whether the problem comes from the migration or from the new content.
Key Takeaways
Nine signals justify a redesign, and none is aesthetic. No enquiries, speed outside thresholds, editing rigidity, positioning mismatch, blocking architecture, end-of-life foundation, trust deficit, competitive lag, excessive maintenance cost.
Three of them can often be fixed without a redesign: speed, trust, and conversion can be addressed with targeted interventions at a fraction of the cost. The other six are structural and call for a rebuild.
Whatever the scenario, preparation determines the outcome. The inventory of URLs with their traffic, the line-by-line mapping to the new pages, and daily monitoring during the month after the switch-over. These three steps take a few days and prevent the only truly irreversible mistake in a redesign project. Projects we have delivered under these conditions are featured in our work, and to gauge a budget, the price simulator provides an order of magnitude in a few minutes.
One last point on internal communication. A redesign changes the habits of the entire business: new editing interface, new URLs to share, new forms whose notifications arrive elsewhere. Planning a one-hour onboarding session with the relevant people prevents the new site from being used like the old one—or worse, not being used at all.
