Most people asking about website redesign services should not buy one.
About sixty percent of the redesign requests we get describe a problem that is fixable in plumbing, for a fraction of what a rebuild costs, and we turn most of those down. So before anything else, two things worth knowing: what a redesign engagement actually contains, and the four conditions that make one worth paying for. Both are below, and if none of the four describes your situation, the honest answer is that you need a repair rather than a rebuild and you can stop reading after the next section.
If you would rather find out which one you are in before reading two thousand words, run your site through the scanner. It takes about a minute and it checks performance, structure, functionality and design against what Google actually measures.
Here is what the work looks like when a rebuild genuinely is the answer.
We took on the Aussie Grill rebuild in early 2024 because Bloomin' Brands had outgrown the previous site three times in two years. Fast-casual rollout, sixteen locations live, two more openings every quarter, an international expansion into Saudi Arabia and Hong Kong. The previous site couldn't load fast enough on Google's mobile crawler to register Core Web Vitals as Good. The brand had pivoted toward chef-driven QSR positioning and the imagery still read as generic burger chain. Editorial workflow was broken because the CMS the previous agency had picked couldn't handle multi-region rollout cleanly.
Eight weeks on the build. One week on location for the photo shoot, actual food, actual location, art-directed and color-graded to the brand palette so nothing on the new site looked like stock. Total elapsed time from kickoff to live deploy: nine weeks. The site went from CWV-failing on mobile to CWV-Good across every template, with the locations module now driven by a real data layer instead of a developer hand-coding each new opening.
That experience is the reason this article exists. Most websites do not need a redesign. The ones that do need a redesign mostly don't get one that's worth the money. There's a narrow band where the rebuild is the right answer, and an even narrower set of conditions for what "rebuild done properly" actually looks like.
What website redesign services should contain
Before the triggers, the contents, because "redesign" is used to describe two very different products and the word does nothing to distinguish them.
A visual refresh restyles what exists. A redesign engagement replaces the structure underneath it. The second one contains all of the following, and a proposal missing any of them is the first one wearing the second one's name:
- Discovery and site structure, meaning how the pages are organised and how someone finds what they came for, agreed on paper before any design work starts
- A design system delivered as working code the developers actually build from, not a brand guidelines PDF nobody opens
- Original photography, art directed and owned by you outright, rather than stock
- The build, with your analytics, search markup and lead-tracking wired in during the build rather than bolted on afterwards
- Content migration with a tested redirect map, every old URL to its correct new equivalent
- Editorial training, with your team using the CMS while we are still on the engagement
- Pre-launch QA, Core Web Vitals confirmed per template, schema validated per entity type, forms triggered end to end into the CRM
The two line items that get cut first are the photography and the instrumentation, and they are the two that determine whether the site is still working in a year. The redirect map is the one that, if dropped, quietly destroys rankings that took years to accumulate.
When a redesign is the right answer
I want to start with the negative case because it's the most expensive mistake firms make. About sixty percent of the redesign requests we get are for sites that don't need to be rebuilt. The conversion problem the firm has identified is fixable in plumbing for a fraction of the cost of a rebuild. We turn most of those down, because spending a client's budget on a rebuild that doesn't move their number is the wrong work regardless of who books the revenue.
The legitimate triggers for a real rebuild break into four categories.
The platform is structurally hostile to performance, security, or instrumentation. WordPress on a sixteen-plugin page-builder Frankenstein. A Sitefinity install that's been customized so heavily that upgrades break it. A proprietary CMS the original agency built and won't transfer credentials for. We can't make these sites pass Core Web Vitals. We can't add proper schema. We can't reliably wire closed-loop attribution. The right answer is a rebuild on a platform we choose, not a patch on a platform we inherited.
The brand has fundamentally changed. Acquisition. Vertical pivot. New service line that needs equal billing with the old ones. Aussie Grill went from "international chain test" to "domestic fast-casual brand" between the previous site and ours. SayClean rebranded entirely to EBS Maintenance. Bloomin' Brands launched Aussie Grill as a separate consumer brand that needed its own digital presence rather than a tab on the corporate site. When the brand has changed at the level of positioning, the existing site can't carry the new story. Patching it costs more than starting over.
The IA is genuinely wrong. Practice areas buried three clicks deep. Service categories that don't match how prospects actually search. A homepage hero that explains what the company is rather than what it does for the visitor. This is rare and easy to mistake for a content problem instead of a structural one. The diagnostic question: would re-skinning the existing site fix the problem? If yes, it's content. If no, the IA itself is wrong, and that requires a rebuild.
The site is producing measurable revenue impact below a baseline you can name. This is the highest-bar trigger and the one I trust most. The firm can point to a real number, cases per month, revenue per visitor, conversion rate on a specific path, and that number has been below baseline for two consecutive quarters despite plumbing fixes. The site itself is the constraint. Now we're talking.
If a redesign request doesn't fit one of these four, we walk back into the audit conversation before the rebuild conversation. Most of the time the cheaper fix is the right fix.
What eight to ten weeks of "properly" actually looks like
The week-by-week below is the sequence we run on every bespoke build, regardless of platform. Sitefinity, WordPress, Next.js, .NET. The platform decision changes the implementation but not the rhythm.
Weeks 1 to 2: discovery and IA. Stakeholder interviews on the actual operations of the business, not the marketing version. We sit with the intake person. We watch the editorial team try to update the current site. We ask the sales team what questions prospects ask before they convert. The deliverable from this phase is the information architecture document, sitemap, page templates, content models, conversion paths. Approved on paper before any pixel decisions get made.
Weeks 2 to 3: typography and palette. One display family, one sans, one mono. Four color tokens. We refuse Tailwind's default ramp. We refuse stock typography pairings. The deliverable is a design system in Figma plus a real-code Tailwind config the dev team will actually ship. Not a PDF. Not a brand guidelines deck. Working tokens.
Weeks 3 to 4: imagery direction and shoot prep. This is where most agencies skip and ship stock photography. We don't. The week before the shoot we storyboard the hero compositions, source location and props, brief the photographer, and write shot lists for every page. The shoot itself runs the following week, typically three to five days on location depending on the scope. The Aussie Grill shoot was four days across two of their flagship locations, with food styling for the menu hero shots and location photography for the brand storytelling on the about page. Color-graded, retouched, and licensed to the client outright. They own the negatives.
Weeks 4 to 6: build. Component library, page templates, content modeling, editor UX. Every component gets a Figma + code pair so the design system is what ships, not a translation of it. We instrument analytics, schema, and conversion paths during the build, not as a Phase 2 afterthought. The closed-loop attribution wiring back into Google Ads, GA4, and the CRM is on the critical path of week six.
Week 7: content migration and editorial training. All page content moves to the new site. Old URL → new URL redirect map gets generated and tested. The editorial team takes over the CMS while we're still on the engagement, so the handoff isn't a surprise. We write the editorial playbook for them, voice rules, schema rules, image rules, and they edit live with us watching for the first two days.
Week 8: QA and instrumentation verification. Core Web Vitals confirmed Good on every template. Schema validated against Google's Rich Results test for every entity type. Forms triggered with traceable IDs to confirm routing all the way through to the CRM. Phone tracking confirmed source-attributed. Closed-loop conversion upload tested with a known case-management record. Cross-browser, cross-device, accessibility audit. Everything that has to be true before launch.
Weeks 9 to 10 (when needed): copy work, feedback rounds, and integration scope. This is the buffer the audit calls "longer if we're writing your copy from scratch, longer if feedback cycles need room." It's also where custom integrations land, the Salesforce sync that needed a custom adapter, the proprietary case management API that needs a webhook handler, the multi-region content sync. If the project is clean on copy and integration, we're done at week 8 plus the photo week. If it needs the buffer, weeks 9 to 10 is where it goes.
That's the timeline. Eight weeks for the build, plus one on location, plus one to two for copy and integration when needed. We do this for two to four clients a year. The reason we do it slowly is that we run the entire engagement with one senior, the senior who scopes the work is the senior who ships it. Two-handoff agencies break things at handoff. We don't have handoffs.
What you get for the eight to ten weeks
Concrete deliverables from a rebuild we'd actually be proud to ship:
- Working Core Web Vitals at Good across every template, monitored by Orbit for life of the engagement
- Schema markup for every page, validated against Rich Results, mapped to the right entity types
- Closed-loop conversion attribution from the site, through the form or phone, into the CRM, and back to the ad bidder weighted by case value
- A photography library that's yours outright, art-directed and color-graded to the brand palette, no stock licensing trail
- An editorial CMS your team can actually use, a WYSIWYG that previews like the live site, content models that match how you write, role-based publishing if the team needs it
- A redirect map from every old URL to the right new URL so the SEO equity transfers cleanly at launch
- 90 days of post-launch monitoring inside Orbit, with weekly check-ins on rankings, performance, and conversion deltas
That's the spec. Not the marketing version of the spec, the spec we hold ourselves to.
Keeping your rankings through the launch
This deserves its own section because it is the failure that hurts most and shows up latest.
A redesign changes URLs. Every URL that changes without a redirect pointing to its correct replacement is a page Google has to rediscover from scratch, and every inbound link pointing at it stops passing anything. Do this at scale and you lose rankings that took years to build. The damage is invisible on launch day and obvious about six weeks later, when the traffic report arrives and nobody can explain it.
What has to happen, in order:
- Crawl the existing site before anything changes, so you have a complete inventory rather than a list of the pages someone remembered.
- Map every old URL to its correct new equivalent. Not to the homepage. A redirect map that dumps everything on the homepage is technically present and functionally identical to having none.
- Test the map before launch, not after. This is the step that gets skipped under deadline pressure.
- Preserve or improve the schema per entity type, validated rather than assumed.
- Watch Search Console for the first six weeks specifically for crawl errors and coverage drops, because that is the window where a missed redirect is still cheap to fix.
Ask any firm you are considering what happens to your existing URLs, specifically. A vague answer is disqualifying. It is the single most reliable way to tell whether you are talking to people who have launched sites or people who have designed them.
What it costs
We do not publish a price for a rebuild, because the range across the ones we have run is wide enough that any figure here would mislead most of the people reading it. What drives the number, template count against page count, migration volume, integrations, whether the site needs a real data layer, is covered in what a website redesign actually costs, along with the eleven questions any real estimate has to answer before the number in it means anything.
What we will say is what a cheap rebuild is buying by not buying. A redesign priced well below the work is skipping the photography and shipping stock, or skipping the closed-loop attribution and leaving you with the same blind-bidder problem, or skipping the platform decision and inheriting the broken stack, or skipping the editorial training and losing the gains by month four. Those are the four line items that get cut, in roughly that order, and each one is invisible at launch.
If the budget is not there for the full engagement, the right answer is not a smaller rebuild. It's the audit conversation, the plumbing fixes, the GBP reactivation, the ninety-day cadence on the existing site. We can scope that separately, or recommend someone who will. Money spent on a rebuild that ships rough does less than the same money spent on the operational layer.
The two-to-four-slots-a-year rule
We take on between two and four bespoke web platform builds a year. Total. We do not run parallel builds. The senior who scopes is the senior who ships. We turn down work when we're full because the math of running two builds in parallel is the math of shipping both rough and burning out the team that scoped them.
This is a real constraint. If you're talking to us about a rebuild, the next slot opens when the next slot opens. Sometimes that's six weeks out. Sometimes it's eleven months out. We'll tell you honestly which it is during the first conversation, and we'll tell you what to do in the meantime if your business need is urgent.
If the timeline doesn't work for you, that's a real reason to hire someone else. We're not the right shop if the rebuild has to ship in three weeks. We're the right shop if you want it shipped properly and the business case justifies the wait.
What we won't redesign
I'll close with the negative space, because it's where most of the agency-rebuild conversation goes off the rails.
We won't rebuild a WordPress site that's running an Elementor or Divi page-builder stack onto another WordPress site that's running an Elementor or Divi page-builder stack. The platform decision is the engineering decision. If you're committed to the same stack that produced the current problem, the rebuild won't change the outcome.
We won't ship a site that uses stock photography for the brand-defining shots. Custom photography is on the critical path, not a budget option. If the project budget can't absorb the shoot, the project budget can't absorb the rebuild. Pick the version where the visual brand is yours.
We won't run the engagement without the closed-loop attribution. The attribution wiring is what makes the site optimizable after launch. Without it, we're shipping a brochure that you'll watch decay over twelve months.
We won't run parallel builds, and we won't quietly hand the engagement to a junior. The senior who scopes is the senior who ships. If we're full when you call, we're full.
If those constraints land as the right kind of strict for your situation, the next step is a 30-minute conversation about your specific case. We do this audit free for two to four firms a quarter. If we find work, we'll scope it. If we don't, you'll have a punch list for whoever does the work.