Website redesign services usually bundle five things: a strategy and audit phase, UX/UI design, development and performance work, content and SEO migration, and launch plus post-launch support. That is the agency-marketing version of the package. What shows up in your contract is narrower, because copywriting, photography, SEO implementation, and post-launch fixes are the line items vendors cut first when a quote needs to look cheaper. If you are shopping for a website redesign right now, the gap between those two lists is where budgets blow up around week six.
What a redesign engagement is supposed to include
The five-bucket structure shows up across most agency pitches: audit and strategy, UX/UI, development and performance, content and SEO migration, then testing and launch. One agency roundup lists it almost exactly that way, with the caveat that exact scope “varies by project,” which is doing a lot of work in that sentence. A more buyer-focused breakdown gets specific about what each bucket needs to produce: a redirect map that covers every old URL, analytics and search markup wired in during the build (not after), and editorial training while the agency is still on the clock. Skip the redirect map and you do not get a cheaper redesign, you get a slower-motion traffic collapse that shows up in Search Console three weeks after launch.
Two items get cut first on tight budgets: original photography and instrumentation (analytics, schema, CRM hooks). Both are invisible on launch day and expensive to retrofit. A proposal that treats website redesign and custom websites as a single line item without naming these deliverables separately is a proposal you should ask more questions about before signing.
The honest tension here is that “redesign” gets sold as two different products wearing one name. Sometimes it is a visual restyle: new fonts, new colors, same site structure underneath. Sometimes it is a structural rebuild with new information architecture, content migration, and training. Agencies rarely volunteer which one they are quoting. Ask directly, because the price difference is the difference between repainting a room and moving the walls.
How to compare providers without comparing apples to orchards
A $38,000 bid migrating 50 pages with client-written copy and a $210,000 bid migrating 400 pages with a full content rewrite and a year of support are not two prices for the same job. They are two different jobs that happen to use the word “redesign.” The fix is a scope grid: rows for discovery, design, development, content migration, copywriting, accessibility testing, training, post-launch support, and hosting, one column per vendor. If a proposal stays silent on a row, treat that as excluded and make them say so in writing.
A few scoring tells are worth stealing. Accessibility work should name a standard and a verification method (WCAG 2.2 AA plus manual keyboard and screen-reader testing on key journeys) rather than “fully ADA compliant” bolted to an overlay widget, which courts know is not the same thing. Migration plans should state a page count and a pricing assumption for overages, something like “up to 400 pages, $X per additional 100.” And ask who does the weekly work. The senior who scopes the project and the junior who builds it are frequently different people, and a proposal that names the delivery team by role is telling you more than one that leans on “certified experts.”
This is also where I’ll land an opinion instead of hedging: only about a third of agencies pitching redesign work are exceptional at the craft, reliable on delivery, and staffed close enough to your time zone to understand your compliance and brand context without a translation layer. That is not an anti-outsourcing argument, there is real talent everywhere. It is a base-rate warning about picking from a redesign proposal stack on price and portfolio screenshots alone, because screenshots do not show you who answers the phone when the redirect map breaks.
What gets left out unless you name it
Buyer-side scope-of-work templates and agency marketing pages disagree on the default package, and the disagreement costs money. SOW templates list copywriting, photo and video production, SEO implementation, translations, and third-party integrations as things that need an explicit line item or they are out. One agency’s own commercial checklist agrees: demand copywriting, photo/video, translations, and integrations in writing, because a design-only purchase is the wrong buy when the real problem is technical debt.
Hosting and maintenance get the same treatment. They should appear as a stated ongoing cost, not something you assume is baked into the one-time redesign fee. And the accessibility substitution mentioned earlier shows up again here: an overlay plugin gets quietly swapped in for the manual testing that actual compliance requires. None of this is universally deceptive, it is just unstated by default, which means it is on you to ask. If you want the full range of what projects like this actually run, what website redesign services cost breaks down the line items that drive the spread between a five-figure site and a six-figure one.
Timeline, and who is on the hook for delays
Published timelines disagree by close to an order of magnitude, and the reason is that “redesign” covers a 3-day Wix template swap and an 18-month enterprise migration under the same word. A focused marketing site with content mostly ready runs 3 to 4 months. A mid-sized organizational site with real IA work and CMS development runs 4 to 9 months. A large platform rebuild with governance and procurement attached runs 9 to 18 months or more. One agency’s own process, run by a small studio taking on a handful of clients a year, compresses to 8 to 10 weeks because the senior who scopes the project is the senior who ships it, with no handoff between a sales team and a delivery team.
The delays, when they happen, are rarely technical. Content not ready, too many people with veto power and nobody with decision rights, scope creep after design is already approved, accessibility postponed to “phase two,” and a launch date tied to a marketing campaign with no schedule buffer. A proposal that only shows vendor tasks on the timeline is fiction. One that names client dependencies, content approval within five business days per round, for instance, is telling you the vendor has shipped enough of these to know where projects actually stall. We aim for a first working proof of concept within 7 days of kickoff specifically so clients can react to something real instead of a slide deck, and we land on-budget on about 94% of projects because scope gets locked before anyone starts designing, not after.
How to tell a rebuild from a reskin wearing a rebuild’s price tag
“Custom” on a quote can mean three different things: original design work on a purchased theme, a customized child theme with some hand-written code, or a from-scratch build with no theme underneath at all. All three get called custom. Only one of them is what most buyers think they are paying for.
The checks are mechanical, not aesthetic. View source on the live site and look for platform fingerprints: a Shopify theme_store_id of null versus a numeric ID tells you uploaded-and-customized from off-the-shelf. A WordPress theme folder matching a marketplace name like “astra” or “hello-elementor” instead of a business-specific name is a tell. So is a footer linking to features the site does not use, a leftover “Shop” or “Events” tab from a template nobody bothered to strip out. We had one e-commerce client who had genuinely outgrown their Shopify template. Instead of talking them into a from-scratch rebuild they did not need, we rebuilt the homepage and product page structure around the brand assets they already owned and got a 3x conversion lift without touching the platform underneath, which is the cheaper and correct answer more often than agencies with a bigger project to sell will admit.
The timeline is a tell too: a custom 5 to 10 page site built from scratch typically takes 6 to 12 weeks. A “custom” site delivered in one or two weeks was template customization, priced like something it was not. None of this makes template work illegitimate. Plenty of sites do not need a ground-up rebuild. It becomes a problem only when custom-build rates get charged for a logo swap, and the way you catch that is checking the source code, not the mockups. Before you sign anything, make sure SEO in a redesign is named as its own deliverable rather than folded silently into “development,” because a redirect map that only covers the homepage is, functionally, no redirect map at all.