The Website Relaunch Guide: redesign without losing your traffic

A website relaunch means replacing the platform, the structure, or both. Done properly it takes weeks to a few months, costs between $10,000 and $50,000 on the platforms we build on, and follows the sequence in this guide: strategy, content, platform, migration, launch, measurement. Done carelessly, it's the fastest way a healthy business loses years of accumulated search equity in a single afternoon. This guide exists to make sure you're in the first group.

Start by checking you need one, because a relaunch is the most expensive way to fix a website. There are three sizes of project hiding inside "we need a new website", and half the briefs we see are one of the first two wearing the third one's budget:

• Refresh: the strategy and structure are right, the execution is dated. New design on the existing bones. Weeks of work, the cheapest of the three, and often all that's needed
• Restructure: the platform is fine but the site's architecture no longer matches the business. Pages reorganised, navigation rebuilt, content remodelled, design largely retained. The middle path
• Relaunch: the foundations are wrong. The platform fights your team, the structure can't be saved, or the brand has genuinely moved on. New platform or structure, months not weeks, and the full checklist in this guide

The diagnostic is about who's suffering. If visitors find the site dated but your team publishes without friction, you're a refresh. If your team can't ship a landing page without a developer, or every change requires a workaround someone has to remember, the platform is the problem and you're headed for the full project. And if you're relaunching because a stakeholder is bored of the design, stop: boredom is the worst reason to touch working infrastructure, because the people who visit your site weekly are not the people who built it and stared at it for two years. Your buyers see it fresh.

Half the "we need a new website" briefs are smaller projects wearing a relaunch's budget.

One timing note worth the paragraph: the best moment to relaunch is when something true about the business changed, new positioning, new service architecture, a brand project just completed. A relaunch that expresses a real change pays twice. A relaunch that's just newer pays once, at full price.

The single best predictor of a smooth relaunch is whether content was modelled before design started. Decide what the site needs to say, structure it as reusable content types, and only then design the layouts that express them. Run it the other way, look first, and somewhere around week six the real content arrives and doesn't fit: headlines too long for the hero, case studies that each need a different layout because nobody agreed what a case study contains. The design flexes, the system fragments, and the editors spend year one fighting the layouts. This is the case for starting your redesign with structured content.

The first concrete task is the content inventory, and it drives everything later. List every URL on the current site with its traffic and rankings attached, then give each one a verdict:

• Keep: earning its place, carries over largely intact
• Improve: right page, underperforming content, rewritten during the project
• Merge: three thin pages that should be one strong one
• Retire: deliberately killed, with eyes open

That list becomes your new sitemap, your redirect map, and your writing schedule, all from one afternoon in a spreadsheet. Sites that skip it discover their most-trafficked page three weeks after deleting it, and by then the rankings have transferred to a competitor rather than to nowhere.

Then model the content types. For each kind of page the new site needs (service, case study, article, guide, definition), answer four questions: what entity is this page about, what fields does it always contain, what question does it answer for a visitor, and who updates it, how often. The answers are your content model, and the design brief largely writes itself from there, because every layout decision now has something concrete to lay out.

Design is how the content looks. Decide what the content is first.

There's a visibility dividend hiding in this chapter. Content modelled as fields is content machines can read: the answer engines that increasingly stand between you and your buyers lift claims from structured pages far more readily than from designed prose. A relaunch is the one moment you get to rebuild every important page answer-first at no extra cost, because you were rewriting them anyway. Take it. Retrofitting the same structure two years later is a second project at a second price. Liftable Content, chapter seven of our book Be the Answer, covers the page structure in full.

Choose for the five years of maintaining, not the three months of building. The question that settles most platform debates in ten minutes: who touches the site weekly after launch? If marketing owns it, publishing pages, changing copy, running campaigns, pick the platform they can drive unaided, with visual editing and instant publishing and no deploy queue. If the site is the product, with real application logic and engineers on staff, build it as software, with the team that maintains software.

In practice the decision lands in one of three places:

• Marketing site, marketing-owned: a visual platform. Webflow is our default, because copy ships in minutes, there are no dependency updates to babysit, and the constraint keeps the site simple enough for the people who own it. If you're weighing the platform choice itself, Shopify or Webflow comes down to what your team updates weekly.
• Commerce at the core: Shopify. Everything around the transaction (carts, payments, tax, shipping, fraud) is solved and maintained by a platform whose entire business depends on checkouts working. You trade some design freedom on deep commerce pages, and that trade is worth it the night something breaks at checkout and it isn't your problem
• Structured content feeding multiple surfaces, with engineering capacity: a headless CMS with a modern front end. Content lives as data, renders anywhere, and the site you're reading runs exactly this way. The real price is that you now maintain software, so only pick it if someone is genuinely on the hook for that

For what it's worth, we've made this call at the agency level: we build on Webflow and Shopify, and when a project genuinely needs custom, we say so on the first call and point you somewhere good. Fewer tools, fewer problems, and no incentive to talk you into the architecture we happen to sell.

Whatever you pick, confirm the boring things in week one, not launch week: who owns the hosting accounts, the domain registration, and the DNS, and can you log into all three today. Access chaos has sunk more go-lives than technology ever has, and it's always discovered at 5pm on launch day, because that's the first time anyone needed the login.

This chapter is the insurance policy, and it's the part most relaunches skip because nothing in it is visible in a design review. Search equity lives at specific URLs. Years of rankings, backlinks, and trust are attached to addresses, not to your brand, and if the addresses change without a forwarding plan, the equity transfers to nobody. Work through this list completely before anything goes live:

• Full URL inventory of the old site, with traffic and rankings attached (you built this in chapter two; now it earns its keep)
• Benchmark capture before touching anything: current rankings for your money keywords, monthly organic traffic by page, conversion numbers. You cannot prove recovery, or diagnose a problem, without the before picture
• A destination decision for every URL: kept as-is, redirected, or deliberately retired
• One-to-one 301 redirects for everything that moved. The lazy blanket redirect to the homepage tells engines the old pages are gone, and their equity mostly evaporates instead of transferring
• Backlink check: find the old URLs that other sites link to (your analytics referrers and any SEO tool will surface them) and make sure every one of them redirects somewhere sensible. These are the URLs where equity concentrates
• Title tags and the substance of the content carried over for the pages that earn your traffic. A redirect preserves the address; it can't preserve rankings for content that no longer exists on the other end
• Canonicals, sitemap, and robots.txt reviewed on the new build
• The staging site's crawl-blocking removed at launch. Shipping the live site with noindex still on is the classic self-inflicted wound, it happens to professionals every month, and it's why launch week has a checklist
• Analytics, consent, and conversion tracking re-verified on the new stack, with a test conversion fired and confirmed in the reports

A redirect map ships with the launch, not after it. Equity doesn't wait around while you tidy up.

None of this is glamorous. All of it is the difference between a launch and a loss, and it's the first thing to ask about when comparing proposals: an agency that can't talk you through its migration process fluently is planning to learn it on your traffic.

Launch on a quiet day, and never a Friday, because problems found on Saturday get fixed on Monday. Have the old site restorable, DNS access confirmed in writing, and one person clearly holding the checklist with the authority to delay if it isn't done. The run-sheet:

• Ship the redirect map with the launch itself, live within the hour, not "this week"
• Crawl the new site immediately with any standard crawler: broken links, missing pages, accidental noindex, orphaned URLs
• Submit the new sitemap in Search Console and watch the indexing report daily for the first two weeks
• Click every form, every purchase path, every conversion, on a phone, as a stranger, in a private browser window. Not the developer's test rig; the actual thing a customer touches
• Spot-check the top twenty old URLs from your inventory by pasting them into a browser and confirming each lands where the map says it should

Then hold your nerve. Expect rankings to wobble for a few weeks even when everything was done right, because the engines are re-evaluating a changed site, and that re-evaluation is noise, not verdict. What you're watching for is errors: a spike in 404s, pages excluded from the index, a redirect chain looping. Those are problems, and they're fixable fastest in week one while the before-picture is fresh. Small position shuffles are weather. The teams that panic and start reversing decisions in week two usually do more damage than the launch did.

The relaunch isn't finished at launch; it's finished when the numbers say the move paid. Keep the old inventory beside the new reality and check three things monthly:

• Recovery: are the pages that carried traffic back at or above their benchmark? Most recover within four to eight weeks when the migration was done properly; a page still down at week ten has a specific problem worth finding
• Conversion: is the new site converting at or above the old baseline? A prettier site that converts worse is a failed project wearing good photography, and you want to know early
• New ground: which new pages are earning attention that didn't exist before? This is the growth the project was for, and it's the number that justifies the invoice

Fix what the data flags, then bank the real dividend of a good rebuild: the site is now easy to change. That was the point of every platform and modelling decision in chapters two and three, and it only pays if you use it. Set a publishing rhythm, keep the content models tidy rather than bolting on one-off fields under deadline, and put someone's name on owning the site as a channel.

The pattern we see over and over: the businesses that get relaunch value are the ones still shipping improvements in month six, publishing into the structure they built, watching the new pages compound. The ones who exhale at launch and walk away are back in three years, asking for another relaunch, to fix the same neglect at the same price. The website you just built is a living asset. The whole project was the act of making it one; the return comes from treating it that way.