A failed redesign rarely looks like a failure on launch day. The new site looks good and colleagues like it. The numbers start slipping a few weeks later.
Enquiries fall. Search traffic halves. The site feels heavy on a phone. The English version still shows old content. Most of these problems were built in months before launch, during planning and procurement.
This article sorts the most common redesign mistakes into ten patterns. For each one it covers warning signs, causes, prevention and recovery. We do not name real companies as failures; we only cite published research and official guidance.
Main sources for this article
- Google Search Central guidance on site moves, multilingual sites and indexing control
- web.dev Core Web Vitals thresholds and the Vodafone case study, HTTP Archive Web Almanac 2024, WebAIM Million 2026
- W3C WCAG 2.2, the European Accessibility Act (Directive 2019/882), the Dutch government's business portal business.gov.nl
- The McKinsey and University of Oxford study of IT projects (2012), plus guidance from Japan's Cabinet Office, Fair Trade Commission and IPA
| What this article answers | Short answer |
|---|---|
| What counts as a failed redesign? | Results drop, budget and time overrun, no one runs it, or it breaks rules (section 1) |
| Which mistakes are most common? | Ten patterns, from no goal to weak contracts (section 2) |
| How do you prevent each one? | Signs, causes, prevention and recovery in each section (sections 3–12) |
| What if it already happened? | Stop the bleeding, isolate the cause, fix, prevent repeats (section 13) |
| How do Japan and Europe differ? | Legal deadlines, languages and contract types differ (section 14) |
1. Defining a "failed" redesign
Failing to launch is not the only failure
Far more common than a collapsed project is a site that launched on time but did not achieve its purpose. This article treats failure as four types; if any applies, the redesign shows signs of failure.
| Type of failure | What it looks like | When it shows |
|---|---|---|
| Results drop | Fewer enquiries, sales or search visits than before | 2–12 weeks after launch |
| Budget and time overrun | Repeated change quotes, launch keeps slipping | Mid-project onwards |
| No one runs it | Updates stop, no owner | 3–6 months after launch |
| Breaks the rules | Accessibility or cookie complaints | Any time |
Failures usually surface long after their causes were introduced, which is what makes them hard to see.
Large IT projects show the same pattern
In 2012 McKinsey and the University of Oxford studied more than 5,400 IT projects. Large projects with initial budgets above USD 15 million ran on average 45% over budget and 7% over time, and delivered 56% less value than predicted. The causes it lists, from missing focus and shifting requirements to skills and execution, apply almost unchanged to much smaller website projects.
2. The ten failure patterns at a glance
Warning signs and where each one starts
Here are the ten patterns, with the early signs that reveal them and the stage at which they are introduced.
| Failure pattern | Early warning sign | Introduced at |
|---|---|---|
| 1. No clear goal | No one can say in one sentence what will improve | Planning |
| 2. Design-only change | Meetings are mostly about colours and photos | Planning to design |
| 3. Content left until last | No owner or deadline for copy | Design |
| 4. Tracking gap | No list of key actions to measure | Build |
| 5. Search traffic drop | No URL mapping | Design to launch |
| 6. Slower pages | No speed target in the requirements | Design to build |
| 7. Legal gaps | No one knows the accessibility standard | Planning to build |
| 8. Broken languages | English is "translate later" | Design |
| 9. No owner after launch | No one assigned for post-launch work | Planning |
| 10. Weak procurement | Quote says only "complete website" | Procurement |
The table shows that many of the ten are introduced during planning and design. The figure below maps where each is introduced and when it surfaces.
3. Mistake 1: Starting without a goal
Warning signs and causes
This is the most common mistake and the starting point for many others. The sign is simple: no one can say in one sentence what the redesign will improve.
"It looks dated" or "a competitor relaunched" are triggers, not goals. Without a goal, decisions rest on the taste of whoever is in the room.
Prevention: turn the goal into numbers
Before you start, turn the goal into numbers you can measure. The table pairs common goals with the numbers that measure them.
| Common goal | Number to measure | Save before launch |
|---|---|---|
| More enquiries | Monthly enquiries, form completion rate | Last 12 months of counts |
| Stronger hiring | Applications, careers page views | Last 12 months of applications |
| More European customers | Traffic and enquiries by country | Figures by country |
| Update content in-house | Days needed per update | Current days and request flow |
| Faster pages | LCP, INP, CLS on mobile | PageSpeed Insights results |
The right-hand column matters most. If you do not save the old site's numbers, you cannot tell whether the new site is better.
Recovery
If you realise after launch that the purpose was vague, it is not too late. Pick one or two target numbers now and pull what you can of the old site's figures from Search Console and GA4, then re-check the site with the measurements in ten signs your website needs a redesign.
4. Mistake 2: Changing only the design
Warning signs and causes
Most meetings end with discussions of colours, fonts and the homepage photo. That is the sign. Visuals are easy for everyone to comment on, so the conversation gravitates there.
What drives enquiries, however, is usually structure and paths, not looks. Where do visitors arrive, what do they read, and where do they get in touch? If that flow is the same as before, new visuals will not move the numbers much.
Prevention: separate looks from what drives results
Split the design discussion into parts that drive results and parts that do not. The table ranks common redesign elements by their effect on results.
| Element | Effect on results | How to check |
|---|---|---|
| Page structure (information architecture) | Large | Tree testing: can people find things? |
| Enquiry path and forms | Large | Steps and drop-off to completion |
| Body content and case studies | Large | Search terms and reading time |
| Page speed | Medium to large | Core Web Vitals |
| Colours, fonts, photos | Small to medium | Readability and contrast |
The elements that matter most are the least visible. If you change the look, review the structure and paths at the same time.
Recovery
If you already launched a looks-only redesign, review the paths on your top ten enquiry pages; a full redo is not needed. Shorter forms, a contact button within the body text, and links from case studies to the contact page often work better than another redesign.
5. Mistake 3: Leaving content migration until last
Warning signs and causes
Design and build are moving, yet no one owns the copy and there is no deadline. That is the sign.
Content migration means checking each page and deciding to keep, update, merge or remove it. It looks simple but scales with page count. Left until last, it creates a pile of manual work just before launch, which either delays launch or forces a poor migration.
Prevention: start with an inventory
Make a content inventory the first task. Nielsen Norman Group recommends first listing every page in an inventory and then auditing its quality.
The inventory lists title, URL, owner, topic, last update and metadata. Adding search traffic and enquiry figures makes decisions easier.
| Decision | Typical condition | Old URL handling |
|---|---|---|
| Keep | Gets traffic or enquiries and is accurate | 301 to the new URL |
| Update | Gets traffic but is outdated | Update, then 301 |
| Merge | Several similar pages | 301 to the merged page |
| Remove | No traffic and no remaining purpose | 410 or 404; do not redirect to unrelated pages |
Avoid redirecting every removed page to the homepage. Google says redirects to irrelevant pages may be treated as soft 404s, meaning pages that effectively do not exist.
Recovery
If you launched with many old pages missing, check the URLs returning 404 in the Search Console Pages report. In order of past traffic, add redirects to new pages or restore the content.
6. Mistake 4: A gap in tracking
Warning signs and causes
After the redesign someone says "enquiries seem lower", but no one can check it with data. That is a tracking gap.
The usual cause is that the tag was reinstalled on the new site, but key actions such as a completed enquiry are no longer recorded. When the form changes, the old tracking setup stops working.
Watch GA4 retention too: standard properties can keep user- and event-level data for 2 or 14 months, and explorations are affected by this setting. If no one checked it, older data may be gone when you want to compare.
Also, Universal Analytics stopped processing data for standard properties on 1 July 2023. If the old site only had the legacy tag, nothing after that date was recorded at all.
Prevention: list the actions to measure before launch
Before launch, list the actions to measure and confirm each one on staging.
| Item to check | Before launch | Impact if missed |
|---|---|---|
| Key actions (e.g. completed enquiry) | Use the same event names old and new | No old-new comparison |
| GA4 retention period | Confirm it is set to 14 months | No year-on-year comparison |
| Search Console ownership | Verify new domains and URLs | Post-launch errors go unnoticed |
| Cookie consent setup | Confirm no tracking before consent | Legal breach, or no data at all |
| Old site's numbers | Export 12+ months and keep them | No baseline to compare |
The fourth row matters most for European sites: tracking before consent is a legal problem, and a bad setup can stop measurement entirely.
Recovery
When you find a gap, fix the tag that day. Lost data will not come back, but you can often recount enquiries from email or CRM records.
Lining up the recounted enquiries by month before and after launch at least tells you whether they fell. To automate monitoring, see automating Search Console and GA4 monitoring.
7. Mistake 5: A sudden search traffic drop
Warning signs and causes
One or two weeks after launch, search traffic visibly drops.
Google's site move guide says visibility in Search may fluctuate temporarily during a move, and that it can take a few weeks or more for new URLs to show. Some movement is expected.
Large drops, however, are usually configuration gaps rather than fluctuation. The four common causes:
| Cause | Where to check | Fix |
|---|---|---|
| Missing redirects | Crawl the list of old URLs | 301 to the matching new URL |
| Leftover robots.txt or noindex | robots.txt and page meta tags | Remove staging blocks |
| Wrong canonical or sitemap | Page source and sitemap URLs | Point to live new URLs |
| Much less body content | Compare old and new top pages | Restore removed information |
Google's guide explicitly reminds site owners not to forget to remove noindex or robots.txt blocks that were only needed for the migration. Staging settings reaching production is that common.
Prevention: build the URL map first
The core of prevention is a redirect map from old to new URLs. You can build it from the same inventory as section 5.
Google advises keeping redirects as long as possible, generally at least one year. Redirect types, configuration examples and the Search Console Change of Address tool are covered in our article on 301 redirects.
Recovery: respond according to the size of the drop
Respond differently depending on how far traffic has fallen. The figure shows the rule of thumb DevNJ uses. Because of seasonality, compare with the same period last year or the average of the four weeks before launch.
When the drop is large, work in this order:
- In Search Console, list the 20 pages with the largest click losses
- Open each old URL and confirm it 301-redirects to the right new URL
- Check robots.txt and the noindex and canonical tags on those pages
- Submit fixed URLs for recrawl with the URL Inspection tool
- Record clicks for the same 20 pages weekly to track recovery
As a reported case, Search Engine Journal studied 892 domain migrations and reported an average of 523 days to regain previous traffic. This is an average of reported cases, not a rule, but moves that also change the domain can take long to recover.
8. Mistake 6: Slower pages
Warning signs and causes
The new site is slower than the old one. This is surprisingly common. Large photos and video, animation, and third-party chat or ad tags pile up without a check.
Teams test on fast connections and laptops, so they miss it. Real users visit on phones on the move.
Prevention: put pass marks in the requirements
web.dev defines "good" Core Web Vitals at the 75th percentile of real users as follows. Write these straight into the redesign requirements and acceptance criteria.
| Metric | What it measures | Good threshold |
|---|---|---|
| LCP | Time until the main content shows | 2.5 seconds or less |
| INP | How fast the page responds to input | 200 milliseconds or less |
| CLS | Layout shift while loading | 0.1 or less |
Slowness is not rare. According to HTTP Archive's Web Almanac 2024, only 43% of sites on mobile and 54% on desktop passed all three metrics. Around half of sites do not pass.
A published case where speed moved results
web.dev describes a Vodafone case published in 2021. A version with a 31% better LCP saw 8% more sales, a 15% higher lead-to-visit rate and an 11% higher cart-to-visit rate.
This is a large e-commerce case, and a small business site will not necessarily see the same percentages. It does show that speed affects business numbers, not just perceived quality.
Recovery
If the site got slower, measure your top pages with PageSpeed Insights and tackle the biggest causes first. Three fixes usually help most:
- Replace large hero images with correctly sized, modern formats
- Remove unused third-party tags, such as those from old campaigns
- Replace above-the-fold video or sliders with a still image
9. Mistake 7: Accessibility and legal gaps
Warning signs and causes
"Accessibility next time." "We'll sort the cookie banner after launch." Postponements like these are the sign. Both cause rework in design and build if added later.
WebAIM's February 2026 check of the home pages of the top one million sites detected WCAG failures on 95.9%. The most common was low contrast (83.9%), followed by missing alternative text and missing form labels. A redesign is exactly the moment to fix these in one go.
Deadlines are piling up in Japan and Europe
In recent years, deadlines affecting websites have followed one another in both Japan and Europe. The figure lists the main dates relevant to redesign requirements.
The table sorts them by Japan, the EU and the Netherlands.
| Region | Law or guidance | Effect on the website |
|---|---|---|
| Japan | Amended Act on the Elimination of Disability Discrimination (from 1 April 2024) | Reasonable accommodation mandatory for businesses; web is part of the environment |
| Japan | External transmission rules, Telecommunications Business Act | In-scope businesses must disclose data sent to third parties |
| EU | European Accessibility Act (applies from 28 June 2025) | Consumer e-commerce, booking, banking and more |
| Netherlands | National supervision of the Act | ACM supervises e-commerce; an accessibility statement is required |
| Netherlands | Cookies and GDPR | AP checks banners; tracking before consent may be unlawful |
Under the European Accessibility Act, services provided by microenterprises with fewer than 10 staff and annual turnover or balance sheet total of no more than EUR 2 million are exempt (Directive 2019/882, Article 4). A Japanese company can still be in scope if it offers services such as e-commerce to EU consumers.
Prevention: write the standard into the requirements
For accessibility, the usual approach is to write WCAG 2.2 level AA from W3C into the requirements. Without a stated standard, an agency cannot quote for it.
For cookies, add two pre-launch checks: is rejecting as easy as accepting, and do analytics or ad tags stay off until consent? The Dutch AP says it scanned the cookie banners of 10,000 sites and sent warnings to more than 200.
Recovery
If gaps are found after launch, fix form labels, keyboard operation and contrast first. W3C's WAI publishes evaluation approaches from quick checks to expert reviews.
10. Mistake 8: Broken languages
Warning signs and causes
Only the Japanese version is new while English still runs on the old site. English URLs do not match Japanese ones. The language switch sends you to the homepage. These are broken languages.
The cause is treating multilingual work as "translate after Japanese is done". URL structure per language, the switching mechanism, translation ownership and update flow must be settled in design, or aligning them later becomes hard.
Google's guidance and common breakages
Google gives clear guidance on multilingual sites. The table pairs common breakages with that guidance.
| Common breakage | Google's guidance | Prevention |
|---|---|---|
| Auto-redirect by browser language | Do not redirect based on a guessed language | Let users choose |
| One URL switches language | Use different URLs per language | Separate /ja/ and /en/ |
| Only boilerplate translated | Body in one language makes a poor experience | Translate the body too |
| Translation pairs out of sync | Use hreflang to link language versions | Link versions to each other |
| Switch jumps to homepage | — | Link to the same page in the other language |
The first row is especially common on European sites.
Prevention: make language a unit of design
In a multilingual redesign, list pages per language and decide which page translates which at the start. That mapping feeds both hreflang and the redirect map.
Design and pitfalls of automating translation are covered in automating translation for a bilingual site.
Recovery
If languages are broken, stop auto-redirects first. Then rebuild the mapping between language versions and align hreflang and the switch links.
11. Mistake 9: No one runs the site after launch
Warning signs and causes
Launch is treated as the finish line, and no one has decided who does what afterwards. It shows as news that stops for six months, no new case studies, or form errors no one notices.
The cause is that the redesign gets a project budget while ongoing work is left as someone's side job.
Prevention: put roles and frequency in a table
Before launch, put post-launch roles and frequency in a table. Any row without a person's name will stall.
| Role | Frequency | Main tasks |
|---|---|---|
| Content updates | Weekly | News, case studies, vacancies |
| Numbers | Monthly | Enquiries, traffic and speed, shared with the team |
| Technical maintenance | Monthly | CMS and plugin updates, backups |
| Legal review | Once or twice a year | Cookies, privacy, accessibility statement |
| Translation | Each update | Reflect changes in other languages |
Recovery
If upkeep has stalled, do not try to restart everything at once. Restart the monthly numbers check first and use it to choose one or two pages to improve next.
12. Mistake 10: Weak procurement and contracts
Warning signs and causes
The quote says only "complete website" and you cannot tell what is included. The number of revision rounds and how changes are handled are not written down. Such vagueness surfaces mid-project as extra costs and delays.
The cause is comparing prices before the scope is fixed. How to read a quote is covered item by item in redesign costs and how to read a quote.
Prevention: five things to settle in the contract
Before signing, settle at least these five in writing.
| What to settle | What happens if you don't | Example wording |
|---|---|---|
| Scope | Migration and redirects fall outside scope | Number of pages migrated, number of redirects |
| Deliverables and acceptance | Disputes over what "done" means | Speed and accessibility pass marks |
| Change handling | Extra costs pile up | Hourly rate and approval steps |
| Post-launch maintenance | Unclear whether fixes are paid | Scope, period and response times |
| Ownership and rights | You cannot take over domain or code | Domain, hosting and CMS in the client's name |
In Japan, liability differs depending on whether the contract is a fixed-result contract (ukeoi), where the supplier answers for completion, or a service contract (jun-inin), where it answers for performing the work. IPA publishes model contracts that explain responsibilities by phase for system development.
When hiring an individual freelancer in Japan, the Freelance Act in force since 1 November 2024 requires the client to state the work, fee and payment date. Dutch agencies usually quote by the hour rather than a total; surveys put the agency average at €106 per hour and freelance web developers at €73 (both excluding VAT).
Recovery
If you are already disputing scope, list the remaining work and sort it into in scope, out of scope and undecided.
If the relationship cannot continue, secure ownership and admin access for the domain, hosting, CMS and source code first. Without them, another agency cannot take over.
13. Recovering after the fact: four common stages
Don't roll back in a panic
When numbers fall after launch, someone will suggest going back to the old site. But rolling back changes the URLs again and makes search signals even less stable. Investigating the cause first is rarely too late.
The table summarises the recovery flow common to all ten patterns in four stages.
| Stage | Typical duration | What to do |
|---|---|---|
| 1. Stop the bleeding | Same day to 3 days | Check robots.txt, noindex, tracking tags and forms |
| 2. Isolate the cause | 1–2 weeks | Identify affected pages and timing; match to the ten patterns |
| 3. Fix | 2–6 weeks | Fix redirects, content, speed and paths, biggest impact first |
| 4. Prevent repeats | 1–3 months | Update the checklist and post-launch roles |
Durations are DevNJ's rule of thumb and vary with site size and cause.
Effort and cost ranges for recovery
Recovery costs vary widely by cause. The table converts typical effort into money using Dutch hourly rate surveys (€73–€106).
| Fix | Typical effort | Cost range (excl. VAT) |
|---|---|---|
| Fix robots.txt and noindex | 1–2 hours | €73–€212 |
| Reinstall tracking tags and events | 4–8 hours | €292–€848 |
| Add redirects for top pages (up to 100) | 4–12 hours | €292–€1,272 |
| Speed up top pages | 8–24 hours | €584–€2,544 |
| Main accessibility fixes | 16–40 hours | €1,168–€4,240 |
Effort figures are DevNJ's estimates; real quotes depend on the platform and volume. Convert to yen at the current rate. The point is that configuration gaps are cheap to fix, while design-level rework is expensive.
14. How failures differ between Japan and Europe
Same mistake, different weight
For sites aimed at both Japan and Europe, the same mistake can have different effects, especially in law, languages and tracking.
| Area | Sites for Japan | Sites for Europe |
|---|---|---|
| Accessibility | Reasonable accommodation duty; JIS X 8341-3 is the usual web reference | European Accessibility Act; WCAG 2.2 is the practical benchmark |
| Cookies and tracking | External transmission rules (in-scope businesses) | Tracking before consent may be unlawful; regulators check |
| Languages | Often Japanese only | Two or three languages are normal |
| Search engines | Google and Yahoo! JAPAN (Google technology) | Mostly Google; Bing has a share |
| Contract type | Ukeoi or jun-inin; Freelance Act | Usually hourly, built up from estimates |
On European sites tracking before consent is a legal issue, so mistakes 4 and 7 often occur together; one person should own both consent and measurement.
15. Case studies (fictional examples)
The following uses fictional companies to show how failures cascade and how to recover. They are not real companies or cases.
Case A: A manufacturer that changed the design and halved its enquiries
Company A, a Japanese parts maker, prompted by a competitor's new site, completely changed its homepage photos and design. Two months after launch, enquiries fell from 20 to 10 a month.
The "technical documents" pages that had generated many enquiries had been removed during migration, so engineers arriving from search hit dead ends. Company A restored the pages, added 301s from the old URLs, and added a path from the documents to the contact form.
Case B: A service company whose auto-redirect confused Dutch users
Company B, a Japanese service firm based in the Netherlands, added auto-redirects by browser language during its redesign. Japanese users on Dutch-language computers could no longer reach the Japanese version.
Company B removed the auto-redirect and showed a language choice on the first visit instead. It also reset hreflang between the versions.
Case C: A retailer whose launch slipped three months over scope
Company C, which sells online to EU consumers, hired an agency on a "complete website" quote. Mid-project it emerged that migrating 800 old product pages was out of scope, causing extra cost and delay.
| Case | Main mistakes | Core of the recovery |
|---|---|---|
| A (manufacturer) | 2 Design-only, 3 Late content | Restore removed pages with 301s, add paths |
| B (services) | 8 Broken languages | Stop auto-redirect, reset hreflang |
| C (retail) | 10 Contract, 7 Legal | Redefine scope, add accessibility requirements |
16. Pre- and post-launch checklist
Finally, here are checks that prevent the ten patterns. Use them one month before launch, on launch day and for 90 days after. The project flow itself is covered in redesign process and schedule.
Planning and procurement
- The goal fits in one sentence and the numbers to measure are set
- At least 12 months of the old site's numbers are exported and saved
- The quote includes migration, redirects, languages, accessibility and maintenance
- Post-launch roles have named owners
Design and build
- Every page is in an inventory with a keep, update, merge or remove decision
- There is a map from old URLs to new URLs
- Speed (LCP, INP, CLS) and accessibility (WCAG 2.2 AA) pass marks are in the requirements
- No analytics or ad tags run before cookie consent
Launch day and the first 90 days
- No staging blocks remain in robots.txt or noindex tags
- Top old URLs 301-redirect to the correct new URLs
- Key actions such as completed enquiries are recorded
- Goal numbers are compared with the old site at weeks 4, 8 and 12
Summary
- Think of redesign failure as four types: results drop, budget and time overrun, no one runs it, or it breaks the rules
- Most of the ten patterns are introduced in planning and design and surface within 90 days of launch
- The usual starting point is no goal. Before starting, decide which number should move and by how much, and save the old site's numbers
- A content inventory and URL map prevent three failures at once: migration, redirects and languages
- When procuring, settle scope, acceptance, changes, maintenance and ownership in writing
Sources
Search and analytics (Google and others)
- Site moves with URL changes — Google Search Central: fluctuation, timing, redirects for at least one year, leftover blocks
- Managing multi-regional and multilingual sites — Google Search Central: no auto-redirects, separate URLs, full translation
- Tell Google about localized versions of your page — Google Search Central: hreflang setup
- Block search indexing with noindex — Google Search Central: how noindex works
- Introduction to robots.txt — Google Search Central: the role of robots.txt
- HTTP status codes and network errors — Google Search Central: how soft 404s are handled
- Change of Address tool — Search Console Help: domain moves
- Universal Analytics has been replaced — Google Analytics Help: processing ended 1 July 2023
- Data retention — Google Analytics Help: GA4 retention of 2 or 14 months
- Study of 892 domain migrations — Search Engine Journal: average 523 days (reported cases)
Speed and accessibility
- Web Vitals — web.dev: LCP, INP and CLS thresholds at the 75th percentile
- Vodafone case study — web.dev: 31% better LCP, 8% more sales
- PageSpeed Insights: measuring speed
- Web Almanac 2024 Performance — HTTP Archive: pass rates 43% / 54%
- WebAIM Million — WebAIM: failures on 95.9% of home pages
- WCAG 2.2 — W3C: the accessibility standard
- Evaluating Web Accessibility — W3C WAI: evaluation approaches
Law and regulation
- Directive (EU) 2019/882 — EUR-Lex: the Act, microenterprise exemption, application date
- European Accessibility Act — European Commission: products and services covered
- European Accessibility Act — business.gov.nl: Dutch supervisors and accessibility statement
- AP pakt misleidende cookiebanners aan — Autoriteit Persoonsgegevens: cookie banner checks
- Cookie banner follow-up — Autoriteit Persoonsgegevens: fixes after warnings
- Amended disability discrimination act — Cabinet Office of Japan: reasonable accommodation mandatory from 1 April 2024
- External transmission rules — Ministry of Internal Affairs and Communications: Telecommunications Business Act disclosure rules
- Freelance Act — Japan Fair Trade Commission: duty to state terms in writing
Projects and procurement
- Delivering large-scale IT projects on time, on budget, and on value — McKinsey: 5,400+ projects, overruns
- Model IT transaction contracts — IPA: model contracts
- Content Audits — Nielsen Norman Group: inventory versus audit
- Benchmark report Agencies Q4 2025 — Simplicate: average Dutch agency hourly rate (€106)
- Web developer zzp hourly rates — Knab: freelance web developer hourly rate (€73)
All sources were checked on 26 October 2026. Rules and figures change, so check the official sites for the latest information before making decisions.
FAQ
- What is the most common website redesign mistake?
- Starting without a goal and the numbers that measure it. Without a goal, decisions follow taste, and quiet work such as content migration, tracking and redirects gets pushed back. If enquiries or search traffic then fall after launch, you cannot tell why or even whether the redesign worked.
- Search traffic dropped after our redesign. What should we check first?
- Check three things: redirects from old URLs to new ones, leftover robots.txt or noindex blocks, and where your sitemap and canonical tags point. Google says visibility may fluctuate during a move, but a sharp drop right after launch is usually a configuration gap. Checking your top pages first often brings traffic back quickly.
- Why do redesigned websites often get slower?
- Large images, video, animation and third-party tags pile up without a performance check. Developers test on fast connections and miss it, while real users visit on phones. Write LCP 2.5 seconds, INP 200 milliseconds and CLS 0.1 into the requirements as pass marks, and measure before launch.
- How do we avoid losing analytics data during a redesign?
- Before launch, list the key actions such as a completed enquiry and confirm on staging that the new site records them under the same names. Include the GA4 tag, the cookie consent setup and Search Console ownership in the migration scope. Export and keep the old site's numbers before launch.
- Which laws are easy to miss when redesigning a site for European users?
- The European Accessibility Act and cookie consent. The Act has applied since 28 June 2025 to consumer services such as e-commerce. Services from microenterprises with fewer than 10 staff and turnover or balance sheet of no more than EUR 2 million are exempt. In the Netherlands, the data protection authority (AP) keeps checking cookie banners.
- What should a contract with a web agency settle to prevent a failed redesign?
- Five things: the scope (especially content migration and redirects), deliverables and acceptance criteria, how changes are estimated and approved, post-launch maintenance, and who owns the code, domain and accounts. In Japan, whether the contract is a fixed-result contract or a service contract changes the liability, and IPA's model contracts are a useful reference.
Planning a site that has to work in both markets?
Get in touch