The Only Honest Answer: You Need a Before-and-After Baseline
You can tell if your website redesign worked by comparing specific, pre-defined metrics from before the redesign against the same metrics after launch, measured over a long enough window to account for seasonal variation and traffic fluctuations. If you didn’t capture baseline data before you launched the new site, your ability to evaluate the redesign drops dramatically. You’ll be left relying on gut feel, anecdotal feedback, and vanity metrics that tell you very little about business impact.
That might sound blunt, but it’s the reality we encounter on a significant number of projects. A company spends £40,000 to £200,000 on a website redesign, launches it with genuine excitement, and then three months later someone in the leadership team asks “so, did it work?” The room goes quiet. The marketing team pulls up some Google Analytics screenshots. Someone mentions that bounce rate went down. Someone else says the sales team likes the new look. But nobody can connect the redesign to actual business outcomes with any confidence.
This article is about making sure that doesn’t happen to you. Whether you’re mid-project, about to launch, or evaluating a redesign that went live months ago, we’ll cover what “worked” actually means, which metrics matter, how to set up a proper comparison framework, and what to do if you launched without baselines.
Define What “Worked” Means Before You Measure Anything
The first problem with evaluating a redesign is that most teams never agree on what success looks like. The CEO wanted the site to “feel more premium.” The head of marketing wanted more leads. The sales director wanted better case studies. The UX designer wanted faster page loads and improved accessibility. Everyone got something they wanted, but nobody established a shared definition of success that could be measured.
In our projects, we require this definition before design begins. We sit down with stakeholders and force a specific conversation: what does this website need to do differently after the redesign that it’s not doing now? Not “look better” or “be more modern,” but specific, observable changes in user behaviour or business outcomes. The answers typically fall into a few categories.
Conversion improvements are the most common. You want more demo requests, more contact form submissions, more downloads of a key resource, or more signups. These are straightforward to measure if your tracking is in place.
Engagement improvements come up often for content-heavy sites. You want visitors to read more pages, spend more time with your content, or engage with interactive tools like calculators or configurators.
Efficiency improvements matter for sites that serve existing customers. You want fewer support tickets because the knowledge base actually works. You want customers to self-serve instead of calling.
Acquisition improvements relate to getting more of the right people to the site in the first place. Maybe the old site was invisible in search results for your most important terms, or your paid campaigns were sending traffic to pages with a 90% bounce rate.
The key discipline here is choosing a primary metric. You can track all of the above, but you need one number that, if it moves in the right direction, means the redesign justified its investment. For most B2B companies in the 10-to-250-employee range, that primary metric is qualified lead volume or lead-to-opportunity conversion rate. Everything else is a supporting indicator.
The Baseline Problem: What You Need From the Old Site
A proper evaluation requires at least three months of clean data from the old site, ideally six to twelve months to account for seasonality. “Clean” is doing a lot of heavy lifting in that sentence. What we typically find on mid-market sites is that the analytics setup on the old site was incomplete, misconfigured, or tracking the wrong things entirely. Form submissions weren’t being tracked as conversions. Internal traffic was inflating the numbers. Key pages had broken tracking scripts. The data existed, but it was unreliable.
This is why we build measurement into website projects from the start, as described in our measurement systems guide. If you’re planning a redesign and haven’t launched yet, the single most valuable thing you can do right now is audit and fix your current tracking so you have trustworthy baselines to compare against later.
Here’s what you actually need to capture from the old site before it goes away:
- Monthly conversion counts and rates for every goal that matters (form fills, demo requests, downloads, purchases)
- Traffic volume by channel (organic, paid, referral, direct, email) with at least six months of history
- Behaviour flow data showing the most common paths users took through the site, especially paths that ended in a conversion
- Page-level performance for your top 20 landing pages (sessions, bounce rate, conversion rate, average time on page)
- Technical performance baselines including Core Web Vitals scores, average page load time, and mobile usability scores
- Search visibility data from Google Search Console showing impressions, clicks, and average position for your target keywords
Export this data. Screenshot it. Put it in a shared document that your entire team can access. Analytics platforms change their interfaces, old properties get archived, and people forget login credentials. Treat your baseline data like a legal document.
How Long to Wait Before Evaluating
One of the most common mistakes is evaluating too early. Someone pulls up the analytics dashboard a week after launch, sees that sessions dropped 15%, and panics. Or they see a spike in traffic from all the social media posts announcing the new site and declare victory. Neither tells you anything useful.
You need a minimum of 8 to 12 weeks of post-launch data before drawing any conclusions about whether the redesign worked. Here’s why that window matters.
The first two weeks after launch are essentially noise. You’ll see unusual traffic patterns from internal users exploring the new site, from announcement campaigns driving temporary spikes, and from search engines re-crawling and re-indexing your pages. Google’s ranking algorithms need time to process significant site changes, and you might see temporary fluctuations in organic visibility that have nothing to do with the quality of your new site.
Weeks three through six give you the first useful signals. Traffic patterns start to normalise. You can begin comparing conversion rates on key pages. But you don’t have enough data yet to distinguish a real trend from a blip. A single strong week can skew your averages dramatically if your overall traffic volume is modest.
Weeks seven through twelve give you a genuinely comparable dataset. You have enough volume to calculate meaningful averages, enough time to spot consistent patterns, and enough distance from the launch to filter out the novelty effect. For B2B sites with longer sales cycles, you might need even longer. If your typical time from first visit to closed deal is 90 days, you won’t see the full revenue impact of the redesign until at least 90 days post-launch.
The Seasonal Trap
Be careful about which months you’re comparing. If your old site baseline is from October through December and your new site launched in January, you might be comparing your busiest quarter against your quietest. Always compare like-for-like periods: this January versus last January, this Q2 versus last Q2. Year-over-year comparisons are far more reliable than month-over-month for detecting the true impact of a redesign.
The Metrics That Actually Tell You Something
Not all metrics are equally useful for evaluating a redesign. Some are genuinely informative. Others are what we call “comfort metrics,” numbers that make everyone feel good but don’t connect to business outcomes. Here’s how to distinguish between them.
High-Value Metrics
Conversion rate by traffic source is probably the single most important metric for evaluating a redesign. Your overall conversion rate can be misleading because it blends together visitors with very different intent levels. Someone arriving from a branded Google search is far more likely to convert than someone who clicked a display ad. If your traffic mix changed at the same time as your redesign (which it often does, since teams frequently adjust ad spend around launches), your overall conversion rate can move without the site itself being responsible. Breaking it down by source isolates the site’s performance from changes in your marketing spend.
Goal completion rate on key pages tells you whether the specific pages you redesigned are performing better. If you rebuilt your pricing page because the old one confused visitors, compare the percentage of pricing page visitors who proceed to a demo request before and after. This is where the redesign’s impact is most directly visible.
Organic search traffic to non-branded terms reveals whether the redesign helped or hurt your visibility for the keywords that bring new prospects. A redesign can easily damage SEO performance if URLs changed without proper redirects, if content was removed or consolidated poorly, or if technical SEO elements were overlooked. Monitoring non-branded organic traffic in the months after launch is essential.
Lead quality indicators require connecting your website analytics to your CRM, but they’re worth the effort. If your form submissions went up 30% but your sales team reports that lead quality dropped, the redesign might have attracted more tyre-kickers rather than genuine prospects. Track the percentage of website-generated leads that progress to a qualified opportunity stage.
Lower-Value Metrics (Handle With Care)
Bounce rate is frequently cited in redesign evaluations, but it’s one of the least reliable indicators. Google Analytics 4 has replaced traditional bounce rate with “engagement rate,” which measures the percentage of sessions that lasted longer than 10 seconds, had a conversion event, or had two or more page views. Even this improved metric is easily skewed. A one-page site with a clear call to action might have a high “bounce rate” because visitors complete the form and leave. That’s success, not failure.
Average session duration is similarly tricky. Longer sessions could mean visitors are engaged and exploring. Or it could mean they’re lost and frustrated, unable to find what they need. Without context, session duration tells you almost nothing about whether the redesign worked.
Pageviews per session has the same ambiguity. More pageviews might indicate genuine interest, or it might indicate that your navigation is confusing and people are clicking around desperately looking for the contact page. Always interpret engagement metrics alongside conversion metrics, never in isolation.
What To Do If You Launched Without Baselines
This is the situation we see most often. The redesign is already live, nobody captured proper baselines, and now the team needs to figure out whether it was worth the investment. You’re not completely out of options, but honesty about the limitations is important.
Check whether your old analytics data still exists. If you were using Universal Analytics, that data was sunset by Google in July 2024 and may no longer be accessible. If you had GA4 running on the old site, you’ll still have historical data in the same property, assuming the tracking wasn’t replaced entirely. Google Search Console retains 16 months of data, so you may still be able to pull historical search performance even if your analytics data is gone.
Use your CRM as a proxy. Your CRM probably has records of every lead that came through the website, timestamped and tagged by source. Compare the volume and quality of website-generated leads from the six months before the redesign against the six months after. This won’t tell you everything, but it tells you the thing that matters most: did the website generate more business?
Talk to your sales team. This is qualitative, not quantitative, but it’s surprisingly useful. Ask whether the quality of enquiries has changed. Ask whether prospects mention the website during sales conversations. Ask whether the sales team is actually using the website as a tool in their process (sending case study links, referencing the pricing page, sharing comparison guides). A redesign that makes the sales team’s job easier is delivering value even if you can’t pin an exact number on it.
Start measuring properly now. Even if you can’t evaluate the redesign retrospectively with full confidence, you can establish today as your new baseline. Set up comprehensive tracking, define your key metrics, and commit to reviewing them monthly. In six months, you’ll have a solid dataset that you can use to evaluate any future changes.
The Hidden Damage a Redesign Can Do
Most evaluation conversations focus on whether things got better. Fewer teams ask whether the redesign caused damage they haven’t noticed yet. But some of the most costly redesign failures aren’t visible in a conversion report.
SEO regression is the biggest risk. When URLs change, when pages are consolidated or removed, when internal linking structures shift, your organic search visibility can take a significant hit. We’ve seen companies lose 30% to 50% of their organic traffic after a redesign because redirects were incomplete or implemented incorrectly. This doesn’t always show up immediately. It can take two to four months for the full impact to become visible in your traffic data, by which point the agency that built the site has moved on and the trail is cold.
To check for this, compare your top 50 organic landing pages from before the redesign against your current organic landing pages. Are the same pages still ranking? Have any high-value pages lost significant traffic? Use Google Search Console to compare impressions and clicks for your target keywords across the old and new periods.
Accessibility regression is another common issue. A visually striking new design might inadvertently create barriers for users with disabilities if colour contrast was reduced for aesthetic reasons, if interactive elements lack proper ARIA labels, or if the site relies heavily on animations that affect users with vestibular disorders. Run your new site through an accessibility checker like axe or WAVE and compare the results against the old site if you have them.
Third-party integration breakage often goes unnoticed for weeks. Chat widgets, analytics scripts, marketing automation tracking, heatmap tools, and A/B testing platforms can all break during a redesign without triggering any obvious errors. Audit every third-party integration within the first week after launch. The longer a tracking gap persists, the more data you lose permanently.
Building a Proper Evaluation Framework
If you want to get this right, either for a current redesign or a future one, here’s the framework our team recommends. It’s not complicated, but it requires discipline.
Six weeks before launch: audit your current tracking, fix any gaps, and export baseline data for all key metrics. Document your primary success metric and three to five supporting metrics. Get stakeholder agreement on what “success” looks like in numerical terms. Not “more leads” but “a 20% increase in qualified demo requests within six months.”
Launch week: verify that all tracking is firing correctly on the new site. Check every conversion event, every goal, every integration. Use Google Tag Manager’s preview mode, check real-time reports, and manually test every form and interaction. Annotate your analytics with the launch date so you can find it easily later.
Weeks two through four: monitor for technical issues, broken tracking, and SEO disruptions. Don’t try to evaluate the redesign’s success yet. Focus on making sure data quality is solid and nothing is broken.
Week eight: pull your first comparative report. Compare weeks three through eight of the new site against the equivalent period from the old site (ideally the same calendar period from the previous year). Flag any metrics that moved significantly in either direction.
Week twelve: run a comprehensive evaluation. By now you have enough data to make confident statements about conversion rates, traffic patterns, and engagement changes. Present findings to stakeholders with clear before-and-after comparisons, acknowledging any confounding factors (seasonal changes, marketing campaign adjustments, market conditions).
Month six: conduct a final evaluation that includes downstream business metrics. How many of the leads generated through the new site converted to revenue? What’s the cost-per-acquisition compared to the old site? This is where you connect the redesign to actual business outcomes, not just website metrics.
When the Answer Is “It Didn’t Work”
Sometimes the data tells you the redesign didn’t deliver the expected improvements. This is uncomfortable but valuable. The worst outcome isn’t a redesign that underperformed; it’s a redesign whose impact nobody ever measured, so the same mistakes get repeated in three years when the cycle starts again.
If your metrics are flat or negative, resist the urge to blame the data or move the goalposts. Instead, dig into why. Which specific pages underperformed? Which user segments behaved differently than expected? Did the redesign address the right problems, or did it solve problems that didn’t actually exist?
Often, a redesign that doesn’t improve conversions was focused on the wrong things. The visual design was overhauled, but the messaging stayed the same. The navigation was restructured, but the content that visitors actually needed was still missing. The homepage looks stunning, but the landing pages that receive 80% of your paid traffic were barely touched. Design changes without strategy changes rarely move business metrics.
The good news is that a well-measured redesign, even one that didn’t hit its targets, gives you a clear roadmap for what to fix next. You know which pages aren’t converting. You know which traffic sources are underperforming. You know where users are dropping off. That’s infinitely more useful than a vague sense that the site “isn’t working.”
Making Your Next Evaluation Easier
The single best thing you can do, right now, regardless of where you are in the redesign process, is to treat measurement as a core project deliverable, not an afterthought. When measurement is designed into the site architecture from the beginning, when tracking requirements are defined alongside wireframes and content plans, evaluation becomes straightforward. You’re not scrambling to figure out what to measure after launch. You already know, and the data is already flowing.
If you’re about to kick off a redesign, insist that your measurement plan is documented and approved before the first design comp is created. If you just launched, spend this week verifying every piece of tracking on the new site and establishing clean baselines going forward. If you launched months ago and still aren’t sure whether it worked, pull together whatever data you can from analytics, Search Console, and your CRM, and be honest about what the numbers show.
A redesign is one of the largest investments a mid-market company makes in its digital presence. Knowing whether it worked isn’t optional. It’s the only way to ensure the next investment is smarter than the last.
