Web Design Agency: Site Changes Over Four Weeks
A web design agency came to me through a referral. Someone they knew suggested they should take a look at their website. They had traffic coming in. Around 300 visitors a month. But they weren’t getting many inquiries from it.
Looking at the data, they received maybe one or two inquiries per month. That seemed low for the amount of traffic they were getting. They mentioned the site felt slow when they used it. The layout was cluttered. The call to action wasn’t obvious. When I looked at the page load time, it was 4.8 seconds. Not unusually slow, but noticeable.
There’s something a little uncomfortable about a web design agency’s own site having exactly the kind of problem they’d presumably flag on a client’s site. Nobody said that out loud during our first conversation, but I suspect it was part of why the referral came with a bit of embarrassment attached. Agencies are busy running client work, and the one site that never gets prioritised is often the agency’s own.
We made changes over four weeks. Here’s what we did and what happened.
They came through a referral, which meant they were already thinking something could be better. They weren’t defensive about feedback. They just weren’t sure what to change.
When we looked at the site, several things stood out. The page load time was slow enough that people might leave before the page fully loaded. The homepage had a lot of competing elements. The form to request a quote took multiple steps to complete. They tracked their analytics so we could see patterns. Visitors came, spent a short time, and left without going deeper. The bounce rate was high.
The homepage in particular tried to do too much at once. A hero section with three different competing calls to action, a portfolio grid immediately below it showing a dozen projects with no obvious hierarchy between them, and then a services list further down that repeated some of what the hero had already claimed. None of it was individually wrong. Together, it gave a visitor no clear next step, just a lot to look at.
Four weeks is a short runway for a website project, and it’s worth being upfront about why that was realistic here in a way it usually isn’t. Nothing about this required waiting on Google. No new content needed to get indexed and ranked. No technical foundation needed rebuilding from nothing. Every change was something a visitor would experience the moment it went live, which is a genuinely different category of fix from the SEO-driven work most case studies here involve.
- Looked at what was slowing the page down
- Images were large and unoptimised
- Some scripts were blocking page rendering
- Compressed images and adjusted script loading
- Load time went from 4.8 seconds to around 2.5 seconds
- This was the first change, not the complete picture
The portfolio grid was the biggest single offender on the images side, a dozen full resolution project screenshots, some pushing several megabytes each, all loading at once above the fold. None of them had been compressed or resized for web display, they were just the raw exports someone had uploaded once and never revisited. That’s an easy thing to miss when you’re looking at your own site, since the images look fine to you, you already know what’s coming and you’re probably on a fast connection anyway.
The render-blocking scripts were a mix of things that had accumulated over time, an analytics snippet, a chat widget that loaded even though almost nobody used it, a font loading approach that made the browser wait before it could paint any text at all. None of these were individually dramatic, but stacked together they added real, measurable delay before anything on the page became visible or usable.
Most of that first jump, 4.8 seconds down to 2.5, came from installing WP Rocket and letting it handle caching and script timing, the same plugin this site runs. It’s rarely the only fix a slow site needs, but it’s usually the fastest win available.
- Simplified the homepage layout
- Made the call to action more straightforward
- Changed the inquiry form from three steps to one
- Removed some visual elements that weren’t essential
- Continued working on speed, brought load time to around 1.2 seconds
- Changes went live by the end of week two
Cutting the form from three steps to one was the change that took the most convincing. Multi step forms have a real place, there’s a reasonable argument that breaking a long form into stages reduces how overwhelming it feels. But this particular form was three steps for four total fields, name, email, project type, and budget range. Splitting that across three screens added friction without actually reducing how much information someone had to provide. The single step version asked for exactly the same four things, just without three extra clicks in between.
The homepage rework also meant picking one call to action instead of three. That’s a harder decision than it sounds for an agency, since there’s a real temptation to give a visitor every possible path, a contact form, a phone number, a portfolio link, a newsletter signup, all fighting for attention above the fold. We kept the phone number visible in the header, since that’s low cost and some visitors genuinely prefer it, but everything else pointed at the same single form.
- Watched to see how the site performed with the changes
- Inquiries came through the new form
- Bounce rate appeared lower in the data
- The number of pages people visited per session seemed to increase
- By week four, there was a noticeable shift in how the data looked
Weeks three and four were mostly about resisting the urge to keep changing things. It’s tempting, once a site is live with new changes, to immediately start tweaking again based on early data that hasn’t had time to settle. A few days of numbers can look like a trend when it’s actually just noise. We agreed to leave everything alone for the full two weeks and look at the data properly at the end, rather than reacting to every daily fluctuation.
The four weeks was faster than many website projects because we weren’t waiting for rankings or building content. We were modifying what was already there.
Before the changes, they received roughly one or two inquiries per month from 300 monthly visitors. After the four weeks, the inquiry count appeared to increase. They started receiving more contact form submissions. How many is uncertain because I didn’t track it closely afterwards, but they mentioned noticing a difference in their inbox.
The traffic volume didn’t change dramatically. Still around the same monthly visitors. But the proportion of visitors who inquired appeared to shift. There’s information about how page speed can affect whether people stay on a site, though whether that was the main factor here is difficult to isolate.
What made this project easier to read than most is that the traffic itself stayed roughly constant. When both traffic and conversion behaviour move at once, which is common when SEO work and conversion work happen together, it becomes genuinely hard to say which change caused what. Here, since the source of visitors didn’t change and the volume didn’t move much, whatever shift showed up in behaviour was more plausibly connected to the site changes themselves rather than a different mix of visitors arriving.
A note on proof: client confidentiality means I can’t share this business’s name or their actual analytics screenshots. The numbers here are real, reported in writing only.
They mentioned one inquiry turned into a client project. Whether that would have happened anyway, I don’t know. But the timing aligned with the changes.
In this case, two things changed at the same time. Speed improved and the layout simplified. It’s hard to say which had more effect. A faster site gives people less reason to leave immediately. A clearer layout makes it easier to find the contact form once they’re there. Both could have played a role.
If I had to guess which mattered more, I’d lean toward the form change, mostly because the drop from three steps to one removed an actual decision point, whereas the speed and layout changes mostly reduced friction that was more passive. But that’s a guess, not something the data here can actually confirm, since all three changes shipped in the same short window.
This situation was different from other cases. They weren’t trying to get Google traffic. They had traffic. The issue was that visitors weren’t converting into inquiries. You can read about how conversion works more broadly if you want the wider context.
The timeline moved quickly partly because the work was focused. We weren’t starting from zero. We were adjusting what existed.
A website with traffic can appear to get more inquiries when the site is faster and simpler. This one seemed to. Whether this would happen the same way for a different business, I can’t say. The circumstances matter.
The referral brought them in, which may have helped. They were open to changes. They had data to measure against. They had traffic already. All of those things made it possible to see a shift relatively quickly.
Worth naming the parts of this that made it an easier case than most, since not every project has them. There was already traffic, so no waiting on rankings. There was already analytics set up, so no month spent just getting measurement in place before any real work could start. And there was a clear, narrow problem, people arrive but don’t inquire, rather than a vague sense that “the website isn’t working” without a specific mechanism behind it. Most projects don’t arrive this clean, and it’s worth being honest that this one’s four week timeline isn’t really a template for how fast conversion work normally goes.
There’s also a version of this story that’s slightly less flattering than “speed and layout fixed it,” worth sitting with rather than skipping past. An agency’s own website getting more attention for four straight weeks, after however long it had been neglected, might have produced some of the shift on its own, simply because someone was finally paying attention to whether the form actually worked, whether the copy still made sense, whether anything had quietly broken. Some of what looked like a design and speed win could partly be a maintenance win in disguise. There’s no clean way to separate the two from the outside.
They could continue optimising after week four. I don’t know if they did. But the basic changes were there. The site was faster. The path to inquiring was clearer. If you have traffic but few inquiries, you can explore what affects whether visitors take action.
If there’s one takeaway worth applying elsewhere, it’s probably this: a slow, cluttered site with a three step form isn’t three separate problems competing for attention, it’s one underlying pattern, too much asked of the visitor at every stage, showing up in three different places. Fix the pattern rather than the individual symptoms, and the individual fixes tend to fall out of that naturally. That’s also, incidentally, the same reason a caching plugin alone rarely fixes everything on its own, it closes the technical half of the gap and leaves the layout half exactly where it was.
Have a site getting visitors but not inquiries? You can discuss what might be happening.