How Website Design Affects Your Traffic

Traffic and sales are different problems, and design affects them differently.

Conversion is about what happens once someone arrives. Traffic is about whether they arrive at all, and that’s decided by things most people don’t think of as design: how the page is structured, what’s in the HTML, how deep the navigation runs, and how fast it loads.

This is the part of design that search engines and AI systems actually read. It’s covered less often than the visual side because it isn’t visible, and it’s where most of the damage gets done.

For the conversion half, we’ve written separately on website ROI.

Speed is a ranking input, and it’s set during design

Core Web Vitals are a confirmed ranking signal, and around 49% of the mobile web currently fails them. That’s not a technical footnote. It means roughly half your competitors are carrying a handicap, and so might you.

The three metrics measure loading, responsiveness, and visual stability. Interaction to Next Paint is the hardest to pass, with about 43% of sites missing the 200ms threshold, and unlike loading speed it can’t be fixed by compressing images. It requires changing how the page handles interaction.

What matters here is that these are design decisions. A full-bleed video, six typeface weights, a scroll-triggered animation, and a carousel of high-resolution imagery each add weight. Together they’re a site that fails. By the time a developer receives those screens the cost is committed.

Setting a performance budget during design, the way an architect works to a load limit, is the only reliable fix. We’ve covered why the handoff matters in web design vs development.

What search engines can’t see

A design decision that regularly costs traffic: putting content behind JavaScript.

If your service descriptions load through a framework after the initial page render, there’s no guarantee a crawler sees them. Google renders JavaScript, but not always, not immediately, and other crawlers are less capable. AI retrieval systems in particular tend to work from the initial HTML.

The same applies to accordions, tabs, and click-to-reveal sections. Content hidden behind an interaction is generally indexed, but it carries less weight than visible text, and it’s less reliably extracted into AI answers.

The practical rule: anything you want found should exist in the HTML that arrives with the page. If a design pattern requires hiding your main content, it’s costing you.

Heading structure decides what gets extracted

Headings look like a styling choice. They aren’t. They’re the map search engines and AI systems use to understand what a page contains and which section answers which question.

This matters more now than it used to. Zero-click searches reached 68% of Google queries between January and April 2026, up from 60.5% in 2024, driven largely by AI summaries. Those summaries are assembled from extracted passages, and extraction works section by section.

Three rules:

Headings should mark real content boundaries, not visual rhythm. An H2 that exists because the page needed a break has told the crawler something false.

One H1 per page, describing what the page is about.

Don’t skip levels to get a smaller font size. Use CSS for that.

A page with a clean heading structure gets segmented correctly. One where headings were chosen for how they looked is harder to parse, which makes it a weaker candidate for citation.

Navigation depth decides what gets indexed

Every page on your site sits some number of clicks from the homepage. Pages buried four or five levels deep get crawled less often and rank less well, because crawl attention follows link structure.

The design decision here is the navigation itself. A mega-menu exposing your service pages directly puts them two clicks from home. A menu that requires going to a category page first pushes everything one level deeper.

Same for internal linking within content. A blog post with no links out is a dead end. One that links to relevant service pages passes authority to them and tells crawlers those pages matter.

This is directly relevant to a common problem: businesses whose homepage outranks their own service pages for commercial terms. Usually the cause is that everything links to the homepage and nothing links to the service page.

Images

Two things cost traffic here.

Weight. Oversized images are the most common cause of a failing Largest Contentful Paint score. A hero image exported at 3000px wide and displayed at 800px is wasting most of what it downloads. Serving appropriately sized versions per breakpoint is a build decision that follows from a design decision about how many images the page needs.

Descriptive text. Alt text exists for accessibility first, but it’s also how images get understood and surfaced. A site full of IMG_4471.jpg with no alt text is invisible in image search, which matters for visual categories like jewellery, interiors, and property.

Mobile is the version that counts

Google indexes the mobile version of your site. If your mobile site shows less content than desktop, the missing content effectively doesn’t exist for search.

This catches sites where a designer simplified mobile by removing sections rather than restacking them. It looks cleaner and quietly deletes ranking signals.

In the UAE this compounds. Mobile accounts for roughly 75.3% of web traffic and 96% of internet users access the web via mobile, with Android holding around 81% of the device market. A site signed off on a laptop and checked briefly on an iPhone has been tested on neither the primary platform nor the primary device.

What design can’t fix

Worth being clear about the limits, because a redesign gets sold as a traffic solution more often than it should be.

Design gets you a fair hearing. It doesn’t create demand. If nobody searches for what you offer, a faster site with cleaner headings won’t change that.

Design also can’t substitute for content. Pages rank because they answer questions. A beautifully built site with four thin pages will lose to an ugly one with forty substantial ones.

And structural changes carry risk. If a redesign changes URLs without a complete redirect map, documented traffic losses of around 40% follow. Design improvements that lose your existing rankings are a net negative regardless of how the new site looks.

The short version

If you want a redesign to help traffic rather than harm it:

• Set a performance budget before design starts
• Keep primary content in the initial HTML
• Use headings to mark real sections
• Keep important pages within two or three clicks of home
• Link from content to the pages you want ranking
• Size images properly and write real alt text
• Show the same content on mobile as desktop
• Produce a full redirect map before launch

None of that constrains what a site can look like. It constrains how it gets built, which is a different thing.

Our SEO work starts during design for this reason, and it’s part of why we build custom rather than templated where the structure matters.

Frequently asked questions

Want to know whether your current site is helping or holding back your traffic? Talk to us and we’ll run the checks.