Why Your Website Is Costing You Clients Before They Even Contact You

There’s a version of your business online that you’ve probably never seen the way your customers see it. They land on your website, form an opinion faster than they can consciously read a word of it, and either stay or leave. Most leave. And almost none of them will tell you why.

This isn’t theory. It’s the pattern we see in nearly every discovery audit we run at Gravity Projex, across every industry. A business doing everything right in real life, quietly undermined by a digital presence that stopped keeping up somewhere around 2021.

What follows is what we actually look for, why each thing matters, and what the research says it’s costing you. Some of it you can fix yourself this week.

The first fifty milliseconds decide more than you’d like

Researchers at Carleton University ran a now well-known study on how quickly people form an opinion of a web page. The answer was 50 milliseconds. Not fifty milliseconds to read your headline. Fifty milliseconds to form a visual judgement that then colours everything they read afterwards.

That’s the uncomfortable part. It isn’t a considered assessment of your credentials. It’s a snap aesthetic reaction that happens before conscious thought, and it acts as a filter on everything that follows. A visitor who has already decided your site looks dated will read your genuinely excellent case study through that lens, and find reasons to agree with the judgement they’ve already made.

You cannot argue your way out of this with better copy. The look and feel has to carry its own weight, and it has to load fast enough to be seen at all.

Speed, and the numbers most people get wrong

Everyone knows they should have a fast website. Fewer people know the size of the effect, which is much larger than the vague “it’s better” most business owners have absorbed.

Google commissioned Deloitte to study this properly. They tracked mobile site data across 37 retail, travel, luxury and lead generation brands in Europe and the US, isolating speed as the variable. A 0.1 second improvement, one tenth of one second, produced an 8.4% lift in retail conversions and a 9.2% increase in average order value. Travel conversions rose 10.1%. For lead generation information pages, the sort most service businesses run, bounce rates improved by 8.3%.

Portent’s analysis of over 100 million pageviews found the same curve at a coarser scale: sites loading in around one second converted at roughly triple the rate of sites taking five seconds. And in a case study published on Google’s own web.dev, Vodafone Italy ran an A/B test on two pages that were visually and functionally identical. The only difference was a 31% improvement in load performance. It produced 8% more sales.

Three things follow from this that most business owners haven’t been told.

The penalty is front-loaded. The difference between one second and two seconds costs you far more than the difference between four and five. Most of the damage is done early, which is exactly why “it feels fine to me” is such an unreliable test.

Google grades you on real visitors, not your laptop. Core Web Vitals are the three metrics Google uses to measure real-world experience: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, and Cumulative Layout Shift under 0.1. Critically, these are field measurements taken from actual Chrome users at the 75th percentile over a rolling 28-day window. A perfect score in a developer tool means nothing if a quarter of your real visitors on mid-range phones are having a slow time. Worth noting too: INP replaced First Input Delay back in March 2024, so any advice you’re reading that still mentions FID is out of date.

You can’t blame the network anymore. Australia’s median mobile download speed now sits above 120 Mbps, ahead of the median fixed broadband connection. Mobile is faster than home internet here. If your site is slow on a phone in Brisbane, that’s your site, not the carrier.

The usual culprits are unglamorous: cheap oversubscribed hosting, images uploaded straight off a camera at 4MB each, a plugin count in the high thirties, and a bloated theme carrying code for features you’ll never use. None of that requires a rebuild to fix. It requires someone who knows where to look.

Trust signals, and the ones you’ve stopped seeing

Visitors are unconsciously auditing you for reasons to be cautious. Not consciously, and not fairly, but constantly. Each of these is small on its own. Stacked, they create a feeling that something is off, and almost nobody will stay long enough to work out what.

• A copyright date stuck three years ago. The most common single tell of a neglected site, and the easiest thing in the world to fix properly with one line of code so it never happens again.
• A contact form that silently fails. This one is genuinely alarming, because it’s invisible. Forms break when hosting changes, plugins update, or an SMTP setting expires, and the failure mode is usually silent. The enquiries just stop, and the business assumes it’s a quiet month. Test yours from an outside email address on the first of every month. Every single month.
• Stock photography that looks nothing like you. The smiling headset call centre, the four colleagues laughing at a laptop. People recognise these instantly now and read them as a signal that there’s nothing real to show.
• No address, no phone number, no human being. A business with no visible location and no named people reads as a risk, particularly to Australian customers who want to know you’re local and reachable.
• No pricing signal at all. You don’t have to publish a rate card. But a page that gives no indication of whether you’re a $500 or $50,000 proposition wastes the visitor’s time and yours, and it reads as evasive.
• Dead links and orphaned pages. A 404 on your own site tells a visitor that nobody has looked at this in a long while.

The pattern behind all of these is the same. Your website is the only staff member who talks to every single prospect, and it’s the one nobody performance-manages.

Mobile, honestly

You’ll read a lot of confident claims that mobile makes up 60% or more of Australian web traffic. The honest answer is that it depends enormously on your industry, and Australia is actually more desktop-heavy than the global average because so much of our browsing happens at work. Some professional services sites see a desktop majority. Local service and hospitality businesses routinely see 70 to 80% mobile.

So don’t take a number from a blog, including this one. Open your own analytics, look at the device breakdown for the last 90 days, and use your figure. Then go and use your own website on your own phone, on mobile data rather than the office wifi, and try to do the thing you most want a customer to do.

What usually surfaces when a site was designed for desktop and made responsive afterwards: body text too small to read without pinching, tap targets crowded so close that thumbs miss them, navigation that requires two hands, forms that trigger the wrong keyboard, and a phone number that isn’t tappable. None of these are cosmetic. Each one is a point where a ready-to-buy customer gives up and rings someone else.

The contact journey is where most of the money leaks

You’ve done the hard part. Someone has found you, formed a decent impression, read enough to be interested, and now wants to make contact. This is the moment businesses casually throw away.

Form length has a cliff in it, and it’s not where people expect. Benchmark data across a large sample of 2026 forms shows completion dropping gently from around 23% at three fields to about 17% at five. Then it falls off: roughly 11% at seven fields, and under 7% once you’re asking for ten or more. Every field past five costs you close to double what the earlier ones did.

There’s a nuance here that most advice skips, and it comes from a well-documented test by conversion specialist Michael Aagaard. He shortened a client’s form and conversions fell by 14%. On investigation, he’d removed the fields people actually wanted to fill in, the ones that let them describe their situation, and left only the ones that felt like data harvesting. The lesson isn’t “fewer fields wins.” It’s “remove the fields that serve you and keep the ones that serve them.”

For most service businesses, the honest minimum is a name, a way to reach them, and a box where they can say what’s going on. Everything else you can ask in the reply.

Then there’s what happens after they hit send. The foundational research here, a behavioural study by Dr James Oldroyd covering more than 15,000 leads and 100,000 call attempts, found that the odds of successfully contacting a lead dropped roughly 100 times when the first attempt came at 30 minutes instead of five, and the odds of qualifying that lead dropped by around 21 times. It’s an odds ratio from one dataset rather than a universal guarantee, so treat the exact multiples with appropriate caution. The direction of the finding, though, has been replicated repeatedly, and Harvard Business Review’s follow-up work put the average business response time at around 42 hours.

Sit with that gap for a second. The research says minutes. The average business takes nearly two days. If your website funnels an enquiry into an inbox nobody checks until Thursday, the site isn’t the only thing costing you clients.

The legal layer nobody audits

This is the section that surprises people, and it’s the one I’d most like Australian business owners to take seriously, because the ground moved recently and almost nobody was told.

In April 2025 the Australian Human Rights Commission released new guidelines on equal access to digital goods, services and facilities, replacing advisory notes that had been in place since 2010. The guidelines name WCAG 2.2 Level AA as the minimum accessibility benchmark, and they apply to public-facing digital services in the private sector, not just government.

The legal backbone is section 24 of the Disability Discrimination Act 1992, and the principle that it covers websites isn’t new or theoretical. It was established in Australia back in 2000 when Bruce Maguire took the Sydney Olympics organising committee to the Human Rights Commission over an inaccessible website and won. Coles faced action over its online shopping service more than a decade later. Complaints have been climbing steadily since.

Meanwhile, WebAIM’s annual analysis of a million home pages found detectable WCAG failures on 94.8% of them. Almost everyone is exposed, which is precisely why so few people worry about it.

Here’s why I’d fix it even if there were no law at all: the accessibility work and the conversion work are largely the same work. Readable text sizes, sufficient colour contrast, properly labelled form fields, sensible heading structure, keyboard navigation, descriptive alt text. Every one of those makes the site easier for everyone, and several of them also make your content easier for search engines and AI tools to parse. It’s the rare compliance job that pays for itself in performance.

Two other things worth having on your radar, from 10 December 2026, businesses covered by the Privacy Act must disclose in their privacy policy where automated decision-making is used in ways that significantly affect someone. If your website is running tracking pixels, collecting form data, or feeding an AI tool, the privacy policy sitting on your site since 2019 is probably no longer accurate.

The real question

The question was never whether your website has problems. Every website has problems, including ours, which is why we keep auditing it.

The question is how much those particular problems are costing you in clients you’ll never know you had. Nobody rings to tell you they left because the page took four seconds, or because the form threw an error, or because your site looked like it hadn’t been touched since the last federal election. They just go somewhere else, and you never see the enquiry that didn’t happen.

That’s the expensive part. Not the problems themselves, but their silence.