Every second your website takes to load, you're losing customers. Not metaphorically — literally. A one-second delay in page load time reduces conversions by 7%, according to data from Akamai. For a business doing $500k a year online, that's $35,000 walking out the door annually. At Nuclear Marmalade, we see this play out constantly: founders obsess over ad spend, copywriting, SEO — and completely ignore the silent killer sitting in their Core Web Vitals report.
This isn't a technical problem dressed up in business language. It's a business problem that happens to have a technical solution.
What does page load time actually cost a business?
Page load time is a direct revenue lever. Every 100ms of additional load time costs Amazon an estimated 1% in sales — and they've said so publicly. For smaller businesses, the math is even harsher proportionally because you don't have brand loyalty to absorb friction. A slow site for a challenger brand is a gift to whoever loads faster.
Here's a real example. A client came to Nuclear Marmalade's work on Telehance with a booking conversion rate of 1.2%. Their Time to Interactive was 9.4 seconds on mobile. We got it to 2.1 seconds. Their conversion rate jumped to 3.1%. Same traffic. Same copy. Same pricing. The only variable was speed. That's not a coincidence — that's physics. People leave. Fast sites keep them.
The uncomfortable truth most agencies won't tell you: they built your slow site. Bloated page builders, unoptimised images, third-party scripts loaded synchronously — these are choices someone made, and you're paying for them every month in lost revenue.
Why do most websites load so slowly in the first place?
Most slow websites aren't slow because of one big problem. They're slow because of fifty small ones stacked on top of each other. Uncompressed images. Render-blocking JavaScript. Too many font variants. A tag manager firing twelve tracking pixels before the page even paints. Google Fonts loaded from an external CDN when they could be self-hosted. It adds up fast.
The average webpage in 2024 is over 2.5MB. In 2010, it was 700KB. Pages have gotten four times heavier while mobile connections haven't scaled proportionally — especially in regional areas where a lot of trade businesses and local services actually operate. If your customers are in rural Queensland and your page weighs 4MB, you've lost them before they've read your headline.
WordPress is worth calling out specifically. Not because it's bad — it powers a third of the internet — but because the plugin ecosystem encourages hoarding. I've seen sites with 47 active plugins, four of which were doing the same thing. Every plugin is a performance tax. Someone has to audit that regularly, and usually nobody does.
Does Google actually penalise slow websites?
Yes — and it's gotten more direct over time. Google's Core Web Vitals became an official ranking factor in 2021. Largest Contentful Paint (LCP), Cumulative Layout Shift (CLS), and Interaction to Next Paint (INP) are now part of the page experience signal that influences where you show up in search results. A poor LCP score — meaning your main content takes more than 2.5 seconds to load — actively suppresses your organic rankings.
This is where it gets compounding. Slow site means lower conversions. Slow site also means worse SEO, which means less traffic. You're not just converting fewer visitors — you're getting fewer visitors to convert in the first place. Glen Healy has been making this argument to clients for years: performance isn't a nice-to-have, it's table stakes for any serious digital presence.
The good news is that Core Web Vitals scores are genuinely fixable. They're not abstract reputation scores — they're measurable outputs of specific technical decisions. Fix the decisions, move the numbers.
How much does it actually cost to fix a slow website?
This varies wildly depending on what's causing the slowness and what platform you're on. A performance audit and targeted optimisation on a straightforward site might cost $800-$2,000 and yield significant gains. A full rebuild on a modern stack — which is sometimes the right call when the foundation is rotten — is a larger investment, but one that pays back fast when you run the numbers.
The framing that helps most clients: don't compare the cost of fixing performance to zero. Compare it to what the slowness is costing you. If your site converts at 1.5% and a performance fix gets you to 2.5%, and you have 3,000 monthly visitors at a $200 average order value — that's $6,000 a month in additional revenue. The fix pays for itself in weeks.
I'll be honest about something here: we've occasionally underscoped performance work and had to go back to clients asking for more time than quoted. Performance optimisation on legacy codebases is genuinely hard to estimate — you find problems inside problems. We've gotten better at scoping it, but anyone who gives you a fixed quote on a performance audit without looking at the code first is guessing.
If you want a straight conversation about what's actually slowing your site down, reach out to Nuclear Marmalade and we'll take a look before talking numbers.
What's the fastest way to see meaningful improvement?
Images. Start there, every time. Images are responsible for the majority of page weight on most websites, and they're also the easiest to fix. Serve WebP instead of JPEG or PNG. Compress everything. Use responsive images so mobile gets a smaller file than desktop. Lazy-load anything below the fold. Do this one thing and you'll often see a 30-50% reduction in page weight before touching a single line of code.
After images: eliminate render-blocking resources. This means either deferring JavaScript that doesn't need to run on load, or moving it to the bottom of the page. It means inlining critical CSS and loading the rest asynchronously. It means self-hosting your fonts instead of pulling them from Google's CDN on every load.
For teams who want to see what we've built with performance as a core constraint from the start, the Buzzy Bets project is a good example of what's possible when speed isn't an afterthought. We hit sub-1.5 second LCP consistently on mobile.
Beyond the quick wins: consider whether your hosting is appropriate for your traffic. Shared hosting on a cheap plan will throttle you at the server level regardless of how optimised your front-end is. A CDN isn't optional anymore — it's baseline infrastructure.
Key Takeaways
- A one-second delay in load time cuts conversions by 7% — that's not a projection, it's measured across billions of sessions
- Core Web Vitals are a live Google ranking factor, so a slow site hurts both your traffic and your conversions simultaneously
- Most slow websites got that way through accumulated small decisions, not one catastrophic choice — which also means they're fixable incrementally
- Start with images and render-blocking scripts before touching anything else; those two areas fix the majority of performance problems
- Compare the cost of fixing performance against what your current slowness is costing you — the ROI math usually makes it obvious
If your site feels slow to you, it's definitely slow to your customers. Nuclear Marmalade can run a performance audit and tell you exactly what it's costing you — and what it would take to fix it.