I Audited My Own Website. It Failed.
I spent 2 days relaunching my website.
Why? Because I’m preaching light & performant websites to clients, helping them to improve… while mine was lagging behind.
I was too busy to care. Client work first, my own site last. That’s the classic consultant trap, and I fell into it for about a year.
It probably cost me a few leads: the layout was old, the contact form wasn’t performing, and the offer was so broad that a visitor couldn’t tell in 5 seconds what I do or what to do next.
So I ran on my own site the exact audit I run for clients. Here’s what I found, what I changed, and the numbers before and after. If you run a wellness or fitness business, most of this applies to you too.
1. Refine the audience & offer
My services are quite broad, from digital strategy to sales funnel optimisation to SEO. That’s fine for the work. It’s terrible for a website. A visitor couldn’t tell at a glance who it was for and what the next step forward was.
The numbers said it too. The old site had 9 different offers spread over 6 pages, 3 of them selling programmes I don’t run anymore. 4 different calls to action: book a call, get a consultation, enrol, contact us. Prices on none of them.
What I changed:
- 1 audience, stated on every page: longevity, wellness and fitness businesses.
- 3 packages instead of 9 offers, on 1 Services page, with visible prices: an Implementation Package from €500, a Marketing Strategy Sprint at €1,200, and a Fractional Marketing Lead from €1,500 a month.
- 1 call to action everywhere: a free Sales Funnel Audit. It’s in the header, the homepage, the footer of every blog post, and it’s the only form on the site.
- The 3 old programme pages now redirect to Services, so nobody lands on something I’ve stopped selling.
The lesson for your own site: if you can’t say who it’s for and what to click in 1 sentence, neither can your visitor.
2. Optimise for SEO and AI-SEO
A website no one can find is useless. And I had neglected the basics for months, the same basics I flag in client audits every week.
What the audit found:
- 6 of 12 pages had no meta description. 4 had no page title beyond the default “Page name – Pallas Wellness”.
- 2 pages had no H1 at all, 1 page had 2, and the Services page H1 was a slogan with no search term in it.
- 16 images with no alt text.
- The site still declared itself as http:// in WordPress, so the sitemap and every social share link pointed to the wrong version.
- No schema beyond the default: search engines and AI assistants had no structured way of knowing what I sell, for whom, at what price.
- Old programme pages, author archives and legal pages all indexable, competing with the pages that matter.
What I changed:
a title and description written for search on every page, 1 H1 per page with the term people actually type, alt text everywhere, https everywhere, 301 redirects on the 4 dead URLs, noindex on the 5 pages that shouldn’t rank, and schema for the organisation, for me as a person, and for each of the 3 services with its price.
The AI-SEO part is the schema and the plain language. When someone asks ChatGPT or Perplexity for a marketing consultant for gyms in Europe, the answer comes from pages that state who they’re for, what they offer and what it costs, in words a machine can parse. Mine now do.
Result: Lighthouse SEO score from 92 to 100 on every page tested.
3. Optimise the speed
A slow website is a website people leave immediately. Google’s own threshold: the main content should appear in under 2.5 seconds on a phone. Over 4 seconds is “poor”. Mine was at 4.6.
My loading speed wasn’t too bad on desktop. On mobile, where most of my visitors are, it was. 4 things were doing the damage:
None of this required a new site, a new theme or a developer. It was 1 hour in plugin settings and 1 image export.
Result on mobile: performance score from 67 to 82, and the main content now appears about 1 second sooner. Page weight from 1,079 KB to 964 KB, and that’s with the analytics tag I added, which alone is 172 KB.
If you do 1 thing after reading this: open your homepage, right-click your hero image, check its size. If it’s over 300 KB, that’s your next 20 minutes sorted.
4. Revise the look & feel
This one was a personal choice. I believe that we can get performance without changing every image or font, and I’d say the same to a client: don’t redesign when you can fix.
But in my case, it felt like I had outgrown my brand guidelines. The old site had 3 heading styles, 2 button colours, a dark header on some pages and a white one on others, and the mobile menu was, to be polite, chaos. I wanted to start fresh.
What I kept: the logo, the 2 photos, the fonts. What I changed: 1 background colour, 1 button style, 1 header on every page, and every section rebuilt on the same grid. It looks calmer because it is simpler.
Simpler also happens to be lighter, which is why point 3 and point 4 are the same job when you do them together.
5. Fix the plumbing
My contact form was too generic. Name, email, subject, message: the same form as every other website. It told me nothing about the business, so every enquiry started with a call to find out what they actually needed. And it gave the visitor nothing in return for filling it in.
I replaced it with a Free Sales Funnel Audit request: 9 questions, about 3 minutes. Website, socials, booking page, what they sell and at what price, how clients find them today, where they think it breaks, who handles marketing, revenue band. That’s everything I need to send back 3 findings and a short video within 5 working days, and it’s a much better reason to fill in a form than “we’ll be in touch”.
Nothing was tracking conversions either. Jetpack counted page views, which tells you nothing about leads.
Fix: Google Analytics with an event on the audit form submission and on every click to the audit page, so I can see which page sends people to the form.
If your contact form asks for a name, an email and a message, ask yourself what the visitor gets out of it, and what you learn from it. Usually the answer to both is not much.
What’s left to do
Posting regularly value-added content. Not AI slop: insightful content that search engines can find and use as reference, and that AI assistants cite when someone asks them a question I’ve already answered. 1 article a week, 12 weeks, starting with this one.
Everything above took 2 days. The content will take 3 months to show in search. That’s an honest timeline for any wellness or fitness business too: fix the site in days, earn the traffic in months.