Conversion rate optimisation, usually shortened to CRO, is the work of getting more enquiries, bookings or sales from the visitors your website already has. Instead of paying for more traffic, I find the points where interested people give up, work out why, and fix the causes in a sensible order. Your conversion rate is the visible number, but the real aim is more paying customers for the same marketing spend.
Good CRO is mostly diagnosis. On a typical UK small business site, the first gains usually come from removing obstacles: a form that fails on a phone, a price that never appears, a call to action that sits below three screens of introduction. Testing has its place, but only once the obvious faults are gone and the site has enough visitors to give a reliable answer.
Who CRO suits, and who it does not
CRO is worth paying for when traffic is not your main problem. The usual signs:
- Search or paid ads bring people to the site every day, yet enquiries or orders have stayed flat.
- Your cost per lead from Google Ads or Facebook ads keeps rising and better targeting has not helped.
- Visitors start the checkout or the quote form and then disappear.
- The team argues about the homepage or the pricing page and nobody has evidence either way.
It is the wrong starting point if very few people visit the site at all. With a handful of visits a day there is too little behaviour to study, and the money is better spent on search visibility or advertising first. It is also the wrong job if the site is so dated or hard to edit that every fix needs a developer; in that case a website redesign planned around what the current site already earns may be the better investment, with CRO following once the new site settles.
What the work covers
Tracking you can trust
Every CRO decision rests on counting enquiries and sales correctly, and that counting is often wrong. Forms fire a conversion on page load, phone clicks are missing, or thank-you pages are counted twice. I check this before anything else, and if you want to look yourself first, my guide on how to test whether your conversion tracking records real enquiries covers the same checks. UK sites with a proper cookie banner only see analytics data for visitors who accept, so reports show a sample rather than everyone, and I read the numbers with that in mind.
Funnel analysis: finding where people drop out
Before touching a single page, I map the route a visitor takes from landing to converting and measure how many are lost at each step. A funnel exploration in GA4 shows whether the leak sits on the landing page, the service page, the form or the checkout. Splitting the same funnel by device and by traffic source often changes the picture completely: a checkout that works well on desktop can lose most mobile visitors at the delivery step.
A page-by-page review of friction
For the pages that lose the most people, I review them as a first-time visitor would, on a phone and on a desktop. I look at whether the page says plainly what you offer and where, whether the price or a price guide appears, whether reviews and credentials sit near the decision, and whether the next step is obvious. Where consent allows, I also watch anonymised recordings and click maps from a tool such as Microsoft Clarity to see where people hesitate, rage-click or scroll past the important part.
Forms, checkout and the moment of contact
The final step is where most of the value sits and where faults are easiest to miss. I test every form and checkout myself on common phones and browsers, look at which fields people abandon, and check what happens after submission. Form abandonment often traces back to something small: a required phone field for people who prefer email, a postcode lookup that rejects valid addresses, or an error message that clears everything already typed.
A/B testing, when the numbers allow it
An A/B test shows two versions of a page to different visitors and compares results. It only works when enough people convert during the test to separate a real difference from chance, which is what statistical significance measures. Many UK small business sites simply do not have that volume, and I will tell you so rather than run tests that cannot produce an answer. Where testing is viable, I write the hypothesis down first, fix the sample size and the stopping point before launch, and do not stop a test early because one version looks ahead. You can check any result yourself with my A/B test significance calculator.
How I run a CRO engagement
- Agree what counts as a conversion. An enquiry, a booking, a sale or a call, and what each is worth to you, so the work aims at revenue rather than form fills.
- Audit tracking and fix what is broken. Nothing else is reliable until this is right.
- Map the funnel and find the biggest leak. I rank problems by how many people they affect and how much each lost conversion is worth.
- Fix the clear faults. Broken forms, slow pages and missing information are corrected directly, without testing, because there is nothing to learn from testing a fault.
- Test the open questions. Where the right answer genuinely is not clear and traffic allows, I run a structured test.
- Measure and report. I compare conversion and lead quality before and after each change, and I record what did not work, which is as useful as what did.
Problems I see most often on UK sites
- The page answers the wrong question first. Long company histories at the top, with the service, area covered and price guide pushed far down.
- Prices hidden or revealed late. Delivery costs or booking fees that appear only at the last step lose buyers, and since the Digital Markets, Competition and Consumers Act 2024, drip pricing of mandatory fees is unlawful, not just poor practice.
- Pressure tactics that cause problems. Fake countdown timers and invented “only two left” messages are dark patterns that the CMA treats as misleading, and they erode trust with the people who notice.
- Ads sending people to the homepage. A paid click lands on a general page that does not mention the offer in the ad. A landing page built for the specific ad campaign usually fixes this better than any amount of homepage tweaking.
- Slow, heavy mobile pages. Large images, chat widgets and several tracking scripts loading before the content; technical SEO work on speed and page weight helps search and conversions together.
- Session recordings capturing personal data. Recording tools need form fields masked and must respect your cookie banner, or you create a UK GDPR problem while trying to fix a sales one.
What you receive
- A tracking check with every fault listed and fixed, or handed to your developer with clear instructions.
- A funnel report showing where visitors drop out, split by device and traffic source.
- A ranked list of changes, each with the evidence behind it, the expected effort and who should make it.
- A testing plan with hypotheses and the traffic each test needs, or a plain explanation of why testing is not yet worth it for your site.
- Before-and-after reporting on every change made, including the ones that made no difference.
What it costs
Each quote is written in pounds after a free audit or a first call, and agreed in writing before work starts. Cost depends mainly on how many funnels the site has, whether tracking needs rebuilding first, and whether I make the changes myself or brief your developer. A single lead form on a service site is a much smaller job than a shop with several product ranges and a custom checkout.
Next step
Send me your website address, the action that matters most to you (an enquiry, a booking or a sale) and roughly how many visitors and conversions you get each month. I will tell you where I think the biggest leak is and whether CRO, a landing page or more traffic is the right place to spend first. Get in touch to arrange a call.