Website optimisation without guesswork

How do you know if a website change actually worked?

Written by Neil Webley · Last updated 19 July 2026

You changed your headline, button, offer or layout. Enquiries went up. But did your change really improve conversions, or did something else happen at the same time?

This guide explains why simple before-and-after comparisons can mislead small businesses, and how A/B testing helps you make website decisions with more confidence.

Start Testing Free See how FreeCROTool works Why is my website not converting?

You found something that looks wrong. Now what?

Maybe your call-to-action is too weak. Maybe your pricing page creates doubt. Maybe your product image is not doing enough. Spotting a possible problem is useful, but it is only the start.

You notice friction

Visitors reach an important page but do not click, enquire, book or buy as often as you hoped.

You form an opinion

You decide the button, headline, offer, trust message, form, image or wording might be putting people off.

You make a change

The page looks better, so it feels like the conversion problem has been solved.

The dangerous part: the change feels like progress

Updating a website gives you a sense of completion. You have done something visible. The page looks cleaner, stronger or more persuasive.

But the feeling of improvement is not the same as evidence. If conversions rise afterwards, it is tempting to assume your change caused it. If conversions fall, it is tempting to blame traffic, the market or bad luck.

A website change is an opinion. An A/B test turns it into evidence.

The aim is not to remove judgement. Good testing starts with judgement. The aim is to stop guessing whether your judgement was right.

Why before-and-after comparisons can fool you

A common mistake is to change a page, compare this month to last month, and decide the change worked. The problem is that the two periods were not identical.

Traffic changed

You may have had more visitors, fewer visitors, or a different mix of people from search, ads, email or social.

Advertising changed

A stronger campaign, extra impressions or better targeting can make the website look better than it really was.

Timing changed

Payday, school holidays, weather, events, fashion, seasonality or the economy can influence buying behaviour.

Luck changed

On a low-volume website, a few extra enquiries can make a big percentage difference without proving the page improved.

How A/B testing gives you a cleaner answer

Instead of changing the page for everyone and hoping the next month tells the truth, an A/B test shows different versions to different visitors during the same period.

Version A stays the same

Your original page keeps running, so you still have a fair baseline to compare against.

Version B tests your idea

Your new headline, button, form, offer or trust message is shown to a comparable group of visitors.

✓

You compare behaviour

Because both versions run at the same time, they are exposed to the same market, campaigns, season and timing.

Learn about the visual A/B testing editor A/B testing without developers

What should a small business test?

You do not need a huge experimentation programme to start learning. Small businesses often get the most value from testing obvious points of friction near important actions.

Homepage headlines and opening messages
Button text and call-to-action placement
Pricing explanations and offer wording
Trust signals, reviews, guarantees and reassurance
Contact forms, booking prompts and enquiry journeys
Exit intent messages, announcement bars and promotional banners

Keep the test simple

The more things you change at once, the harder it is to know what made the difference. A good first test usually asks one clear question.

For example: “Will changing the hero button from ‘Learn More’ to ‘Get a Free Quote’ increase enquiry clicks?”

What if your website has low traffic?

Low traffic does not mean you should never test. It means you should test carefully, avoid overclaiming, and focus on practical learning rather than pretending every result is perfect science.

Look for meaningful signals

For smaller sites, results may be directional rather than definitive. That can still be better than making blind changes.

Test bigger changes first

A tiny wording tweak may be hard to detect. A clearer offer, stronger CTA or better trust message may create a more visible signal.

Measure the right action

Track the action that matters: enquiries, bookings, purchases, quote requests, form starts, calls or important clicks.

A simple website testing process

When you think something on your website is less than optimal, use a repeatable process instead of jumping straight to a permanent change.

Spot the problem

Use analytics, feedback, common sense or customer behaviour to identify where visitors may be hesitating.

Write a hypothesis

Describe what you want to change, why it might help, and which action should improve.

Run the test

Show the original and new version during the same period so the comparison is fairer.

Read the result carefully

Check whether the change appears to help, hurt or make little difference.

Make a decision

Keep the winner, reject the loser, or run a better follow-up test if the result is unclear.

Keep learning

One test teaches you something. A series of tests helps you understand your visitors properly.

FreeCROTool helps you test instead of guess

FreeCROTool is built for small and growing websites that want to make better decisions without enterprise CRO software, developers or complicated workflows.

Edit pages visually without touching code
Run practical A/B tests on real website visitors
Track results using your own Google Analytics
Start free and upgrade only when you need more room

Start Free Watch a real A/B test demo

Good first test idea

Pick one important page and one important action. Then test one change that could make that action more likely.

For example: change a vague “Submit” button to “Get My Free Quote”, or test a trust message above a booking form.

Related guides

Keep learning how to improve your website with a more practical, evidence-led approach.