A/B Testing Basics: Start Improving Your Website Today

You have opinions about your website. Your designer has opinions. Your staff have opinions. But the only opinion that matters is your customers’—and A/B testing lets them tell you what works.

A/B testing is not complicated. It is simply showing two versions of something to different visitors and measuring which performs better. This guide covers the basics so you can start testing today—an essential part of any conversion optimisation strategy.

What A/B Testing Actually Is

In an A/B test, you create two versions of a page or element:

• Version A: The original (control)
• Version B: The variation with one change

Half your visitors see version A, half see version B. After enough visitors, you compare conversion rates and the winner becomes your new default.

The power is that you remove guesswork. Instead of debating whether a green button or orange button is better, you test it and let data decide.

Why Testing Matters for Business

Small improvements compound. Consider:

• Your website gets 1,000 visitors per month
• Your conversion rate is 2% (20 leads)
• You run tests that improve conversion to 3%
• Now you get 30 leads from the same traffic

That is 50% more leads without spending more on marketing. Over a year, those extra leads add up significantly.

Even modest improvements—going from 2% to 2.3%—are valuable when sustained across many visitors over time.

What to Test

Not everything is worth testing. Focus on high-impact elements:

Headlines

Your main headline is often the first thing visitors read. Test:

• Different benefit statements
• Including vs excluding specific numbers
• Question headlines vs statement headlines
• Addressing pain points vs promising benefits

Call-to-Action Buttons

The button people click to convert deserves attention—small changes here can significantly impact your website’s conversion rate. Test:

• Button text (“Get Quote” vs “Request Pricing” vs “Start Now”)
• Button colour and size
• Button placement on the page
• Surrounding context and supporting text

Forms

Forms are where conversions happen. Test:

• Number of form fields
• Order of fields
• Required vs optional fields
• Single-step vs multi-step forms
• Placeholder text vs labels

Social Proof

Trust signals influence decisions. Test:

• Placement of testimonials
• Photo vs no photo with testimonials
• Specific numbers (“helped 127 Brisbane businesses” vs “helped businesses”)
• Third-party badges and certifications

Images

The images you use affect perception. Test:

• People images vs product images
• Stock photos vs authentic photos
• Image placement and size
• Hero image variations

Page Layout

Structure affects how people engage. Test:

• Long-form vs short-form pages
• Single column vs multi-column
• Video above fold vs below
• Pricing displayed vs hidden

The Testing Process

Step 1: Identify the Problem

Start with data. Look at:

• Which pages have high bounce rates?
• Where do people abandon forms?
• What is your current conversion rate?

Use Google Analytics to find problem areas. A page with 70% bounce rate has more room for improvement than one with 30%.

Step 2: Form a Hypothesis

A good hypothesis follows this format:

“By [making this change], I expect [this outcome] because [this reason].”

Example: “By reducing the contact form from 8 fields to 4, I expect form completions to increase because shorter forms have less friction.”

The “because” matters. It forces you to have a reason beyond “I think it might work.”

Step 3: Create Your Variation

Make one change at a time. If you change the headline, button, and layout simultaneously, you will not know which change affected results.

This feels slow, but it generates reliable insights.

Step 4: Run the Test

Use A/B testing software to split traffic evenly between versions. Popular tools include:

Free options:

• Google Optimize (though being sunset, alternatives emerging)
• VWO (free tier available)

Paid options:

• VWO
• Optimizely
• AB Tasty
• Convert

Let the test run until you have statistical significance—typically 95% confidence that the difference is real, not random chance.

Step 5: Analyse Results

After reaching significance, compare:

• Conversion rate (primary metric)
• Secondary metrics (time on page, scroll depth, etc.)
• Segment analysis (does variation work for all traffic sources?)

Step 6: Implement and Iterate

If the variation wins, implement it permanently. If it loses or is inconclusive, you learned something—now form a new hypothesis.

Then test the next element. Continuous testing creates continuous improvement.

Common Mistakes to Avoid

Ending Tests Too Early

You see the variation is winning after two days and want to declare victory. Do not do this.

Small sample sizes produce unreliable results. What looks like a 20% improvement might be random variation. Wait for statistical significance.

A good rule: minimum 100 conversions per variation before drawing conclusions.

Testing Too Many Things at Once

Changing five elements simultaneously might improve conversions, but you will not know why. You cannot replicate the success on other pages.

One change per test. Yes, this means more tests and more time. But each test produces actionable insights.

Testing Low-Traffic Pages

A/B testing requires sufficient traffic. If a page gets 50 visitors per month, it will take years to reach statistical significance.

Focus tests on your highest-traffic pages first.

Testing Trivial Changes

Testing whether the button should be #FF5733 or #FF5734 is not worth your time. Test meaningful differences that could plausibly affect behaviour.

Not Learning from Losers

A test that loses is not a failure—it is information. If the variation performed worse, ask why. The insight might inform your next test.

Getting Started Without Tools

You can run basic tests manually:

Sequential testing: Run version A for two weeks, then version B for two weeks. Compare conversion rates.

This is not as reliable as simultaneous testing (external factors like time of year affect both periods differently), but it is better than guessing.

Google Ads ad testing: If you run Google Ads, you can test different landing pages by splitting traffic between them.

What Good Looks Like

For reference, here are typical conversion rates and realistic improvements:

Lead generation sites:

• Average: 2-5%
• Good: 5-10%
• Excellent: 10%+

E-commerce:

• Average: 1-3%
• Good: 3-5%
• Excellent: 5%+

A single test rarely doubles your conversion rate. Improvements of 10-30% per winning test are more realistic. But compound enough small wins and the impact is significant.

Your First Test

Here is a simple test to start with:

Test your primary call-to-action button text.

Take your main conversion button (probably “Contact Us” or “Learn More”) and test it against a more specific, benefit-focused alternative.

Original: “Contact Us” Variation: “Get Your Free Quote”

This tests a psychological principle: specific, benefit-focused CTAs typically outperform generic ones. But do not assume—test it.

Building a Testing Culture

The businesses that improve fastest are those that test continuously. Some guidelines:

• Run at least one test at all times
• Document all tests (hypothesis, results, learnings)
• Share results with your team
• Celebrate learnings, not just wins
• Make decisions based on data, not opinions

Over time, you build knowledge about what your specific audience responds to.

When to Get Help

A/B testing can get complex:

• Multi-page funnels require careful tracking
• E-commerce testing involves revenue calculations
• B2B with long sales cycles needs proper attribution

If you are serious about conversion optimisation but lack time or expertise, professional help can accelerate results significantly.

---

Ready to start improving your conversion rate? Platform21 provides conversion optimisation services for South East Queensland businesses, including A/B testing strategy and implementation. Get in touch or explore our conversion optimisation to discuss your optimisation goals.

Related Articles:

• 5 Signs Your Website Is Costing You Customers
• Heat Maps and User Recordings: Understanding How People Use Your Site