Business & Growth

What Is A/B Testing and How Do You Run One?

Learn how to compare two versions of something — like a webpage or email — to find out which one actually works better.

Scroll to start

What's the Best Version? Let the Data Decide.

A/B testing is a way to compare two versions of something to see which one performs better. You show version A to one group of people and version B to another group, then you check which version gets the result you want — more clicks, more signups, more purchases, whatever your goal is.

The "A" is usually your current version (the control). The "B" is the new version you want to test. You split your traffic randomly, let the test run long enough to get trustworthy data, and then you pick a winner.

The key rule is: change only one thing at a time. If you change the button color AND the headline at the same time, you won't know which one made the difference. That discipline is what makes A/B testing actually useful.

Stop Guessing. Start Knowing.

Most decisions in business — website redesigns, email subject lines, ad copy, checkout button colors — get made based on someone's gut feeling. Sometimes the gut is right. Most of the time it's not.

A/B testing removes the guesswork. Instead of arguing about whose idea is better, you let real visitors decide. A tiny 5% improvement in your checkout conversion rate could mean tens of thousands of dollars in additional revenue every year. That's why serious businesses run tests constantly.

💡 Key Insight

The difference between a good website and a great one is usually not a new design — it's small, data-driven changes. A better button color, a clearer headline, one fewer form field. These small wins add up fast.

The Four Steps of a Good A/B Test

Running an A/B test isn't complicated, but you need to follow the steps in order to get results you can actually trust.

The A/B Testing Process
📊
Split
Divide visitors into two random groups
🧪
Show
Group A sees original, Group B sees the new version
⏱️
Measure
Track the goal (clicks, signups, sales) for each group
🏆
Pick Winner
Use the version that got the best results
repeat with next idea

Important rules:

  • Run long enough. Not enough visitors means noisy data. You need enough people to see each version that the difference is statistically clear.
  • Test one thing. Change the button color OR the headline — not both. If you change two things and results improve, you can't figure out which one was the winner.
  • Pick one primary metric. Decide before you start what counts as "success." Is it clicks? Form submissions? Revenue? Don't try to optimize for everything at once.
  • Don't stop early. Even if one version looks better after a day, keep going. People behave differently on weekends vs. weekdays.

Testing a Sign-Up Button

Imagine you run an online course website and your sign-up rate is low. You want to test whether changing your button from gray to green gets more people to click it. Here's what that test looks like:

Your A/B Test Setup
GOAL: Get more people to click the "Sign Up" button.

HYPOTHESIS: A green button will stand out more and get
more clicks than the current gray button.

CONTROL (Version A): Gray button, "Sign Up Now"
TEST (Version B): Green button, "Sign Up Now"

SPLIT: 50% of visitors see A, 50% see B
DURATION: Run for 2 weeks (to cover weekdays + weekends)
METRIC: Number of button clicks (primary)

After two weeks, you check the results. Say version A got 340 clicks out of 5,000 visitors (6.8% click rate), and version B got 410 clicks out of 5,000 visitors (8.2% click rate). Version B wins — green button gets 21% more clicks. You update your site and move on to the next test.

Example Results Table
Version A (Gray Button):   340 clicks / 5,000 visitors = 6.8%
Version B (Green Button):  410 clicks / 5,000 visitors = 8.2%

Winner: Version B
Improvement: +21% click rate

Knowledge Check

Test what you learned with this quick quiz.

Quick Quiz — 3 Questions

Question 1
What is the most important rule when setting up an A/B test?
Question 2
Why should you run an A/B test for at least one or two weeks?
Question 3
What does it mean when a result is "statistically significant"?
🏆

You crushed it!

Perfect score on this module.