In simple words, A/B testing is a method of putting two or more versions of something to test, in order to see which performs the best.
You can test anything: a landing page, a specific design element, a title, a CTA color, a banner ad, an email copy,etc, as long as you compare your two variants over the same time period and similar audiences. And wait for a statistically significant period of time so enough impressions are accrued (VWO’s tool is great for calculating your test duration ).
Obviously, an A/B test on your website makes sense when you compare two variants of an important element of your website that may affect conversions, sales, lead generation or such.
So let’s get down to it and see 5 great A/B test ideas you should use on your website.
Calls To Action play a crucial role in converting your visitors, so you should always test variants to find which one convinces them best!
Test the color, copy, size or location of your most important CTAs and see how these changes affect your most important metrics.
Which of the following two forms with different CTA copy seems more appealing to you? Using A/B testing SAS Data Management found that the second variation resulted in 25% more downloads!
One of the most common elements put to the test are headlines. Changing your headline copy length, font style, font size or tone of voice may greatly improve or worsen your performance.
For example, you might want to test headlines on your landing pages, homepage, blog-posts, paid ads… or on your Exit Bee campaign!
Try them out before it’s too late 🙂
For example see the following headline A/B test. Version A led to 32% more users signing up for a trial!
Tweaking and testing your forms is a definite must! The length, copy, buttons, position of your form are all elements that can lead to a major increase in form submissions once you uncover the winning variation.
The following cool A/B tests two different form layouts. The second variation, which is more compact and removes the asterisks, resulted in a 12.1% increase in revenue per visitor.
The actual layout of your pages isn’t just a design choice. Moving things around impacts how your users interact with the page and how many of them actually convert. Layout elements you may want to test could be:
right vs left navigation
long vs short landing pages
with or without carousels (many experts vote against sliders)
other element positioning
Check out how EA used A/B testing for its Sim City game release and managed to get a 43.4% increase in sales by simply removing an offer banner from the page:
Images are another one of those things that marketers and designers often seem to argue about. Why not put your hypotheses to rest with an A/B test in order to see which hero image or product image works best?
For example, VWO performed an A/B test to see whether a stock image or a real photo would lead to more conversions on a driving academy’s landing page.
Guess what! The real-life image of an actual student led to 161% more conversions.
Timing is everything, especially when marketing budgets are at stake. Apart from relying on industry reports and available research, you may want to conduct your own A/B tests to determine which days & times are best for your audience and website.
For example you may want to test:
- Social media posting timing
- Newsletter blasts timing
- PPC campaigns running times
- Cold calling times
The Animal Legal Defense Fund tested the following email sending it at 10am and 7am. It turned out that the later time resulted in 9% more opens and $3,639 more in donations for their cause!
Extracting results from your A/B test
You carefully designed your A/B test and waited patiently for the test to run long enough. The final and most important step is actually making sense of these results!
Unfortunately, not all A/B tests lead to a significant outcome with a clear variant winner. Sometimes our hypotheses lead to baffling results with some key metrics improving while the other just worsen.
For example an A/B test of a CTA could see a variant leading to 5% more clicks but only 0.013% more conversions. Insignificant, no?
Aim for a high confidence level of over 85% so you know you can rely on your results safely and make sure you pay attention to the metrics that matter (eg more clicks or opens may not be as important as sign-ups or purchases).
Turn Lost Visitors Into Customers
Increase conversions and sales by targeting abandoning visitors with tailored messages.