10 A/B tests to consider
In our previous article we have explained what A/B tests are and how to do them. Now let's take a look at a few tests you can perform.
1. Form A/B test
Forms are perhaps the most important element on a website. While most of the elements on a website are one-way traffic, a form facilitates two-way communication: "What's your name?" - "My name is Lichai." Because of this there are many ways to create a form. So what to test:
- Form length (do you really need someones phone number when they sign up for an account?)
- Form layout (I want to ... vs choose from this dropdown)
- Field order
- Field width
- Natural input (My name is Lichai Cohn) vs. system input (My first name is Lichai, my surname is Cohn)
- Labels above or before inputs.
- Help (per field help vs. tooltips vs. general help).
- Error reporting (validate when typing, mark specific fields, give help per field).
- Required and optional fields.
- Multi page form vs single page.
2. Call to action (CTA) A/B test
Call to actions (CTA's) appear on every website and are usually the number one A/B test to consider. The main thing we want to do is make our CTA's stand out. Different target audiences will appreciate different CTA's. Some things you could test:
- Colour (both text and button)
- Visual style
- Copy (Will 'Sign up' work better than 'Become a member'?)
- Text length
3. Title A/B test
Titles act as a visual cue. They introduce the visitor to a section by shortly describing what's coming up next. Titles should be clear enough for visitors to quickly scan a section and decide which part interests them. Some things you could test:
- Font (font, size, leading, kerning etc.)
- Copy (factual vs. free, positive vs. negative containing jargon vs. understandable by everybody)
4. Text block A/B test
Text blocks are used in a variety of ways. From short introductions to lengthy articles, from help texts to terms and conditions. In general you don't want your text blocks to grab too much attention while still allowing a visitor to comfortably read them. Paul Olyslager has a nice article on the subject. On to some testing ideas:
- Font (font, size, leading, kerning etc.)
- Line length (characters per line)
- Line height (space between lines)
5. Image A/B test
Images can be used in a variety of ways. An image can be an attention grabber, but it could also be used as a page background or just an element to give the right atmosphere. You could test:
- Ratio (width vs. height)
- Colour vs black and white
- Contrasting with other content
- Use image as attention grabber vs. accompanying image like a background)
- Visual direction (usage of rule of thirds or golden ratio)
- Persons (should the subject look into a specific direction, should it be a man or a woman, young or old etc.)
6. Video A/B test
Broadband internet and YouTube have made videos on web pages very popular. They are used in a variety of ways, from attention grabbing content to page backgrounds. Some things to consider and test:
- Autoplay vs. click to play.
- Size (both in pixels and in mb's)
- Video length (in seconds).
- Placeholder image.
- Throbber (when waiting for loading).
7. Page length A/B test
In general shorter pages work better, but when displaying in depth articles it might be better to use longer pages. One-pager websites are also content for debates. What could you test:
- In page navigation.
- Visual markers (indicators there's more content lower).
- Visual indicator (a 'you are here' sign).
- Auto scrolling.
- Split up a long page in multiple shorter pages.
8. Menu A/B test
A menu can be any shape and size. With the introduction of responsive design a couple of years ago the 'hamburger' menu is widely used, but this doesn't make it less controversial. Test and see if your visitors understand your menu:
- Iconography vs. text.
- Always visible vs. hidden behind click or hover.
- Type of menu.
- Amount of items.
- Visual hierarchy (main menu and sub menu).
- Position on page.
9. Overlay A/B test
Many visual designers love overlays because they can add extra content to a page without taking up any screen real estate until they are loaded. Anything debatable should provide for enough opportunity to test:
- Using an overlay vs. page reload vs. content appearing in page.
- Make page unusable when overlay is open.
Allow user to easily close overlay.
10. Carousel A/B test
Carousels are both loved and hated. The lovers like the idea of using less screen real estate, while haters indicate content is hidden. Plenty to test then:
- Use a carousel vs. not using one.
- Show thumbnails of slides.
- Show navigational like circles.
- Auto sliding to next slide vs. waiting for click.
Need more ideas? Give us a call
020 - 262 99 09