May 10, 2017 (Updated: May 4, 2023)
A/B testing, also called split testing, is a powerful tactic that exploded with the rise of the internet. For those unfamiliar, it involves comparing the performance of two subtly different versions of something customers will see, such as web pages, newsletters, or landing pages, by sending an equal number of people to both versions at the same time. Every other visitor of a website front page, for instance, would see the B version. This allows you to test any change, so you can stick with the clearly better one and A/B test other factors.
There is no shortage of advice when it comes to A/B testing, but most of it relies on different graphics, organization, changes to video content, types of buttons used, and other things outside of the actual written copy. You could be missing out on major success if you don’t test your written copy, but you must have a solid plan, so let’s take a look at some helpful tips.
Image via Flickr by Alejandro Pinto
Split testing will often push a business to refine its copywriting into the strongest chain of words possible. This usually means brevity: cutting sentences down and cutting paragraphs up to allow more and larger text. It’s considered ideal to make the reader scroll with their mouse, trackpad, finger, etc. as little as possible
However, some websites tell a story in big, spaced text, making the reader scroll on and on to build investment by the end. Others combine headings with huge, beautiful pictures, and others use a horizontal slide-show to deliver lots of information with no vertical scrolling. Different businesses have different audiences, and this is why you should test not only different word counts, but also vertical spacing and length, to see what inspires the most engagement.
Beyond trying to communicate more, less, or different information, how it comes across should also be tested. It’s hard to tell if something is amusing, encouraging, or exciting until you test an alternative version. A piece of writing can be re-done to give a different tone, or to strengthen or weaken the current tone.
For example, a business-to-business company with a friendly attitude may end up not as engaging as a more serious and direct one, but there’s no way to know without testing. Different tones might even be more effective depending on the type of content you’re testing, such as more assertive copy on a purchase page, but gentler copy on the product description or advertisement. This is particularly important for businesses who want to maintain a brand image and have a tone tied to that concept. Don’t be afraid to try adjusting the feel of the message in some areas.
Raising or lowering the grade level of a piece of copy is a good way to gauge what type of writing complexity is most comfortable for your customers. For instance, if your business is for children and parents, it can be tricky finding what grade level appeals to both. Inexperienced writers will sometimes overload their copy with large technical words, pushing the grade level to something even hot leads would find boring or unclear.
The Hemmingway app is a free tool you can use to find the average grade level of any piece of copy. Just paste the text into the app and it will give you the grade level plus information about its complexity, word frequency, and more. Though the app is made to encourage prose of a particular type, the data it offers will give clear ideas on what to consider changing.
This is a good rule of thumb for doing split tests on cold traffic generation. Many companies new to A/B testing use more of a certain keyword on their B version of a webpage, and see if it gets more hits from Google searches. The tip, however, is to maintain a balance.
Search engine optimization is made to draw attention, making your page more likely to show up in a search, but it’s not a conversion tool. The keywords have to be ones that draw the right people, and the copywriting itself still has to feel natural and compelling, which is unlikely if it’s stuffed with the same one or two words or phrases. Just remember that SEO isn’t everything, and there’s no point in grabbing a thousand more sets of eyes if it doesn’t lead to any more clicks.
Image via Flickr by paigggeyy
In A/B testing, the A version is called the control. The point is to compare the control to a B copy that was only changed in a single way. If that proves to be a strong positive change, you then turn the B copy into the control for your next test. However, you must resist the urge to switch your control constantly based on small improvements in performance.
For example, imagine you A/B test a webpage by removing the headline in version B, using only a subtitle, but you also had the idea to test a shorter headline. That would be version C. B performs slightly better and you impulsively make that your new control, moving on to test another element. Now you won’t ever know if the simpler headline, version C, would have done even better. There will likely be multiple times where a test leads to a small improvement, so don’t be too reactive.
As a final tip for all forms of A/B testing, don’t check the data every day. You’ll lose your mind quickly and make snap decisions. The more test subjects who visited saw copy, the more objective conclusions you can make.
Hopefully you’ve gotten some ideas for split testing your copywriting with these tips, and also have a stronger sense of how powerful it can be. Although there are many stories of how changing a tiny visual part of a page led to massive revenue gains, there is equal potential in the readable portions. Keep in mind that A/B testing on any kind of effective scale will require paid software and time. Make it count by hiring skilled and experienced copywriters attuned to your business and goals.
More from the author: