Okay, now back to where we left off…split testing.
You may be asking what is split testing?
The simplest way to think of it is a series of left or right kind of choices. For example, in one test, you may test whether visitors prefer the purple background or the blue background. The way the test is done is by showing a part of the visitors one while another part of the visitors see the other one. Then, we watch how many from each group do the desired action. Or in other words, we split to paths our visitors take to test where the end up, if that makes sense.
Let me say it another way…
If you’ve ever taught children, you may have noticed that they can have very strong preferences occasionally. Sometimes, those preferences are clearly distinguishable.
However, at other times, like if they are trying a new food, they do not yet have a preference established.
The way to learn their preference is to give them some of the new food and see how they respond. If they like it, you’ll know by observing the behavior such as facial expressions, words and actions.
Now, in Internet marketing, the difference is that you most likely won’t hear words or see facial expressions of your visitors. So, the best way to learn what is working best for your website is to just try it and see. But, there is something you should know.
Things You Should Know About Split-Testing
First, you must have choices to choose from. Without a choice, there will be nothing you can learn from your visitors.
Next, make it a simple one or the other choice. If you have too many things that are different in the two many options, you will not be able to say for sure that the background color (for example) was what made the difference.
Say, for example, the following three differences are shown all at once.
A) One has a picture of a tiger while the other is a picture of a horse.
B) one has bold red font on the submit button while the other one has plain black text.
C) One has a free report for people who subscribe while the other one has a “surprise gift.” After two weeks of testing, subscribers increased. But, why? Was it because the horse, or the red font? Was it because of the free report or because of increased advertising that was done during the testing period to increase website traffic?
You see how easily split testing too much too fast can leave you knowing less about your visitors than before the split test?
Finally, always remember to “rinse and repeat.” Over time, people’s experience changes. Their preferences change. In addition, the way your site gets listed in search engines changes. Consider making testing for what works best an ongoing business investment.
Just like any business person would conduct market research, one must be aware of the market conditions to be successful.
Recipe For Success in Split-Testing
Ingredients:
1) Equal and different images
2) Visitors
3) Time
4) Measure or tracking sheet
Instructions:
1) Decide what to test and for how long. At a minimum, I’d recommend testing graphics, fonts or other website visuals on no fewer than 100 visitors. Ideally, this will require less than a day to see results.
2) Set up the test. When using Google, they allow split testing landing pages and their system pretty much does it for you once you’ve told Google what parts you want shown where.
3) Invest some time letting your visitors see the website. Please know that the more visitors you can wait to visit, the more information you’ll be able to gather about your website audience.
4) Record results. For example, record the number of visitors, the number of subscribers as well as which ones reached the final desired destination.
5) Rinse and repeat. Do it again and compare two different things.
Please note that there is a cost to split-testing.
Please be cost conscious when are planning and executing split tests.
Please feel free to add any recommendations or comments below.
And, as usual, we are overcoming challenges one day at a time.
Have a great day,
Rachel.
Leave A Response