Here we are once again discussing what is going on in the marketing world. Not only what’s going on, but what is actually helping businesses succeed when it comes to conversion rates. The topic that I am going to discuss is A/B testing.

First and foremost, A/B testing is pretty cool stuff. In short, A/B testing is basically figuring out which web design will create the most conversions. You might be thinking, “But how does one measure the metrics for such a complex situation that originally requires one to say, ‘we think’, instead of, ‘we know’?”. Well fellow unique visitor, a fantastic company by the name of Optimizely, created a A/B testing system, so that you don’t have to guess anymore. What a great group of people.

A/B testing allows you to figure out whether or not the new additions to your website are going to increase conversions. In the olden days, you had to play the guessing game. You would sit around with your co-workers and say things like, “I think that if we change the background color to something that eluded to happiness then the customer will purchase the item because they are in a good mood, and want to keep that good mood going, because the more you buy the happier you get, right???”. Now you can actually test the waters for the new “happy” design and see if it is worth changing the site code. A/B testing measures customer sign-ups, downloads, purchases, or whatever your goals as a business are. A/B testing takes the guessing game out of the optimization equation and backs up your decisions with data. I once had a wise-marketing professor tell me, “Marketing is all about data now”. He couldn’t have been more correct.

So how is A/B testing performed? A/B testing involves testing the original website version against the updated website version. An A/B tester basically splits down the middle of the visitors that go to your website and interact with it. Half go to the original version, and half go to the updated version. Then once they have sufficient data to say that one version is better than the other, they let you know. If you are still a little confused on the subject, take a look at the diagram below that Optimizely created.

I myself have a clothing company called, SomeFire. During its lifetime as a business I have sold four different articles of clothing. My first collection was released in 2012 which was a Givenchy-inspired tee shirt. My next collection was tank-tops, which was dropped in 2013. My past collection was long-sleeve t-shirts and hoodies, which was released in 2014. This year is going to be a big year for SomeFire, because I am cooking up a special surprise for the consumers. When I release my next clothing collection, I am going to use A/B testing in the checkout funnel. I am first going to try something simple, like changing the color of the “purchase” button. Usually I have a black “purchase” button but I am going to try something small, like a green “purchase” button. Once I have collected data on the most effective “purchase” button, I will move on to bigger things i.e. website layout.

Below is from the last collection of long-sleeve t-shirts.

I like to think of myself as an aspiring soon-to-be marketing master. Which means that I go the extra mile to really learn about marketing topics. For instance, I read an article called, Practical Guide to Controlled Experiments on the Web: Listen to Your Customers not to the HiPPO. When the title mentions HiPPO, it is not talking about the ever-so ugly Hippopotamus amphibius. HiPPO stands for Highest Paid Person’s Opinion. In this article they discuss lots of different ways to make sure that the controlled experiments that you are running, are being done right, and evaluated correctly.

In the article they really get down to the nitty-gritty when it comes to the data portion of A/B testing. They evaluated two different check out processes, one had a coupon option, and then one had no coupon option. When they tested the coupon option they lost 90% of their revenue. The reason this happened was because the customers would reevaluate their order and think about if they are paying too much for this product, or possibly go browse the internet hopping that they would find a coupon. This is a valuable lesson when operation an eCommerce site. You want to make sure that the design choices that you make don’t prevent sales.

While reading this article, they used an example of A/B testing that made me really want to think of my own, so guess what? I did. A good way to think of A/B testing is through parenthood. When parents have their first born child (controlled group) they feed that child certain foods, give him/ her certain responsibilities, give them certain freedoms, and so forth. They then mentally measure the toll that it took on that child whether it was good, or bad. So now when the second child (the treatment group) is born they can try different things on that child. Maybe its less freedom, or more responsibilities, and etc. They then can compare the two and figure out which one seems “all there”. So now when the third child comes along they know exactly what to do and what not to do.

A now famous analystics mastermind by the name of Amelia Showalter opened up a new way to do A/B testing for political campaigns. Showalter was behind the “Hey There” emails from President Obama. During the 2012 election, Showalter and colleagues of her’s tested different ways to create an effective email campaign to raise money. They were very surprised to find that people liked the “ugly” e-mail format. Traditionally one might think that a potential donator wants to see the most professional email from their president, which is in fact the opposite. After extensive A/B testing they found out that the headline of, “I will be outspent” would generate the most donations bringing in $2,540,866.

Businesses can learn a lot from Showalter when it comes to A/B testing. You need to try the most unorthodox things to try and see what the outcome would be. Showalter did not think that the headline of “I will be outspent” would be a hit, until they tried it. Showalter did not think that bright yellow portions of the email to attract the eyes of the potential donator would be a hit, until they tried it out. Showalter did not think that having HTML style “donate here” button’s would be a hit, until they tried it out. Businesses need to realize that they can reach more of an audience if they implement A/B testing into their business, and try out different options. But businesses need to realize that their email list doesn’t want to be bombarded with emails 10 times a day. I myself recently had to unsubscribe to the Nordstrom email list because I was getting blasted with emails 10 times a day. In return, it made me want to do the opposite of press on them.

I hope that you learned some valuable information about A/B testing, and that you use it when you feel that it will be effective, and a great way to get more from you customers. Until next time fellow unique visitor.

//