AB Testing: Introduction To Better Conversion Opportunities

Image Source: https://www.lcn.com/

Which one of the above mentioned Call To Action (CTA) buttons do you think will give you the best result? Already thought of the answer? But before you tell me the answer, I'd like to know the basis of your judgment. Is it your guts or instincts? Well, sorry to break this to you, but there are very less chances of you being correct as your judgment has an inherent bias in it; bias of your personal experiences. While your answer/judgment might work for a particular website, it might also fail for a different kind of website. This is where A/B Testing comes in handy. In the previous blog, we spoke about Conversion Rate Optimization (CRO). For every process, there are tools and equipment that help us ease things or enhance the output and that is what is A/B Testing to CRO.

Wondering what A/B Testing is? A/B testing is a method of comparing two versions of a web page or app against each other to determine which one performs better. For this, we create two versions of the same web page, Version A and Version B (hence the name A/B Testing), which will be shown to the visitors at random. Of the two versions, one would be the original live version, which is already present, while the other would be the changed version, which is to replace the live version. You might be questioning yourself the need for all this shenanigan. Simply put, this is just like a test run which allows us to see the upgrades we are implementing and the conversion and revenue data received for the said changes before the upgrade is finally put in place.

Example- You are an e-commerce website owner and you have a website that is giving you a decent conversion rate. On the product page, the Call To Action (CTA) you have is BUY NOW. While reading a blog about copywriting, you get to know that it’s better to have a CTA button copy as Add To Cart. Soon after you implement this change on your product page and now your conversion rate has taken a hit. If you would have done an ab test between Version 1 (CTA button copy as BUY NOW) and Version 2 (CTA button copy as Add To Cart) you would have had a concrete result backed by data to see if this will be a good change or not after which you take a decision and your conversion rates wouldn’t have taken a hit also.

The founder of the e-commerce colossus Amazon, Jeff Bezos said it right, “ If you double the number of experiments you do per year you’re going to double your inventiveness.”

Image Source: https://www.digimar.com/

A lot of people think of A/B Testing to be a waste of time. Contrary to popular opinion this is one of the most essential processes in the world today. If there is a possibility of going through an A/B Test cycle, you should always opt for it before taking any decision about any proposed change in the structure or design. In times when companies are pulling no stops to optimize their conversion rates, A/B testing is something that is sure to give you an edge in the market. Basing your implementations for increasing conversion rates on guts and instincts can prove to be counter-productive. Hence the need for A/B Testing so that, even though the changes might have their origin in your gut/instinct, they are backed by solid testing and data to prove their viability and effectiveness. What works for one particular website, might not work for another website even in the same vertical. Customer behaviour differs from website to website hence implementing a change from another website without proper A/B Testing is not a good strategy to follow.

Few points to remember for a successful A/B Test run :

There is a big difference between A/B Test and Multi-Variate Testing

Image Source: https://www.convertcart.com/

These are the two types of testings that are usually followed, but skepticism about which one to go for is always afloat. Multivariate testing uses the same core mechanism as A/B testing, but compares a higher number of variables, revealing a whole bulk of information about the variables and the way they interact with one other. While A/B Testing allows you to have only two versions, multi-variant testing allows to you have as many versions as you want. 

Now the question that crops up is, why then shall we go for AB Testing instead of MVT? In a/b testing, ideally the traffic split is 50-50 percent while that of multi-variant testing depends on the number of variants put to test. Let’s assume that there are three versions being put to test applying the MVT process, then the traffic split is going to be 33% for each. Given the fact that we need ample data to see which version is the most feasible, we need to have good traffic on the website, without which the MVT process will take a long time to provide us with the required data to reach a conclusion regarding the variables. hence, one slight drawback with MVT is that the more versions you have, the more split in traffic and the more the need for additional traffic to acquire data.

Let's understand this with an example. Let’s say you want to test your CTA copy, color, and position. The easy way to go about this would be to opt for MVT and get the result and be done with it. You might have completed the test but your results are not going to be anywhere near accurate. The people who land in the traffic split of the CTA copy test might not react the same way as the people in the CTA color test.

The optimal thing to do here would be to go ahead and do an ABT for CTA copy, find the best suitable version, and then let the test for CTA color go live. Once we find the winner from the CTA color test then the process follows for all the variables that need to be tested and updated. IN this way, at every stage we are doing the requisite testing and are being provided by the best possible result without causing ambiguity and redundancy. If you are thinking as to how do we narrow down the variant to use for testing, especially in case of any cosmetic change like color, then I'd suggest you to do a thorough market research and understand the psychology behind all the cosmetic features in a website. It will not only help you in narrowing down your search for a variant but also help you identify any adverse cosmetic features that you need to change as well.

Pro Tip- Market research always adds value to tour marketing strategy. Investing more time in it will always give you an edge.

You get one life but unlimited test

A/B testing is a continuous process and no one test is going to give you the magical conversion rate that you always wanted. People think of doing an ABT in which they check a wide array of variables in one go. They actually intend to see a sea of change in the conversion rate in one go. The problem that surfaces in such a situation is that it will not be possible to pinpoint the element which actually worked and gave you the uplift which you saw. Similarly, if the test fails, you will not be able to gauge the variable(s) responsible for the failure. To avoid such a mishap, start small and precise so that the change and the data can be correlated and implemented accordingly. This will not only give you precise test results but also give you an opportunity to run more and more tests. Never stop after one cycle of ABT. There are varying number of things to play around with and test on your website. Why hesitate when you can do an ABT!

Run one test per page at a time

While there might be many opportunities for you to run more than one ab testing on a particular page, it is logical to avoid such temptations. Running more than one ABT on a page, for the same segment of users, will only result in you being boggled about the element which is giving the push to the conversion rate. For instance, for an e-commerce company, you run one test on all product pages where-in you change the CTA copy, and simultaneously you run another a/b test in which you change the product description style on the product page for the same segment. While these tests would go side-by-side, you will not be able to tell which of these is actually influencing the conversion rate in a positive or negative way. Patience is always the key to success. If the segment is different then you can run more than one ab test on the same page. Plan your conversion rate optimization strategy in such a way that you do not encounter a situation where-in you have to run more than one test on a page.

There are a lot of things that you can play around while a/b testing. Segment users based on traffic source, devices, time spent on the website, etc and then run your ab tests to offer a better experience to all your users. Users are bound to convert more if you provide them with a one of a kind experience. Moreover, if their requirements are met without any hassle, then their conversion would not be an issue. If you have an idea, test it, validate it, and then test some more. The more the number of tests, the better is your CRO strategy.

Check out how the former President of the United States of America, Mr. Barack Obama used a/b testing in his 2012 presidential campaign. A/B testing is a weapon that surely has a lot of benefits for anyone who uses it. There are tools to help you run a/b tests as well. In today's day and age, there is content everywhere on any and everything. In the very same way there is an ab testing for any and every idea that you might come up with. For a better understanding of a/b testing, subscribe to CXL which is thorough and diverse and is sure to clear any doubt which comes up regarding this topic.