What is AB Testing? Application to trading
What is AB testing? Have you ever heard of it? For those of you who are learning about AB testing techniques in trading, this is the article for you, it is strongly applied to trading, for example you are using 2 chart analysis methods at Primexbt exchange. , you will use trading methods in turn to help find an effective trading plan that works for you
I. What is AB Testing?
AB testing is a testing technique, by dividing a main object into two versions, A and B, to test which version the user prefers through how the user interacts with each version. In a word, you can simply think of AB testing it as a horse race, you will put your pairs of horses in turn to the ring, and see which is the fastest horse, you will use Use that horse for your big, official matches. As shown below,
In the illustration above, we create an experiment, by creating 2 versions of the website, and see which website users will prefer. Obviously, we will find that test version B (version B) is better than test version A (version A), where version B has more than 37% interactions, while version A has only 23 %. The engagement here depends on how you test it, like click through rate on the page, how long people stay on the page, or something like that you want to test…
Each version of AB testing is mainly based on a simple formula:
Engagement rate = Total interactions / Total visits. Let's say your website has 100 visitors today and the number of registered members on the site is 20 people. Thus, we will have a user interaction rate with the website of 20%.
AB Testing data example for Primexbt exchange registration page, see which landing page has the most subscribers.
Version Total visitors Registered visitors Engagement rate
Original 1406 356 25.32%
Version A 1.488 372 25.67%
Version B 1392 425 30.53%
II. How to create test versions AB Testing
As mentioned above, there are many ways for you to come up with test samples for your website, here I suggest to you some of the ways I often apply as follows:
Move the layout of the website, or remove some elements of the website
Change website title, change text color
Resize web page
In my personal opinion, you don't need to think of too many test samples, myself with just 3 ideas as above, I have created enough ways to test a certain website.
III. Notes when using AB Testing
When using AB Testing to test a certain website, keep in mind that the conditions for testing must be in a similar environment, I list some notes as follows:
For example, you cannot test Version A on Monday and test Version B on Saturday, Sunday, because maybe Saturday is the weekend, so user behavior will be different from Monday.
You can't experiment with a small number of statistics, for example, you can't just take 10 visits, but you can evaluate user behavior, according to the statistical probability learned in school, the statistics The larger it is, the closer the result will be
IV. summary
So, together, we have learned and understood AB testing techniques.
This is a very useful technique for increasing the usage rate of any website or any application. Large websites, e-commerce sites are applying this technique often, typically we apply this technique a lot and are most often seen in google adsense, used to increase the click rate of users' ads or other websites. E-commerce sites use this technique to know what products and types of ads users like, thereby bringing higher profits for their company.
I pause the article here, in the next article, I will introduce to you how to apply and some examples of AB Testing in Bitcoin cryptocurrency trading. Love!
Comments