A/B Experimental Test | Measurement Fundamentals | App Marketing | Udacity
- Articles, Blog

A/B Experimental Test | Measurement Fundamentals | App Marketing | Udacity


So now that you know what you’re
expecting to change through your hypothesis and what you’re going to
do, you need to run an experiment. You can do this using an A/B test. This will help you compare two different
versions of your webpage or app and learn which one performs better. You can use a Control group and
an experimental group. Having a control group can show you
how effective your campaign is. Now, the control group consists
of randomly selected users who are representative of
the population you’re targeting, and who do not receive the change to
whatever campaign you’re running. So being representative of
the population means that if 50% of your app users are male and
50% are female, you need to have the same demographic in
your randomly selected Control group. Now everything except the one change
you’re targeting should stay the same. For example, if you want to test
the effect of moving your sign-up button to the top of the screen, all messaging, colours, looks and feel of the site
should stay the same, except for moving the sign out button
to the top of the screen and only the people in the experimental
group will see these changes. So the people in the experimental
group will see the sign up button at the top of the page and
the people in the Control group won’t see a change in the location
of the sign up button. This way you can see how the people in
the two different groups perform and which button placement would
result in more sign-ups. Now your Control group
doesn’t need to be huge. The key is that it is representative
of the target population. If you have a really, really large
population like thousands and thousands of people, you might only need about 5% of that
population to be in your Control group. However, if it’s a smaller population, you may need closer to 20 %
to be in your Control group. To determine the best sample size for
your experiment, there are a ton of free online sample
files calculators you can check out. Now to run A/B test, there are a ton
of different tools that you can use. For example, Optimize Me allows you
to make changes on your website and will run and
host the experiment for you. So, let’s hear from
an expert about A/B testing.

About Ralph Robinson

Read All Posts By Ralph Robinson

Leave a Reply

Your email address will not be published. Required fields are marked *