At the goal line

Thoughts and opinions on sales, marketing, social media, digital content, technology and social business.

A/B Testing: The Science of Marketing

Posted by Rick Lambert on Thu, Dec 13, 2012

A B Testing The Science of Marketing     

You may have thought you left chemistry and biology behind in high school, but the skills you learned over bubbling beakers and lab report writing actually do apply to real life – your teacher was right.

Any time a company deals with people there are countless opportunities to conduct empirical studies to evaluate how your company can get a better response from customers.

If you want to make a website that is appealing to visitors than you have to be able to track their feedback when you make changes.  This can be done very simply by following these steps:

1.     Identify your Research Question

  • Before you conduct a study, you must always identify your question.  In order to make the process as simple as possible it is best to pick a simple yes/no choice also known as an A/B Test. Even though every A/B test is unique, certain elements are usually tested:
    • The call to actions (i.e. the buttons) wording, size, color and placement,
    • Headline or product description,
    • Form length and types of fields,
    • Layout and style of website,
    • Product pricing and promotional offers,
    • Images on landing and product pages,
    • Amount of text on the page (short vs. long).

2.     Set up the Study

The study should, like any good lab experiment, have three groups: a control and two test pools.   Decide how you are going to set up your study (are you testing buttons, subscription forms, privacy settings, payment options etc) and evaluate how best to carry out the experiment.  Will you need to create new pages to add to your site, will you need to add buttons, links or reformat visually?  These are all important things to answer before you conduct the test, otherwise there could be bug that can corrupt your data and the entire thing will be a potentially expensive waste of time. 

Generally there are two ways to set up the A/B Test: one, replace an element on the original page or two, redirect to a new page with the new features.  Keep in mind that each change will require two versions – one for each test group.  For each you must assign a conversion goal – essentially a way to track a successful test that the A/B Testing tool can use to record what version was shown to the visitor.

3.     Conduct the Study

  • This is the easy part – run your test making sure to make all changes before the page uploads and then let the test run for the set period of time.

4.     Review the Results

  • The A/B/ Testing tool has collected a valuable set of data for you to review to see what variation garnered a better response from the visitors.  See the results and then make the necessary permanent changes to your website. 

Some Things To Keep In Mind

Though A/B Tests are an easy way to get feedback they are not fool proof and there are some important dos and don’ts to remember:


  • Never conduct your tests at separate times.  The variable and the control group should be active at the same time and with traffic evenly split between them.
  • Don’t rush.  Make sure you give your test an appropriate amount a time so you get an accurate sample reading.  There is no point in rushing the results if you will just have to redo the whole thing.
  • Don’t try to get old dogs to learn new tricks.  You are much more likely to get a better evaluation from a new visitor than an old one who has become accustomed to a certain aesthetic.  Try to differentiate between old and new visitors and direct the latter towards your A/B Test.
  • Don’t let your personal feelings overrule the results.  The public is unpredictable and may give you surprising results, but this doesn’t mean that you know better.  Make changes according to the feedback – nothing more.


  • Do be consistent and show your visitor the same variation.  The test tool should have a way of tracking who has seen what and should be set to keep the variation consistent.
  • Keep trying the A/B Tests out.  The first one may turn out a dud result but the more you do them the more effective they will become until you get reliable results.  The key is to persist and not become discouraged by bad readings.

The point of an A/B Test is to poll your audience without them necessarily knowing.  It probably gives you more honest results than if they were aware they were being tracked and allows you to tailor your website to suit their preferences.  So start A/B testing and see the results start to generate improvements to your site.

Click here for your free e-book on the 7 tricks of the trade to enhance your site:

enhance website

Topics: content creation, lead generation, A/B testing, call to action, effective marketing strategy

Subscribe via E-mail

describe the image

Latest Posts

Posts by category

Stay Connected - Follow Us