|
Table of Contents
• How to Do an A/B Testing
• What to Test During A/B Testing
• A/B Testing Services
A/B testing is a marketing tool that can be used to improve the effectiveness of a website. It is also called split testing. This method is designed to improve conversion rate, average order value, time on site, reduce bounce rate, and so on.
The essence of A/B testing is to compare a control page A (the original) and its created analog page B (the new page) with changes made. The entire traffic is split between pages A and B. Then the main metrics are compared. If the new version meets expectations and, for example, the conversion rate has increased, then all the traffic is directed to page B.
Going forward, when new changes are tested, the improved version will act as the control page A.
When working with split testing, it is important to properly evaluate the metrics of the control and new pages, as well as to properly distribute the traffic (it is not always necessary to split it 50/50). In addition, for proper operation, it is necessary to exclude the influence of segments on each other, reduce the impact of external and internal factors. Testing must be carried out in parallel and under the same conditions.
A/B testing is not always used to improve the landing pages of the site. Its capabilities allow you to run tests in other places as well.
How to Do an A/B Testing
Let's consider an example. The user needs to increase the conversion rate of the landing page, where the traffic comes through contextual advertising. Currently, this indicator is 2.5%. To achieve an increase, the user wants to change the design of the page. He creates a new page B and sends part of the traffic there. As a result, on the page with the new design, the conversion rate increased to 3% instead of the planned 4%. To further increase the conversion rate, you can change the text on the page, for example, come up with a new call to action. Now the page with the new design will act as the control page A. In this example, you can similarly try to change the contextual advertising itself (headlines, texts, etc.), work on their relevance, since the main traffic comes from there. When using A/B testing, it is important to get objective data and statistics.
What to Test During A/B Testing
A/B testing is an applied method. It affects various metrics, so the choice of what to change should be based on the set goals and objectives.
The following page components are most often tested:
Headings. Font, color, placement, content, their number - all this affects the retention and attraction of the visitor's attention. Sometimes just changing the font of the heading helps improve the overall appearance of the page.
Text. Similar to the headings. The text should be readable and visually pleasing. The amount of text, its font, color, placement - all this is no less important than the content itself.
Buttons. As a rule, buttons are the main element that attracts user attention. Therefore, changes in appearance, their placement, design can change the indicators for the better.
Images. Use high-quality and beautiful images. It doesn't matter if it's a product photo or a landing page background image. Even the slightest shift of an image can completely change the visual perception.
Page design. Poor design is often a factor that contributes to a high bounce rate. Test new layouts, templates for both the entire site and its individual elements. Look for the most productive option.
Business offers and prices. You can test not only the appearance of the page. Creating a new business offer is sometimes more effective than painstakingly changing the design, texts, buttons, and for months achieving an improvement in conversion.
These are just the most commonly tested components. In practice, everything is subject to change that, in the opinion of managers, can increase sales.
A/B Testing
Services
To conduct split tests, marketers use the following services:
Content Experiments,
Optimizely,
ABTest.ru,
RealRoi.ru,
Visual Website Optimizer,
Unbounce.
|
