Improving Conversion with A/B Split Testing
How do you know what works on your website and what doesn’t? Usability is a beast that can’t be controlled until you test different scenarios and use hard numbers to make your decision. We as people are very adaptive and fluid, we learn and think differently than the next person besides you. As Software Advice guest blogger Nathan Yerian of Adhere Creative mentioned in this post, marketing is a discipline in which art meets science and companies need to commit to best practices to optimize conversions.
Depending on your target market, conversion rates vary from 1%-6% with an average of 2.3%, the electronics arena converts as low as 1.1% and catalog based e-tailers as higher as 6.1% and that’s probably due to the traffic driven to the website from catalogs. In short, conversion rates are really low and its really hard to get any higher than 1%, therefore, its imperative that you do all you can to decrease your abandonment rate and one way to do so is to test, test, test and test – test everything.
On October 2011, Marketing Charts posted a chart that depicted the use of methods to improve conversion rates. A/B testing was used or will be using by 67% of the participants in the future.
What makes up a conversion?
Your website conversion is simply a goal and the rate is the percentage of people that meet that goal. Its an action that you deem valuable. Once your website is established, you should know your primary objective. For shopping sites, its sales, for blog its time on site and amount of content read or newsletter subscriptions, a download and so on and so forth. A conversion rate truly describes that final outcome of a visit to your site. It could be one or a series of actions and that’s why FoxMetrics was built from ground up to be completely event driven so that you can track these actions easily.
If a conversion on your website is a series of actions, building a conversion funnel would be your best choice to measure the progress toward meeting your goal. From within FoxMetrics Dashboard you can define and create a conversion funnel report with as many actions as you need easily in a few minutes.
What is A/B Split testing?
A/B testing is the terminology used to define the a way to compare a baseline sample against a variety of single variable samples. It short, it gives you the opportunity to test two versions of the same thing, version “A” and version “B”. Your visits are randomly assign to each version, note that its crucial that its done in parallel to eliminate any outside factors that could change anything. The results of your tests are only as good as the consistency of the environment; and randomness is the basis for the probability theory that is the heart of the statistical analysis of your test results.
Version “A” of the test is usually the control or champion – it’s the way it is at this moment, whether it’s a page, button, headline or image. Version “B” on the other hand is the challenger, and its up to prove that it’s a better version that version “A”. If the challenger wins, then its replaces the control – very simple.
Its not always A/B, its not uncommon to see A/B/C or A/B/C/D tests where the control Version “A” has multiple challengers and the winning challenger replaces the control.
There are more complex tests that could be use to test different variables such as multivariate, however, A/B by design is as simple as it gets and the results are usually effective and accurate. Implementation is the usual pain points for most testing platforms – testing with this style should be simple as you all you need to do is specify the control and provide the different challengers.
Its also flexible and easy to analyze, meaning the variables that make up the test are few (number of visits, visits that saw the control, visits that saw the challenger and their conversions).
- Don’t measure just the goal, the path is also a very important
- Once you’ve created a funnel, ensure that you work on decreasing drop-off rates on steps that higher than others, starting with the highest
- Create an effective call to action button as you see fit and test it
Things you can A/B test:
- Page layout
- Headline copy
- Interface and button colors
- Form field layouts
- Images and Graphics
A/B Split testing tools in no particular order:
- Google Website Optimizer
- Visual Web Optimizer
- HubSpot / Performable
- Adobe Test & Target
Other ways to improve your website conversion rates:
- Add a direct phone number on the top of your web pages
- Simplify the order processing funnel
- Improve on your website navigation
- Add Testimonials
- Create content that is very specific to your website and its topic
- Call to action buttons
- Pricing, discounts, free shipping and other promotions
- Social Proof