The practice of A/B testing is becoming so common among marketers that you can easily choose to be part of a huge experiment in web optimization at any given point. That is why we love the tech and science behind improvements!
If you are one of those people who are running their tests every day of every week since the past few months and still aren’t getting any improvement conversions then you have come to the right place.
WHAT is A/B Testing?
A/B testing is also known as split testing. It a procedure of comparing two forms of a webpage against each other to figure out which one performs better. One form of the test A is run against a different form B of your page and the results indicate which one has the highest conversion rate.
A/B testing can be used to test everything including your search ads, sale mails and website copy. You can also implement new design changes, and improve your website’s conversion rate.
WHAT is Conversion Rate?
A conversion rate is the rate at which an action is performed on your site by the visitors.
In simple words, anything that we perform on a website can be termed as conversion rate; such as, logging in to the website, subscribing for its newsletter, downloading an application or making a purchase.
Conversion rates are measured by owners of the website so that they can become aware of how much visitors are actively using their websites. These measurements are then used to further enhance the conversion rate of the website.
Any page can be as tested long as you have a conversion metric to measure.
An example can be of testing a page where visitors leave the page immediately after landing on it without checking out any other pages on your website (high bounce rate).
As a conversion metric to measure, a measurable conversion where visitors view at least one other page after landing on your content page can be used in this example.
Let’ have a look at what the experts have to say about conversion rate and website optimization.
So WHY Choose A/B Testing?
It takes the Guesswork out of Website Optimization
Choosing and combining the most effective methods and of promotion can definitely infuse profitable marketing results instead of doing guess work to figure out what would work for your business.
Knowing how a change affects your business performance is better than guessing if it is going to work. A/B testing allows real visitors in REAL time!
Online marketers opt for a/b testing to optimize their websites because it allows them to learn what your visitors want!
It gives you an insight about what your audience wants. Sometimes new visitors prefer different kind of content than the old visitors and A/B testing can figure that out.
HOW does A/B Testing Work?!
Running an A/B test is actually not that difficult as long as you are familiar with what it means and which software allows you to run the test.
Take a look at Hubspot for additional questions relating to running an A/B test.
So how do you start? Here are Simple steps to design and run a successful A/B Test.
Well the first thing you should be doing is measuring the current performance of your website.
There are several websites that help measure the performance of your website such as Google Analytics, we will also show you how to conduct A/B testing using Google Analytics which is a light-coding-required resource of A/B testing.
Web Page Analyzer is another site that offers free webpage speed test for website optimization.
Next, redefine the goal of your website. Make a list of things that need to be increased or modified by going through the content of your website.
Tools for A/B Testing
Here’s a list of resources for you.
Optimizely is a code free resource of A/B testing.
Virtual website optimizer is a light coding-required resource of A/B testing.
Vanity is a heavy-coding-required resource of A/B testing.
You can also find 15 A/B Testing Tools for Mobile Apps here
Set up Target Metrics for Your Website before your A/B Test
A target metric is the target that a business has to achieve for it to be termed as successful.
For example, if a blogger produces ten fine posts per week, is that good or bad? Is the blogger lagging behind, exceeding expectations or just trotting alongside the goal?
Thus, a target metrics is defined. In the abovementioned example, suppose that the target metric is to publish forty blog posts every month.
Key performance indicator is a metric that can help you figure out how well you are doing against your objective.
For example, the objective of a blog is to produce quality posts, their KPI would be the number of quality posts produced per week.
Thus, it is important to keep your objectives in mind so that they can be constantly checked against KPIs.
Ask for Feedback from Visitors
Ask for feedback from your visitors. Set up surveys for your visitors to fill in to find out about the problems in your product or website.
Also ask for the rationale behind their problems and seek more qualitative data to get an insight about the reactions of your visitors towards your product.
Find out the trends that are addressed in your visitor’s feedback.
Use Segmentation for Digging Deeper when A/B Testing
Some of your experiments might be apparently giving poor results because they are not being individually tested for every segment or element. The experiment might have been successful for one element of the website but failed over all.
For example, an experiment might demonstrate that a variation of an application’s landing page is not working properly for Facebook users. However, on digging deeper into segmentation it is concluded that it is running extremely well for android users but failing to give results for users with Windows phone.
For this purpose, the person conducting the experiment must also be aware of the attributes of the target population of the website.
Are the android users more in number than the iPhone users?
Which country do they belong to?
Which browsers they use and what kind of websites they prefer most.
The following chart can help online marketers divide their visitors.
Here’s an article presented by Avinash Kaushik on further strategies for a good segmentation. In which he describes segmentation by visitors (source) of your website and their behavior (behavior) and by the products that they prefer to buy (outcome).
Figure Out What Should You A/B Test?
In order to conduct A/B testing, you need to find out the answers to the questions that arise from problematic pages that what do the visitors want that they are not getting in turn of which they leave the site?
Why aren’t they motivated to proceed and what hindrance is stopping them from further proceeding on the site?
Which element do you want to test first? Start by prioritizing those testing opportunities that offer the most benefits.
The following are some easy testing ideas. Pick yours!
Choose pages with High Traffic for A/B Testing
You need to look for pages with the most traffic. You might be tempted to optimize those pages that resulted poor in performance but the fact is that those pages that already perform well finish getting tested sooner.
While in order to check those pages which are performing poorly, the tests need to be run for long period of time. Consequently, those pages should not be your priority. Choose from the top landing, most visited pages. And even among those if you are confused on which to pick out, choose the ones with most expensive traffic.
Why should you run A/B tests on your Landing Page?
Because A/B testing offers online marketers with the following benefits,
- More sales and revenue
- Unveil your landing page pitfalls
- Lower bounce rates
- Higher conversion rates
- Eliminates assumptions and guesswork
- And last but not the least, to get leverage over the competition
Devote Time to Your A/B Testing
You can run only one test at a time to gain the most accurate results. So time must be devoted to testing.
Remember that the results can only be accurate if the test is run for sufficient amount of time.
You should be mentally prepared that depending on the traffic of your website, you might need to run tests from a time period of days to potentially a week.
Don’t run out of patience as test results can be inaccurate if the tests are run of insufficient amount of time or even for a too much long period of time. Try to gather as much information as you can on testing procedures so that you can be aware of any unexpected changes that occur in your results.
A/b Testing More than One Thing at a Time
If you want to run more than two versions of one element such as three versions of a layout, it can become complicated, such as testing A vs B, B vs C, and A vs C. it will take more time but running an A vs B test is more advisable.
Running more than one thing at a time, such as headline vs layout, is a multivariate test, and is more complicated to run.
Split Testing & Re-testing
You need to constantly track you performance as there is always room for improvement.
If one out of five tests enhances your conversion rate then it means that you are doing well. However, no matter how better you are doing, the digital market is extremely competitive. So, you should not just stop at one test. Keep testing for upgrading your websites.
Considerations to Keep in Mind When Conducting an A/B Test
Here’s a quick Checklist!
Prefer testing one variable at a time.
Run one test at a time for authenticity of results.
Don’t fall back after one failed attempt.
Don’t your test running for too long or too short period of time.
If you are conducting two versions of your homepage then instead of using a No Index meta tag on your variation page, use rel=canonical as you want the search engine to know that they are the variations of same URL.
Change as the numerical difference between the two variations and the amount by which one variation is larger than the other to avoid percentage confusion. Version 2 is commonly quoted when writing conversion rate improvements, but make sure of it to avoid later problems.
Use a 302 temporary redirection method instead of the permanent 301 version so that Google keeps the original URL in its index.
Avoid A/A testing as it takes more time setting up and running. Consider A vs B vs A split test or segmentation as an alternate.
Integrate A/B testing process with conversion rate optimization methodologies for better results
Do NOT conduct A/B tests when you do not have any traffic. Only conduct the test on highly viewed pages.
Always base you tests on a hypothesis. Segmentation is necessary to be carried out to become aware of the problem.
Conclusion : 10 Ways to Leverage A/B testing to Fire Up Conversions
Keep tracking your progress and don’t back down until you see the numbers moving.
A/B testing is not really that complicated, once you are aware of the procedure; find the right kind of softwares to do your work for you.
Follow our strategies for better conversion rates with A/B testing. Experimentation is the best test itself. Brainstorm, improvise and see what works for you!
We value your feedback so don’t forget to put your thoughts in the comments section!