But what is AB testing?
Simply put, you make two versions, A and B, of a webpage with the components you would like to test, then split your website’s traffic to both. In the end, the page that produces the maximum number of conversions comes out the clear winner. So yes, at its core, A/B split testing is really simple. Even if you are performing such simple Google Analytics, you are already far ahead of your competition. But if you want to leave absolutely everyone behind in the conversions race, there are some things to be kept in mind while A/B testing.
A/B Testing: Getting Started….The Right Way!
So this is the thing: it will take you forever to score any big changes if you keep A/B testing random stuff. You have to learn how to find the big issues in order to score big wins with the audience.
We aren’t completely denouncing the idea of each and everything on your webpage. The suggestion here is that you find out the money-makers first, fix them up, get them running and while everyone else is playing catch-up, you are busy tweaking the smaller things to optimize your website to the max.
A/B testing can give you an increase of up to 20% to your revenue and customer base. And when done right, it can provide another 10% increase in the figures. And the most amazing thing about AB testing is that the increase, once achieved, is permanent and you can reap the rewards for years to come. So you don’t have to do A/B testing over and again. Sweet deal, isn’t it?
So the question here is how do you do it right?
Remember when we said that random testing will only provide you with negligible results? Well, here’s how you streamline your testing to get effective results:
Ask your customers
Qualitative insights from your customers can immediately alert you to the problems prevalent with your website. But more importantly, you also get to find out why something is happening. You can witness the spots where customers quit the website, the features they use the most and who spends most of their money on what. This information is invaluable for the budding A/B tester.
Here are some great ways of asking your customers the questions that really matter: surveys, feedback boxes, user activity from Google analytics, usability tests so you can keep it simple and just ask them directly about which features repel them and which ones they really want to see on your website.
Figure Out How to Improve
Once you have found out the problems, it is time to brainstorm the solutions. Now this step here is not an exact science. So, here we ought to concentrate on the feedback we receive and predict solutions based on the customer’s comments. We are only creating hypothesis here, so there aren’t any guarantees yet about what will give a higher number of conversions and what will drive visitors away.
Launch the A/B Tests
We take the hypothesis developed in the previous step and implement them for the AB tests at the very end. The data obtained at the start has helped us figure out the most important changes. And now we are trying out the brainstormed solutions to obtain the secret to a successful website. This is how to chase big game in the A/B testing industry. Once you are done implementing the changes suggested by the A/B tests, your attention can easily be diverted to the smaller issues for fine-tuning the website.
So the A/B testing of taglines and banners will have to wait till the big stuff is out of the way. First it is important to figure out the best ways of acquiring and keeping customers that aren’t afraid of parting with their money at your website.
Tips to Remember
- Don’t go back and forth to the brainstorming session. Instead, once and for all decide which things should be tested first and then put them in a list. Work your way down quickly. Don’t ponder too long on any of them, some things will work, some things won’t. It is the end results that are important to your bottom line, so wasting time is not an option here.
- It is possible that your work gets sidelined because of these pitfalls. If a result doesn’t produce any statistical worth, forget about it. Statistical significance measures how confident we are about getting the same results no matter how many times an A/B test is repeated.
- New additions to your website can sometimes skew your conversion results simply because they are novel. Visitors just might visit them because of this novelty. Don’t let this marginalize your efforts. Give yourself an extra few weeks to figure out if the results are being skewed because of the newness.
- Use the right tools. Optimizely, Visual Website Optimizer as well as Google’s customer analytics are perfect for tracking the entire funnel to help you get better results for A/B testing.
Remember that your data is always evolving, changing with the customer’s demands, your business model and market adjustments. So your perfect website as of now will need some modifications a year or six months down the line. It is important to keep it simple when performing A/B tests for your website as it is very easy o get lost in the wonderful world of Google analytics and the hundreds of other tools that promise to make your website the best. Too many cooks spoil the broth. Just saying! So don’t get into multivariate testing unless you don’t have any other option.