free customer wifi

How to perform A/B tests that help you optimize your web presence for conversions

Nothing is worse than investing heavily in lead generation marketing and branding your business online only to realize that you can’t actually convert any of the customers that come to your website. In today’s frantic online world, the battle for consumers’ attention is even more intense, meaning that if you can’t convert, you’ll lose to the businesses that can. Conversion optimization – the process of creating a web experience that increases the number of conversions a website, landing page, sales ad, etc. gets – is an integral part of online marketing.

split_testing2The heart and soul of the conversion optimization strategy – A/B testing
At the heart of any conversion optimization strategy is A/B testing, or split testing. In an A/B test, you are taking two headlines, emails, sales ads or anything else in your online marketing arsenal and pitting them against one another to see which one performs better when it comes to converting leads into customers.

One of the most commonly performed A/B tests is for headlines. The test is set up so that one group of visitors to a landing page sees one headline, while another group sees a different headline. From there, you can open up your web analytics platform and see which headline led to a stronger performance in whatever metrics you want to measure. In this case, it’s usually your conversion rate.

A/B testing is an extremely important part of the marketer’s toolkit today. Whether you’re building a new website from scratch or have an existing one that you want to improve, split testing can help you make it the best it can be for you and your customers. Here are some tips and tricks you can use to get started running your own split tests and start optimizing your website to maximize conversions.

Do not make assumptions about your customers before you test
As a recent LinkedIn article on the subject explained, you want to start with a clean slate and assume nothing about your customers.

This is essentially step zero. The point of A/B testing is to figure out how your customers really respond to your marketing, not how you assume they do. There’s a strong similarity to the scientific method here: You’re doing an experiment and using the results to drive your knowledge and understanding of how something works. The results of the experiment are all that matter.

Figure out what you want to test
The first step in A/B testing is knowing what exactly you want to test. Kissmetrics wrote that there are two kinds of A/B tests: On-site and off-site. On-site tests, as the name implies, are tests you run on your website: Headlines, body copy, placement of graphics, etc. Off-site tests involve sales ads, emails and anything else that is designed to move leads from wherever they are to your website. Picking a starting point depends on what part of the sales funnel could stand for the most improvement.

Devote sufficient time to testing
Getting good test results isn’t an overnight process. You need to let your tests run for at least a few days to get an accurate picture of how the object of your testing affects the end users. An insufficient amount of time means that you will likely have a small sample size, which could lead to skewed results, and basing any sort of design or writing decision on flawed results can be just as detrimental to your conversion rate as not running any tests at all.

optimize2Know what your baseline is
Before you run any tests, you’ll want to know how your website is performing already so you can see how things change after your experiment. If you’re testing two different headlines, you should have a full understanding of how users interact with your website with the original headlines. Once you’ve established that baseline, you can test your new headline to see how things change. That’s what A/B testing is about at its core, and it all starts with knowing exactly where you stand before the test.

Don’t test more than one variable at a time
As LinkedIn noted, it’s called A/B testing for a reason – you don’t want to overcomplicate things by testing too many variables at once. Going back to the headline example, make sure you have your control and your experimental headline and nothing more. The point is to test one headline against the other, and testing too many variables at once makes this impossible.

Run your tests simultaneously
All of your testing should be done during the same timeframe, Kissmetrics wrote. In other words, you don’t want to test one headline on one day and test another the next because you don’t know how all of the other variables have changed over the course of a day. Be sure to split the traffic that sees your two variations at the same time.

Test small changes
The Harvard Business Review pointed out that inexperienced A/B testers often try to swing for the fences right off the bat by trying to split test an entire website’s worth of content. This takes a lot of time, and it can be a waste if the test ends up not working.

Keep your testing to the details – headlines, CTA placements, background colors, etc. These are quick tests that can net you big results if done properly.

optimize3Repeat some tests
Test results can often be skewed or simply unusable due to a slight nuance or because of seasonal effects. But you shouldn’t throw out an entire test just because it didn’t work when you first tried it.

The HBR recommended that marketers keep a log of all of the A/B tests they perform so they can return to it and re-test some of the things that didn’t turn out correctly on the first go around.

If you’re planning on starting a conversion optimization strategy, what are some of the things you’ll A/B test?

Scroll to Top