Wait, nobody said there would be a test! That’s OK. This isn’t that kind of test. This is more in the nature of a scientific experiment to determine the better of two alternatives. To get the best results, there should only be a difference between A and B of a single variable so you will know what to attribute any changes to.
Forthwith are the 5 W’s of A/B testing.
Who Should Test?
Marketers who want to improve their lead generation and conversion ratios, that’s who. If you are one of the 72% or marketers dissatisfied with your conversion rates, A/B testing is for you.
If you are one of the 28% that is satisfied with your 1-3% conversion rates, you had better read the rest of this before your boss does.
What Should Be Tested?
That’s the great thing about A/B testing; it is a flexible tool that can be used in almost any area of marketing measurement.
· Calls to action · Click through rates · Newsletter open rates · Choices between two pieces of collateral · Ad text
This is by no means a comprehensive list, but you get the idea. A/B testing gives you a tool for real-time measurement of a piece of marketing in a way that gives you immediate feedback on better ways to present your wares.
Instead of testing one variant for a month and then changing to another and waiting to compare, digital marketing makes it easy to send out specific items to particular groups.
When Should You Test?
You should be testing all the time. Each iteration should give you a way to improve the activity rate for your chosen piece of marketing. As the test goes along you find what works and discard the other or use the insight to tweak the weaker performer, possibly to the point of surpassing the leader. Then you go back and tweak the one in second-place to perform better. This is an incremental process that can add up to big wins over time.
Where Should You Test?
This process needn’t be limited to online marketing. Similar tests can be run for telephone campaigns with scripts, print campaigns that appear in different media, and direct mail campaigns. Anytime you have the ability to send out two different versions of the same piece of marketing and can adequately measure impact, you can A/B test.
Why A/B Test?
Because businesses that incorporate A/B testing are twice as likely to see an increase in conversion rates. The whole idea of marketing and business is to get customers to buy your products and services. Thus it follows that the more visitors convert to leads and the more leads convert to customers the better your business.
And all for the time it takes to develop a call to action, an ad, a newsletter, a landing page, or any other marketing content/collateral and to create a second one with a single variable changed. After that you simply gather data until you have a statistically significant sampling and determine whether and how much that variable mattered in terms of conversions.
Bonus! How to A/B Test
Actually, you probably figured it out from the 5 w’s but here are some hints to make the process go more smoothly and for more effective, actionable results.
· Change only one variable between the two versions. This can be a small variable, like the choice of color of a CTA button or a larger one like testing long form marketing against short. Multivariate testing is possible but you need large numbers of data points and more time to set it up.
· If the results don’t seem to be correct, there is nothing wrong with repeating the test.
· Define a specific goal for your test and how to measure it; otherwise you won’t really know anything.
· Your “A” version will be your control. This is the version you typically use or that has always been in your marketing. The “B” version will be what’s called a treatment. It will contain the changed variable.
· Promote the content so you can gather results. This will define a starting line of sorts for the test.
There you have it: the 5 W’s and a bonus H of A/B testing. It isn’t complicated. It doesn’t have to be time-consuming. And you will know fairly quickly how to improve your marketing to boost your numbers.