EMBiz: How To A/B Test

How To: A/B Test

Cody H. Owens

Hi, Cody here, Content Director, Elevate My Brand and in this EMBiz, I'm going to talk about the six steps of strong and smart A/B testing.

First of all, what is A/B testing? Well, it is the term we give to this idea of coming up with 2 or more variables that you want to try out in your content and your advertising, whatever, in order to come away with learning so you can decide how to move forward with the best possible strategy and content in the future. There are 6 steps that we typically follow in AB testing in order to have the best outcomes.

Step 1 - Determine Your Objectives

The first is determining your objective or your hypothesis. If you don't know what you're testing and what you want to find out in the end, then it's hard to go across that journey, right? So you need to determine in the very beginning what exactly it is that you want to know. And that way you can really set yourself up for success, with your goals, with your content, etc. Typically, we recommend that you do this as a team, across departments, cross-departmentally decide what exactly does everyone is going to learn and get from this process, and make sure that there is buy-in.

Step 2 - Determine Your Hypothesis

The second step after you've determined your hypothesis or objective is to set those KPIs. KPI means key performance indicator. Essentially, what is going to make this successful for you? For some people, that might mean we want to reach 50% of a new audience, or that might mean we want to have a certain number of reach, meaning the number of people who actually see your ad or your content, or maybe it's a budgetary KPI. Whatever that looks like, you need to set what success is for this AB testing campaign, so that when you get to the end, you know if you did a good job or not.

Step 3 - Determine Your Variables

The third step is to decide on those variables. So if your objective is that you want to reach a new audience, maybe you could test anything you want because they're new eyeballs and you don't know what they think, but if you're trying to reach a current audience or, a market that you've already infiltrated, then maybe what you want to know is, is this messaging resonating? Do we want to be more playful or do we want to be more serious? Do people care about discounts or do they care about features? Decide what the variables are going to be and that way, when you actually create that content,and you know what you're pitting against one another in terms of content, you can actually, end the campaign knowing what you've tested. This is one of the steps where we see people not do so well, is that they have a clear objective, but then they create content and they come with all these great ideas, and they put all the ideas out there at once. And at the end of the campaign, they're not really sure what they've learned. So it's important to be very clear and put guardrails around what those variables, variables are gonna be. And maybe the variable is literallyjust this one has a purple background, this one has a red background, and this one has a yellow background. Those are three colors, which one do they care about more? And if you see at the end, more people clicked on the yellow one, then you know, yellow is the color to use and all of that kind of content moving forward. So set those variables, and then once you do that, you actually need to run the test.

Step 4 - Determine Your Test Type

Now, the test can be organic or paid. Typically, when we are doing AB testing campaigns, it's paid because what most clients want to know is about new customers. If you want to learn things from the current audience, the sort of Go to strategy there is to do some sort of survey or focus group or something like that. So you can totally AB test with either new or old audiences, or do organic or paid with either one. So determine if you have the budget to do paid, and if not, or if it makes more sense to do organic, try organic. In the olden days of marketing, one of the AB testing campaigns was to make oneminor change to like,, a mailer that would go out through snail mail and send 2000 to this audience, 2000 to that audience, 2000 to that audience. Or have the exact same thing and send it out to different audiences, and then have a different link on each one. That way you can actually track and see what variables are working. So that's an organic campaign. , it's not actually an advertisement, so there are multiple ways to go about it in both organic and paid. You just need to figure out what exactly it is that's going to meet your needs interms of KPIs, hypotheses, and variables.

Step 5 - Determine Your Conclusion

The fifth thing you need to do is develop your conclusions. So when you get to the end, you're gonna have all this data, right? But data is not the end. Data is just data. What you really want and what you can only move forward with is insights. So figure out what's deeper than the data. So maybe the data says that 50% of your audience actually prefer this over the 4 other options over there. But what does that really mean? You need to dig deeper. Sometimes there's like a false positive outcome of things. Maybe you're like, oh, people clearly liked purple more than yellow, but when you dig deeper, you realize thatall of the people who clicked on purple did so after 110 p.m. And so maybe it's more of a time-based thing than a color-based thing, which brings us to our sixth and final step, which is to iterate, test, retest, repeat.

Step 6 - Iterate, Test, Retest, Repeat

Take the insights that you've learned from your campaign and then develop a new campaign built off of that. With each subsequent campaign built on the insights of the previous AB test, you're only gonna get better and better and learn more and more. 

So get out there, take these 6 steps with you, and come up with a new AB testing campaign to determine what you'd like to learn and how you can elevate your own brand. And if you need support, you know who to reach out to.