Why You Should Be Doing A/B Testing | Ep. #21

You know that feeling when the strategy is done, the brief is written, everyone's aligned, and you realize someone still has to sit down and actually create all the content? That someone is you, and it's due tomorrow. Breeze Assistant can help. It works right inside HubSpot, drafting campaign copy, blog posts, emails, all in your brand voice, all grounded in your actual customer data. So you don't just create content. You create content that converts. Check out HubSpot.com, the agentic customer platform for growing businesses. Get ready for your daily dose of marketing strategies and tactics from entrepreneurs with the guile and experience to help you find success in any marketing capacity. You're listening to marketing school with your instructors, Neil Patel, and Eric Sue. Alright guys, before we start, we got a special message from our sponsor. If you want to rank higher on Google, you got to look at your paid speed time. The faster your website loads, the better off you are. With Google's Corvital update, that makes it super, super important to optimize your site for low time. And one easy way to do it is use the host that Eric and I use, DreamHouse. So just go to DreamHouse or Google it, find it, check it out, and it's a great way to improve your low time. Okay, everyone, get ready. It's time for another episode of marketing school, I'm Eric Sue, and I'm Neil Patel. And today, we're going to talk about how to do AB testing. So, Neil, what the heck is AB testing and why should we do it? AB testing is, think of it this way. You go into a grocery store and someone offers you jam to buy. Perhaps the people they may offer a flavor of strawberry, the other half they may offer a flavor of marmalade, and if more people buy with strawberry and in quantity, right, you're talking about thousands of people or a good sample size, then they know that, hey, strawberry is a flavor that most people prefer. When we pitch our jam's product, we should first show strawberry. That's kind of like AB testing, right? On the web, you're trying to figure out, people are coming to your website. Let's say your homepage. Is this version of your homepage better or another version of your homepage? Is adding a video to your homepage, generate more leads or more sales? Or does just having texts with no images cause more sales or more conversions? So with AB testing, it's trying to figure out what copy, images, messaging, videos, what elements can you change or adjust to get the maximum amount of sales? I love it. And I think to kick things off with AB testing, there's a few points to be very cognizant of when you're doing AB testing. So the first thing is, I generally don't recommend any type of AB testing. You're really going to try to AB test when you have a certain volume of traffic and you have certain things rolling in the business already. So things are kicking along already. Let's just say, let's just use a baseline of let's just say 10,000 visits a month. That's a good point to start doing some AB testing on different elements on the site. But that's not to just say, hey, go read a blog post on some conversion rate, blog out there and say, oh, let's just go ahead and change some colors on our buttons. And you know, call it a day. That's not how it how it works. What you might want to do initially is you might use tools like what Neil's talking about or alluded to in other episodes. Qualifier would be a good example. You might use a tool like hot jar, just to see how people are behaving on the site. You can use a tool like crazy egg that will show heat maps on the website. And then also you can run surveys too. There's a lot of great applications out there for surveys, SurveyMonkey is one of them. And there's also one called Yes Insights, which basically is a one click survey where you can email people. And just to give you an example, using a one click survey tool, I got 90 responses. And you know, using SurveyMonkey, I got 10 responses in the past. So that's a nine next increase that that's something to take in mind. When you're running these types of surveys, you're going to get ideas for running a specific test. So you want to look at the data. You want to look at what people are saying. Then you can start to hypothesize different experiments to be running. Neil. When you're running AB testing to maximize your sales, you need volume too. So in general, there's tools like optimizely, crazy egg, visual website optimizer, Google content experiments that help you run the AB test. Things like crazy egg or optimizer, VWL, they have a wizzy wave editor so you can adjust elements within your site, copy without needing a developer. And when you're doing these tests, you need volume. So the tool will tell you is one version feeding out another version. If you're only getting under 1000 sales or conversions on your website in a given month, you shouldn't have more than two versions. You're controlled being your main version of your current site or your current web page. And then be variation being one variation, right? And that's what you want to test at first. The tool will then tell you if something is a winner or a loser. And typically when you're running tests, you need to run them in links. And what I mean by that is if you start on a Monday, you can finish on a Sunday, but you don't want to do a test Monday through Wednesday. You want to make sure you're going in week period of time. A lot of times I like doing it for two weeks, three weeks, four weeks straight. Of course, once it says statistical significance, if it says it too early on like after three days, I will still run it for at least a week. But I want to make sure within my tests that I have each day of the week in there reason being is people interact with the website on Sundays differently than they do on Mondays. And a lot of the people who are on the website on Sundays are different demographic than the people who are on the site on Mondays. For example, Sunday is a work day in Israel, right? Their version of a Monday is actually Sunday. So their work days are from Sunday through Thursday, Friday and Saturday are off. So different demographics view different sites on different times. If you want to get a full sample, make sure you're running tests for at least a week. And don't end a test until the tool tells you it's statistically significant. I love that. And deal, we might defer on this here, you know, I remember running a budget test at a startup I was at before and the, you know, in general, we try to always get to 99% statistical significance. Now, the issue with that is that can take a long time to get there. And it might not be worth waiting that long. So how, you know, when this is statistical significance for you, what percent are you usually at before you say you're satisfied? It usually depends, but we actually do try to go for 99%. If it's been running for a long time, like months, we usually don't have tests that are running that long. And it's like at 97% with thousands of conversions on each variation, the 97% is good enough. Like we'll take it. Usually the tool will call it a winner by the end anyways. But the main thing that we're looking for is how likely is it going to stick when you only have 30 conversions on each variation, even if it's 99% statistically significant, like one version had 30 conversions to other one only had five conversions. Well, it's just starting out, it could only be a day's worth of data. You shouldn't stop a test then you should let it keep running. Right. And there's another concept of AB testing and a lot of, it's called multivariate testing where you're running the A, B, C, D version. So you might have four different versions of something. And in general, a lot of people don't talk about multivariate testing. They keep it to A, B testing. Neil, is there a reason for that? Yeah, a lot of people don't do multivariate testing because they don't necessarily understand how it works. Well, I've also found a ran a ton of tests. Multivariate testing, as Eric mentioned, is like when you're adjusting different elements on a page, so you can have one main version, your control, your current site. And then you can also test the headline, the color of the button, the text of the button, and some images. And you can be running the tests all at once. Well, when the tool tells you, "Hey, this headline is a winner by 10%, and this color is a winner by 30%, and this image is a winner by another 10%, that's 50% in total." What we found is then when you combine all of those, you usually don't see it increased by 50%, for that reason we tend to do AB testing. It just makes it more simpler. And when you're running the tests, in most cases, as we were talking about it on this podcast, most people won't have a ton of conversions. You ideally want over 1,000 conversions per month before you start running AB tests. But even if you have a small amount, you can still run tests. And the way we like to run them is you go look in your Google Analytics and you go see which page are people dropping off in your funnel the most. So if the way people buy is going from your home page to your pricing page to then the checkout page and then the thank you page, if a 100 people go or let's say if a thousand people go to your homepage and out of those thousand people, 900 go to your pricing page. And then out of the 900 on your pricing page, if only 50 people go to your checkout page, and this is before they're into your credit card, they're just clicking through one checkout. That means there's bigger issues in your pricing page versus your homepage because a drop off is so much more steeper that you should first try to optimize that pricing page. And when you're adjusting that pricing page, you don't want to just start running random tests. As Eric mentioned, there's tools like QualaRoo out there. You go take a tool like QualaRoo, you go get feedback on the pricing page of why people have issues with it, the objections they're having. And then what you're doing is you're taking all the objections the main ones and you're figuring out which ones would be the easiest ones to fix, you then fix them on a B variation, right? So you keep the original and then you create a B variation and then you run a test. There's a lot of these paid tools out there, but I want to also point out that Google Analytics is a great place to start because in Google Analytics, you can see the funnel drop off If you have goal set up and then also you can see the behavior flow of where people are landing on and what pages They're going to next and how people are converting in general So if you're building an e-commerce brand you should check out DTC pod hosted by Ramon Barrios and Blaine Bulless on the HubSpot podcast network They speak with founders marketers creators agencies and platform experts about what it actually takes to grow a direct to consumer business From paid ads and influencer marketing to conversion email brand building and consumer trends I particularly enjoyed their conversations around scaling a brand without losing what made customers care in the first place Listen to DTC pod wherever you get your podcasts do recommend taking a look at Google analytics And I think maybe Neil we can close it off by talking about some of our favorite conversion blogs So I'll go ahead and start with conversion rate experts There's a hyphen between each of the words so conversion hyphen rate hyphen experts There's a lot of great case studies out there They were kind of you know the forefront at the forefront of conversion rate optimization and they were the first conversion rate blog that I found Conversion excels another amazing blog all they do is just talk about conversion tests and data Even offline versions that people are running. They have great case studies that you can learn from as well Love it great. Well, that's it for this episode We will see you in tomorrow's episode and keep on learning. Bye This session of marketing school has come to a close be sure to subscribe for more daily marketing strategies and tactics to help you find the success You've always dreamed of and don't forget to write in review so we can continue to bring you the best daily content possible We'll see you in class tomorrow right here on marketing school