Split testing is invaluable to website usability and user experience. Most people who run A/B tests understand Conversion Rate Optimization(CRO) which is a staple of A/B splits.
A/B tests are mostly always about revenue or a performance metric. Increasing the user’s behavior to perform some kind of action(sign up, purchase, opt-in). To get started all you need is a great tool and these testing tools are the best I’ve found for anyone new to A/B testing.
Visual Website Optimizer
The VWO tool has been around for years and it’s one of the more popular testing tools for beginners. Pricing is incredibly simple and it’s extremely straightforward too.
Each account comes with the default A/B testing tools so you can study which layouts are performing best. They also have extra stuff like heatmaps and click tracking along with personalized targeting(such as adding the user’s name).
I’m surprised how many features come with the VWO plan because it’s not that expensive. And if you’re only A/B testing one page you can typically run it for a month, then cancel the account once you’re done.
The lowest plan confines your tests to 10,000 visits which is fairly small. But it’s great for a lone test because generally 5k to each variation should be enough to reach statistical significance.
Larger brands and corporate environments prefer Optimizely because their dashboard is larger and they typically offer more customization.
The only downside is that Optimizely’s pricing is not displayed openly on the site. This makes it hard to get small-time bloggers and site owners to join and give it a shot.
But the upside is this does have a free trial period so you could join, run your tests., then quit and never pay a dime. It works for both web and mobile apps so you have the freedom to run tests and target audiences on all platforms.
Optimizely even has recommendation engines that you can add into your sites for product ideas. Definitely a good choice for ecommerce owners but it’s worth contacting them first to get an idea on pricing.
Every webmaster should run and understand Google Analytics. It’s the best free option for tracking user behaviors, page visits, time on page, bounce rates, pretty much everything.
But did you know Google Analytics also works as an A/B testing platform? It’s a platform in the dashboard called Content Experiments and yes this is totally free.
I recommend reading this guide if you want a full overview. It’s not as simple as other A/B tools because you need to do a lot of the setup work yourself.
Not to mention you’ll also need to create the proper changes on your website for the B variation of the page(assuming “A” variation is the control). This means canonical tags and noindexing for the duplicate page.
But the reason this is so cool is you can target any metric and track it through analytics. Conversions, bounce rates, time on page, all of these have strict numeric values and you can see which variation performs best right in your Google dashboard.
I’m a huge fan of free tools and there’s no doubt that Google offers some of the best for webmasters.
Highly recommend giving this a shot if you’ve never run A/B tests before. It’s not the easiest setup process but there are tons of guides online to help you get started.
- Your Step-by-Step Guide to A/B Testing with Google Analytics
- Set Up Your First A/B Test for Free Using Google Analytics Experiments
- How to Do A/B Split Testing in WordPress using Google Analytics
The nice thing about Unbounce is that you can dive in regardless of your background and start testing right away. They let you create landing pages right in the program so there’s no limitation to having your website online.
Their plans offer competitive pricing and for an all-in-one platform it is pretty great.
Another nice feature is that they don’t limit your tests. You can run unlimited split tests and study the data in real time with your Unbounce dashboard.
Plus they have conversion tools you can test like leadgen forms and modal window opt-in boxes.
This is probably the best A/B tool for marketers who launch a lot of landers and want to split test the hell out of ‘em. But this might not be as useful for someone who already has a site online and just needs raw data.
Statistical Significance Calculators
I mentioned statistical significance earlier and it is an important concept. This is the idea of running enough traffic and seeing enough conversions to statistically trust that your numbers are accurate.
Not all A/B tests reach statistical significance. It’s a matter of total visits and total conversions compared against both versions using some complex statistics.
Thankfully you don’t need to know anything about the math to study your data. All you need is a great stat sig calculator and I’ve listed a few of my favorites here.
The stat sig calculator on GetDataDriven is one of the easiest to use. It has a very clean interface and it’s super straightforward.
You have the A & B rows where you enter the visits for both variations, then the conversions for both variations. You’ll see the conversion rate in one column and next to that you’ll get some info explaining if the data is significant or not.
Typically you want at least 90% stat sig confidence but some marketers make decisions with slightly less.
The A/B test calculator on Peak Conversion is another one of my favorites. This one includes a graph with probabilities mapped visually—not super useful if you don’t understand it.
The really useful thing here is the column clearly stating the chance of that variation being the best. You still enter total # of visitors and total conversions, then hit “calculate” and check that column. If you see one variation with 95% or higher then it’s safe to assume that is the best choice.
And if you like this tool you might also wanna check VWO’s testing calculator which estimates how many days you’d need to run a test before you get accurate data.
All of these tools are worth bookmarking if you’re serious about doing A/B split testing. It’s not an easy process but in the long run it can drastically increase your revenue, even sometimes with just a few design changes.