A/B Testing: Why You Should Start Experimenting Today!

So you’ve built your form, it’s collecting leads, everything is peachy. Now let me sow the seeds of doubt…

How do you know your form is living up to its potential, that it’s generating as much traffic through to you as possible? Think tweaking this or that might be an improvement but you’re not sure how to tell? Have a clever idea but a boss or colleague insists it won’t work?

Stick with us for the next few minutes and you’ll learn how to prove it!

What is A/B Testing and Why is it Important?

We’ve spoken extensively on a variety of ways to boost form conversion, and all of those points are tried and true research based strategies. When it comes right down to it though, it’s a big world with a diverse population, and any general strategy is at best only generally applicable.

How can you move beyond generalities and discover specifically what’s best for you and your audience? You have to experiment!

A/B testing lets you take an idea swimming around in your head and test it live right alongside your current setup to verify which one actually works better. You divide your traffic between the original form and a modified version, and see which one gets better results. It’s that easy. It removes any element of guesswork and gives you real data to move from “I think this would work better” to “I have proof this works better”.

What Tools Are There for Me to Test With?

You have a variety of options. Optimizely is what we at WP Ninjas use for our sites, but you might find another one that you prefer. There are a range of selections out there to fit your needs, typically based on the volume of traffic your website receives.

optly-228x152Free options will probably suit you fine if you serve less than 5-10,000 monthly visitors, and if you serve that range or more there are graded options ranging from small monthly fees for small businesses to expansive (and expensive) enterprise coverage.

For the majority of these the only initial setup required on your end is a simple install that may require manually inserting an API or line of code into your website, which they will walk you through. Let’s look at some options.

Optimizely. Free or Premium. Again, this is our preference here at WP Ninjas. There is a fully functional and easy to use A/B testing component here with multiple bells and whistles completely free, and their Enterprise Plan for deep data analysis. Great for just getting your feet wet or going all in on some serious experimentation. They offer a free WordPress plugin that integrates headline testing directly to your dashboard.
Google Analytics. Free. Google Analytics is the standard for data collection and analysis on the web. Rolled into the Analytics package is Content Experiments, from which you can do a variety of testing. The catch for the price: it’s simply not the easiest or most intuitive tool for those not already familiar with Google Analytics. If you’ve got the basics of the software down already, or if you’re not afraid of a learning curve, go for it. It you’re looking for something easy to get started with out of the box, this might not be the tool for you.
unbounce-logo-plus-header-templatesNelio A/B Testing. Starts at $29 billed monthly. This is a solution fully native to WordPress that, for a fee, takes all the technical work out of your hands. No coding required, fully automated, and bonus features like heatmapping make this an option worth exploring.
Unbounce. Starts at $49/month. Unbounce does some really cool stuff with data and form conversion, and offers a range of related services in addition to their testing products. They are working on a WordPress integration, with a beta plugin available in the repository.
Visual Website Optimizer. Starts at $59 billed monthly. Startup, business, and custom enterprise plans with a wide range of options to fit your needs large and small. Plugin available for WordPress integration.

I’m Ready to Go! What’s the Best Way to Start Testing?

Ever wanted to be a scientist when you were growing up? Well you are one now. You’ll be thinking like one, anyway. Designing a great experiment can be hard, and getting it right is critical. Here’s how you do it.

Identify what you want to test. Pick one thing and only one thing! You have to work with one variable at a time, otherwise when you’re looking at your data, how do you know which change you made affected the results? If you are wondering if your red submit button might do better as a green send button, that’s two tests because you’re changing two variables (red to green and submit to send).

Use a control. If you’re A/B testing in the first place, then you’ve got this covered. Your control is the form as it is now, without changes. Call that A. The form with the single change you want to test is B. A/B testing: not just a clever name. Run your test!

Collect as much data as reasonably possible. The more the better within reason. The bigger your sample size, the more accurate the conclusions you can draw from it. If you ask 20 people how they’re voting for president, that doesn’t tell you much about the election. If you ask 20,000 people you’ve got a better indication, and 200,000 is much better. See the trend?

pexels-photoHow long and how much traffic flow you can test is going to be highly situational. It’s hard to set a line and say numbers over this are significant and numbers below it aren’t; it’s very relative to your site, your goals, and your traffic. A 10% change with a sample size of 200 would be pretty big if you’re serving 400 visitors a month. The same numbers would tell you next to nothing if you serve 4,000 people a month. Bear in mind that making decisions based on too little data can be worse than just plain guessing, but you can’t test forever either.

Analyze your results. Analyze thoroughly and consider results carefully. Don’t assume that just because Form B outperformed Form A in conversions, that B is better. Look at the data from as many angles as possible. If quality of useful leads dropped with Form B even though total leads rose, switching to Form B would be a bad move. Take your time and be methodical with your data. Examine it from all angles.

peeking-ninja-300x138A/B testing is the difference between thinking something might work better and knowing it does or doesn’t. Being informed and making data based decisions are at the core of good decision making skills. You might have great business sense, but judgment and instinct will only take you so far. Play smart and let evidence drive your decisions. A/B test your ideas before you make another change!