The Importance of A/B Testing for Web Developers
This article explains the benefits of A/B testing for web developers. After all, the point of a good company website is to bring in more money. So, the more we can do to measure sites’ conversion rates and work to improve them, the easier it is to prove that our website is bringing in ROI.
Join the DZone community and get the full member experience.Join For Free
As web developers, when we get hired to make a website, we might forget the driving assumption behind that expense: companies believe a better website will help them make more money. Why else would they be willing to invest in development, design, and ongoing maintenance other than a belief in the potential of a return on their investment?
Logically, the more we do to measure sites’ conversion rates and work to improve them, the easier it is to prove that return on investment. If we can prove the value of our work, and show benefit that can be expected, the easier it is to justify the expense. The more ongoing systems we put in place to improve conversions over time, the easier it is to sell ongoing services.
In this article, I want to show you some ways that you, as a web developer or development agency, can implement ongoing A/B testing and conversion rate reporting as well as optimization in your projects. These are simple steps you can use to quantify the value of your work to a client in hard numbers.
Start With Analytics Reporting
Google Analytics is pretty standard, but their dashboard is not the easiest thing to digest. Also, just adding it to your site alone doesn't measure conversions. It merely gives you more information.
If your client's site is an e-commerce site, you need to send analytics events for sales. ShopPlugins makes plugins for reporting WooCommerce and Easy Digital Downloads sales to Google Analytics.
If you are developing a lead generation site, make sure to have your lead generation form report a conversion event. FormWorks, a plugin my company develops, can send custom Google Analytic events on form submission from any major WordPress form builder.
If you're building the top of a sales funnel to send users to a third-party e-commerce platform, triggering a custom event on clicks of any of those links is a must.
But remember that Google's analytics dashboard is information overload to most end users. Consider a tool like Analytify or Monster Insights to simplify it. Also, typing up a weekly or monthly summary of the two or three most important pieces of analytics data and sending it to your client can go a long way to illustrating your value.
Reporting Is Not Enough
Reporting changes in conversion rate over time is a good practice, but for your sake, you want it to be constantly rising. This is where conversion rate optimization is just as important as performance optimization.
When you create a website, you're calling on all of your past experience to make the best judgment about what is best for a site. But, in the end, those decisions are just educated guesses.
A/B testing lets you put your best guesses to the test and find the best solutions, in the right combination, for the current audience. On today's internet, we are surrounded by tests. Pretty much every advertisement we see and every recommended post on our feeds is part of an ongoing testing process. On websites, page layout, call to action messages, colors, menu order, and more are often important to find the winning combinations.
Of course, conversion rates will rise and fall. When they are down, having the right data to explain why and develop a plan for how you intend to bring them back up is very important.
There are a lot of ways to do A/B testing on a WordPress site. Choosing the best option for your needs is important, but failing to choose any likely means the site you're developing does not deliver as much value to your client as it could.
It's easy to think "that's just for marketing folks," but content rate optimization, when done right, should make sense to you as a developer. It is a systematic approach to optimizing a repetitive process over time.
Don't Forget This Is a Science
Too often developers jump into testing with the idea that they will build two or more versions of a page, switch them out at random and then choose the one that converts best. What I always ask them is: how they will know what caused the change in conversion rate?
When I started out as a web developer and was trying to fix a bug, I would change a bunch of things and see if they worked. Sometimes it did and sometimes it didn't. I never really knew why. Now I work by methodically changing one thing at a time so I can understand what is actually causing the bug, what fixes it and what has no effect.
A strategic approach to a results-driven multi-variant testing solution is very similar. Instead of making multiple layouts, make a list of all of the possible changes you can make to a layout and try them one at a time. This may sound tedious—that is where automation comes in. If you're new to this, you might doubt the effect of small changes, but they can be big. Changing the color of a button, or a few words in a call to action can have a huge impact, for better or worse. As for little changes that have no effect, the identification of those is valuable information.
For a native WordPress solution, there is Ingot, which I am a developer of, and there is also A/B testing built into the Divi Theme and Builder plugin. A developer could also get creative with different WordPress hooks to make the changes, and then use Google Analytics or similar to track conversions.
Deciding on the best solution comes down to your budget, why you want to test and what kind of traffic you are expecting. Some services are designed with high traffic sites in mind and you may not be able to give them enough data to get any meaningful results. In the last few years, new algorithms that support multi-variant testing and smaller sample sizes, such as the "multi-armed bandit" strategy have gained prominence in automated conversion rate optimization for delivering better results, faster.
It's easy for think that once you have run a test and found a clear winner, that is always the best option. While past performance is a good predictor of future results, not all results are timeless.
You need an ongoing strategy to make sure that your past test data is still valid. Not only do tastes change over time, but as you make more improvements to a site, you do always risk undoing some of the advantages you've gained.
That's OK if you have ongoing testing and reporting to continually monitor and improve conversions and other performance indicators for a site. So, keep testing.
Test All the Things
I hope in this article you have seen the value of ongoing A/B testing for your client projects. Implementing these proven strategies will help you deliver better value to clients and nurture long-term relationships that will be mutually beneficial for you and those you work for. Yes, all of this is more billable time you have to sell your client on, but by its very nature, it is easy to show why it can pay for itself.
Most importantly, it shows that you care about the long-term success of a project, which is invaluable to a client. By using the right set of analytics recording and reporting tools as well as testing tools, you can ensure that the websites you deliver only get better over time.
Published at DZone with permission of Josh Pollock, DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.