A-B Testing Part One: Optimizing User Experience (UX) for Websites - LaunchUX

Captain's Blog

A-B Testing Part One: Optimizing User Experience (UX) for Websites

In his book “Why We Buy: The Science of Shopping,” Paco Underhill revealed (with meticulous research) how people actually behave in retail stores.

He didn’t do it by asking them to fill out surveys about the retail shopping experience.

That would have only told him what people think they want – and what they want the researchers to think they want.

Underhill and his team used clipboards and stop watches to secretly record customer’s interactions with products and displays. (The researchers were often mere feet away from customers but were rarely noticed – which, in itself, is an interesting result).

Did the customer see the display table? Did they stop? Did they pick up the product? Did they put it back? Did they purchase it? How many seconds did this interaction take? If we rearrange the table, do we get different results? Do our efforts to increase product visibility in one area negatively affect another? But selling more of X are we sacrificing sales of Y and is that a reasonable trade-off?

An interesting aside: When the researchers surveyed CEOs and managers, the results revealed that they had little-to-no idea what customers were actually doing in their stores. As with customer surveys, the managers were only guessing.

And they were almost always wrong.

The customers almost never followed the rational, logic-based wayfinding signs and customer pathways in the store. It wasn’t clear why they were doing what they were doing, but there were observable patterns.

They couldn’t control customer behavior, but they could still monitor it and use it to their advantage:

Through cycles of observation and enhancement, the customer experience was improved so that they found what they needed without frustration – and the managers noticed a bump in conversions.

The book was full of practical findings regarding how brick-and-mortar stores could instantly improve their performance, but the most enduring part of the book is Underhill’s comprehensive and analytical approach to customer behavior.

Around the time “Why We Buy” was published, there was a trend in web design to make websites perform like brick-and-mortar stores. It made sense at the time because retail stores were a tried and true frame of reference. By 2018, that trend has started to reverse itself, and stores like Amazon and Warby Parker are incorporating online selling techniques in their brick-and-mortar locations.

The constant cycle of improvement – tweak, measure, record, tweak, measure, record – remains as relevant as ever before.

If you ask managers and customers what they want in a store or website, the results will inherently be faulty.

Visitors can’t tell you how to improve your store or website.

But they can show you.

A lot can be learned about your customers’ current user experience as well as how to improve it by simply watching what they do and measuring it against what you want them to do.

Today, we call it A-B Testing, and digital analytics have made it easier than ever to identify customer behavior and design your website, digital advertising and social media accordingly.

Measuring customer interactions online doesn’t require clipboards and stopwatches. In fact, it’s so easy to do that if you’re online, you’re almost certainly already being tested. Companies like Google and Facebook are famous for running multiple A-B Tests concurrently. They are very interested in understanding how and why you use their services.

In its simplest form, A-B Testing exposes visitors to your website to one version for a predetermined amount of time. Then, they are shown a variation for a predetermined amount of time. The results are compiled and analyzed. Then, based on your company’s specific objectives, the site can be made more effective.

Provide two different user experiences, and visitors will show you which one they prefer.

The differences between the A version and the B version of a site can be as vast as two completely different designs or simply moving the “Sign Up” link to a more prominent area of the site depending on which metrics you are seeking to improve.

And if you have a business, A-B Testing is so easy to perform, that it really doesn’t make sense to not do it.

Like managers of brick-and-mortar stores, business owners who are developing their online presence without metrics can only guess at how their customers are actually using their website.

A-B Testing online is an easy, cost-effective way to ensure that your website is actually serving your customers’ needs while making sure that it is achieving your goals – whether that means sales conversions, e-mail sign-ups, social media engagement or something else entirely.

Here are seven things to keep in mind as you move through the A-B Testing process to ensure that you are maximizing your web building dollars.

1. What is the objective of your website?

Yes, it’s possible to simply observe customer behavior on your website, draw conclusions and start tweaking. However, it’s much more effective to start with a specific goal in mind. The more specific the better. “Generating awareness” is a noble goal, but in A-B Testing it is way too vague and way too hard to measure. Specific goals generally include increasing sign-ups or sales conversions.

2. How successful is your website now?

Testing the “A” version of your site will provide you with the baseline results you need to accurately gauge the impact of the changes that you implement in version “B.” It’s very difficult to make improvements when you aren’t sure where you stand today.

3. What metrics are important to your website?

One you have defined the core objective of the site and you understand how the site is performing at present, you can single out the metrics that are important to bringing your site in line with your vision. Website analytics allow us to study everything from click-throughs, time spent per page and bounce rate.

4. What are you willing to change about your website?

Your current website might be lovingly handcrafted work of art, but if it’s not getting the results you want, you might have to make some very difficult decisions. It’s natural to resist making changes to something so personal, but the data doesn’t lie. Obviously, aesthetic design is important, but you ignore analytics at your own peril.

5. What are you willing to temporarily risk in the name of overall improvement?

Sometimes the “A” version of a website is just plain bad. And sometimes the “B” version is even worse. It happens. However, the information gained from the experiment is still valuable. It’s generally not difficult to revert back to the original version, and now you have a better understanding of why things are the way they are.

6. When are you done A-B testing?

This question ties back to the objectives of your website and the analytics that you’re measuring. As we mentioned above, companies like Google and Facebook never stop testing, and they are often running a variety of tests simultaneously, but this isn’t feasible for many smaller businesses. Starting with clear objectives will let you know when your website is moving in the right direction, and periodic testing of that specific variable will ensure that you are hitting your targets. User behavior is constantly changing, and what worked in the 90s, for example, doesn’t work today. A-B Testing can help you stay up-to-date.

7. Does the perfect website even exist?

All of this talk about improvement and testing and re-testing – sheesh, will your website ever be perfect? Well, if you set clear objectives and you’re hitting your targets with a well-designed, aesthetically pleasing website, that’s about as close to perfect as you can get. But perfection is fleeting. What is perfect today might not work tomorrow. Regular A-B Testing will make sure that your site is keeping up with industry changes.

A good foundation starts with investing in a secure, fast-loading website that delivers quality, regularly updated content whether it is accessed on a computer, a smartphone or a tablet while maintaining an active social media presence.

This blog entry has largely been theoretical so far, but A-B Testing is something that LaunchUX routinely does to our own sites as well as those of our clients.

It might be hard to believe, but we still encounter small businesses that simply display information on static websites as if they are digital business cards or billboards.

The internet was built to connect people and increase engagement. A-B testing ensures that we’re doing it effectively and that we are providing an environment that our customers will want to return to (notice that we’re once again using the language of brick-and-mortar stores).

One example of our own A-B Testing was moving our “online chat” feature, which allows prospective customers to message us directly, to various locations on the page. We discovered where it was most utilized. Today, the vast majority of our customers reach out to us using this chat feature, and we will continue monitoring it to make sure that it is fulfilling our objectives and hitting our targets.

For another site, the A-B Testing consisted of running the site with a certain combination of layouts and color palettes for a few weeks to see if sign-ups for a software-as-a-service company increased. We were able to then use this information (whether they increased, decreased or stayed the same) to determine the tweaks that we implemented for the next testing period.

In yet another case, we simply moved a client’s contact form higher up on the page to a more prominent position, and then we measured and compared the number of sign-ups against the original data to see if “B” performed better than “A.”

For most of our clients, we do a form of A-B Testing to encourage people to stay on the site longer (cut the bounce rate), make a purchase or request additional information.

The entire internet has been built on incremental improvements in user experience from day one. As technology improves, your digital presence needs to be ready to incorporate those developments in a way that provides the best user experience for your visitors.

How do we make sure that we’re creating a space that meets our customers’ needs?

We don’t just ask them.

We watch what they’re actually doing.

We are always tweaking and looking at the data to make sure that the A-B Testing is moving our clients closer to their goals while enhancing the overall user experience on their websites.

Leave a Reply

Your email address will not be published. Required fields are marked *