6 Ways To Never Run Out of Conversion Rate Optimization Ideas

When your A/B testing routine starts to feel stale, it's time to try something new. Here are six proven ways to get great results.

Founder, Expandery

When companies first start A/B testing their website, email, and other channels, it usually doesn’t take long for them to see big results. Just changing the placement, size, and color of their CTA buttons, for example, can yield in huge improvements to their conversion rates. There’s a lot of excitement about conversion rate optimization in those first few months, but over time the momentum and enthusiasm driving all that testing starts to slow down.

One of the most common reasons that A/B testing programs lose steam is that the people who create the tests start to run out of ideas. This can happen quickly, as marketers and analysts tend to start with the lowest-hanging fruit. With the most obvious test ideas out of the way, it’s a little more challenging to know what kinds of things to test next.

People also have their own preferences and biases when it comes to testing priorities. A particular analyst or marketer might be a highly visual person, for instance, and tend to only think about tests that involve visual elements like color variants or layout. Another person might focus on the content of the website or email, A/B testing alternate site copy, headlines, or offers. Other people think in structural terms and focus on elements like which form requirements generate the highest number of conversions.

None of these are bad ideas for A/B tests, but there are only so many variants of each kind that are worth exploring. Eventually, people who stick to these same testing ideas will get to the point where they start to see diminishing returns from their tests. The gains in conversions and other metrics become smaller and smaller, making it seem like the A/B tests aren’t generating worthwhile results. To generate new and exciting results again, they need to try new kinds of tests.

When your A/B testing routine starts to feel stale, it’s time to try something new. Here are six proven ways to get great results.

1. Create an Idea Generation Matrix

One of the most helpful starting points for reviving your A/B testing strategy is to create a list of all the different broad categories that you have the ability to test. Then, combine that list with all the kinds of tests you can run. This testing matrix allows you to quickly remind yourself of what elements can be tested, and what kinds of tests you have available to test them.

These categories might include:  

  • Design Elements
  • Button Size and Color
  • Offers
  • Site Copy
  • Social Proof
  • Page Functionality
  • Suggestion Algorithms

This allows you to create tests easily, and to quickly identify testing categories that you might have overlooked. 

2. Get as many people involved in the process as possible

If your testing team is small — often, it’s just one or two people — coming up with new ideas for A/B tests can quickly become a chore. At the same time, there may be other people in marketing, sales, and management, who have tons of ideas they would love to try.

Get as many of these people engaged in the process as possible. This can be an informal conversation, such as a group email or dedicated Slack channel, or a weekly or monthly meeting. This team can then discuss new testing ideas, and examine the results of previous tests to consider future iterations and improvements.

3. Prioritize your A/B ideas

Some A/B test ideas are better than others. Judging which tests should take priority over others, however, can be a serious challenge when your team includes people from different departments, many of whom will have their own priorities (or outrank each other). One solution to this problem comes from one of my former clients, a large American fashion brand…

Once a month, they would gather their digital marketing team together and hand out sticky pads to everyone. They would then spend five or ten minutes brainstorming ideas for tests, one idea per sticky note. This allowed them to quickly create ideas for as many different kinds of tests as possible.

After that five or ten minutes had passed, everyone would lay out each of their sticky pads in front of them. Each member of the team was then given a number of poker chips, and they would place those chips on the tests they most wanted to see done. One person might put all of their chips on a single test that they really wanted to see done, while others would spread their chips out over several tests they liked. This allowed them to quickly prioritize which tests they should do next. Not to mention, it was always a good time.

4. Remember the five basic questions

Every A/B test is an experiment, and that experiment should answer a specific question. In creating that experiment, however, you should always be able to answer the five basic questions about what you are hoping to learn. We all know these questions: What, Who, Where, When, and Why?

  • What specific concept or preference is being tested?
  • Who is the audience? Are all visitors being tested, or just some select segment? You should be using different tests for different audiences.
  • Where on the site are you testing, and under what circumstances? Is it the homepage, or are you limiting it to targeted landing pages? Why is this test happening in this specific location?
  • When will the test run? Some tests make more sense at specific times. Some tests may generate the same results no matter when they happen, while others will only provide meaningful insights if they happen during peak site hours, on the weekend, or during the holidays.
  • Why are you running the test? What is the hypothesis of the test? What will you learn from the test? What results are you hoping for? How will you incorporate what you’ve learned into the site, and into the next round of tests?

There’s more to these tests than simply seeing what happens when you change the size of headline or replace the default product image. Your real goal should be to use the data from your tests to create an overarching theory about how these elements come together to generate the results you are looking for.

5. Don’t forget to A/B test your emails

When compared to websites, emails only have a handful of testable elements. As a result, many people have come to believe that A/B testing emails isn’t a good use of resources. That’s wrong. Email is actually one of the fastest and easiest ways to see real results from A/B testing. It’s a perfect sandbox to test out segmentation strategies, and even seemingly crazy ideas.

For instance, I once had a client who wanted to see what would happen if they sent out an email with a blank subject line. They already had a large mailing list, but their open rates had been stagnant for a few months. Why not try something radical?

The control email got typical results, but the no-subject email way outperformed it. It’s a strange idea, but it’s not hard to see why it worked. We already know that emails tend to work better if they are informal and prompt curiosity out of the recipient. A company sending out a blank email is just unusual enough that people decided to open it, just to see what it was.

Obviously, you couldn’t send out emails with no subjects all the time. Even if it worked for a while, eventually people would get used to seeing blank subject lines from your company, and the higher open rates would quickly taper off. But maybe sending one blank-subject email from time to time could be a workable strategy.

6: Create an A/B email testing arsenal

While it’s certainly worth A/B testing other aspects of your emails, for example, the focus should often be on the subject lines. Once people open the emails, you have more options, but there’s no getting around the reality that most of the testing should be focused on that one crucial element, especially if you have a smaller mailing list. Unfortunately, there are only so many ways to write an effective subject line. This makes A/B testing of those subject lines an essential part of the process.

What you’re really doing with email A/B testing is building up an arsenal of ideas. If your open rates start to slip, you can use these ideas to shake the recipients out of their rut. It’s not about finding the “one true email subject line” that will always have the highest open rates. It doesn’t work that way. It might work for a few emails, but eventually, the recipients will get used to the new format and lose their curiosity. Your email open rates will then go back to where they were before.

Let’s say that you have an arsenal of three dozen techniques that have been proven to work through A/B testing. Some will be more effective than others, but they all have a statistically relevant effect on open rates. If you only send an email every other week, you have more than a year’s worth of ideas to try.

To help you with this, I’ve put together a simple (but hopefully effective) conversion rate optimization and a/b testing tracker that should help you categories your ideas, come up with new ones, and maintain a list of results for posterity:

A/B Testing Sheet

Click here, or the image above to access.

Wrapping Up

Remember, the purpose of A/B testing is to learn things quickly, and to gather data you can put to immediate use. Over time, your A/B tests should become laser focused and increasingly advanced. They should move beyond things like basic design elements, and move into more complex categories like site functionality. Before long, you’ll be able to A/B test everything, gathering crucial insights into product suggestion algorithms, pre-orders, user personalization, customer surveys, and segment-specific changes. This can even apply to your mobile site, such as the “one-click buy” versus “swipe to buy” tests that Amazon has been running on mobile users lately.

There’s no reason to let your A/B testing become stagnant. Anything and everything about your site, your email, your social media, and even your customers’ behaviors, can be tested. The more you test and learn, the better your results.

6 Ways To Never Run Out of Conversion Rate Optimization Ideas
5 (100%) 2 votes