For most businesses, the measure of a website’s succes it the number of conversions it delivers. By continually experimenting with web pages and measuring which changes increase conversions, these businesses can optimize results over the long term.
According to a study by Marketing Sherpa, overall conversion rates across a variety of web properties are frequently less than 5%. Other studies show conversion rates (for B2C landing pages specifically) to average around 10%. Wherever your web pages fall on this performance curve, the way to achieve better performance is to focus testing on the right things, and test in the right way.
What to focus testing on
Any testing regimen begins with a degree of speculation. You must form a hypothesis to test based at least to some degree on a speculative proposition. A few things you could focus your testing on include:
Your unique proposition. Anyone visiting your landing page should immediately recognize it offers them something valuable. Consider what your offering and how you can make the idea more easily understood. Each variation can be tested.
Images. Visuals work. Testing various visuals, and combinations of visuals, as well as their size and placement, can yield valuable insights.
Input forms. Test form labels, form length, the presence of required fields, the presence of in-field labels or “hints”, and more. Generally, it’s assumed shorter forms are easier to complete, so test simple forms versus those that are slightly more complex but which may provide more value qualifying information to your organization. Consider multi-part or multi-stage forms as well in your testing protocol.
Buttons and other calls to action. There are so many stylistic nuances to test with buttons. Color, size, font, verbiage, placement, mouseover and click actions – all can be tested against one another to arrive at the most effective combination.
Navigational elements. Sometimes, offering a wide variety of navigational options helps reduce bounce rates, since visitors have many options to explore next. The downside is when these options distract from the action you hope visitors will take. Testing can help determine if your navigation options are diverting customers away from converting.
Editorial copy. Test the impact of the length of your opening paragraph as well as the amount of copy on the page, the tone, the keywords, the personality it conveys and the overall value of the information it provides.
Additional testing opportunities await with testing page headers, sidebars, footers, titles, subheads, layouts, themes, and widgets. This makes A/B testing exciting but adds some complexity to the page-testing process.
How to test these elements
Testing a web page boils down to testing its performance against an alternative, nearly duplicate page that has only one difference – the element being tested. There are some things to remember to facilitate a good test, and some things to avoid a bad one.
Conduct A/A tests. First, to gain an understanding of the natural variances your audience my impart to any subsequent testing, conduct an A/A test. Create two identical pages, and divide traffic to these pages into two equal parts. It’s very likely you’ll see a difference in behavior between the two groups, even if they’re chosen randomly.
If these variances are off the charts, double-check audience definitions to ensure there aren’t factors introduced during participant selection that influence test outcomes. Look at traffic sources, devices, age, gender, and other factors that could be influencing outcomes. If one group is not like the other in some respect, redefine them until they’re as alike as they can be.
Divide testers into equal parts. Dividing your test group into equal parts helps shorten the period of time it takes to achieve statistically relevant test results. If you test your new page on only 10% of your site’s traffic, it can take five times longer to reach statistical significance. Achieving that threshold is important, since early A/B test results may show outcomes that don’t hold up when applied to larger audiences. When tests reach, say, the 95% confidence level, there’s only a 1 in 20 chance the outcome was due to random chance.
Eliminate outside factors. There are a host of outside influences that push test results into unreal territory. These could include running (or ceasing) advertising campaigns, initiating price changes, changing privacy or return policies, or changing brick-and-mortar hours during the test. If possible, conduct testing when outside influences are minimal.
Don’t forget AI. Bots can generate a stunning amount of web traffic and warp the best testing protocols. Look closely at your analytics to see where traffic is coming from, and filter for bots by default.
Good testing processes will eventually yield more conversions. Yet with so many page elements contributing to conversion success, testing can be daunting. Thankfully, there are a number of plugins dedicated to helping you with this task. Choose a few to get started, and learn more about how you can work conversion wonders with WordPress.
Work Conversion Wonders
was written by me, Greg Norton – also known as webzenkai. I’ve got more than two decades’ experience building effective websites and powerful email campaigns that yield results. Feel free to contact me regarding this article or anything else you find on this website.