By Paul Parcellin, Staff Writer
November 14, 2014

The word "test" strikes fear into the hearts of many a student, and the mere mention of the word may even make you feel a little queasy.

But let's assume that there are tests that can actually help you rather than make you anxious. If you're an online lead generator, you'll probably want a method of evaluating your website that many marketers swear by, called A/B testing. If you already know about A/B testing and have yet to implement it you may want to give it a try. Here are some tips to get started.

So, Why Test?

Your pages usually contain different kinds of material, and each item is placed on the page for a reason. You may want to increase content sharing, get more email addresses or improve your conversion rate. Ultimately, you probably want to generate leads at a quicker pace.

Although you may not be getting exactly the response you're hoping for, it's not a sign that everything on the page is bad. Fortunately, you can get a better idea of which parts of your pages are working and which ones aren't by using A/B testing.

A/B testing is pretty much what it sounds like. You have two options and you test user response to see which does the best with whatever metric is most important to you - bounce rate, sales or conversion rate, to name a few. You channel half of your traffic to one version of the test page, and the other to an alternative version of the same material. When testing is complete you compare the data and see which version performed best.

So, what sorts of tests should you start with? Stephen Walsh, writing for KissMetrics, names "19 Obvious A/B Tests You Should Run on Your Website." Here are some highlights:

Typography, serif vs. sans serif. Many say that serif type is for print and sans serif is for the Web. But according to Walsh, a serif typeface (Georgia) is the most popular one on the Web.
Colors that work best with your call to action. You'll probably want to stick with good old fashioned black type on a white background when it comes to most text on your Web pages. It's simply the easiest combination to read and it's what most people are used to. However, your call to action is a different story. It's usually going to be short and forceful, so you may want to experiment with colors to see which ones help drive your point home.
The font size that will be most legible. Readable text can have an impact on your click-through rate, notes Walsh, so you'll want to make sure that the point size you choose is appropriate for your font.
Test typefaces, but don't go overboard. There are hundreds of typefaces available. Just choose a few of the major ones and test those. Walsh further advises that you test more than two at a time for expediency purposes.

Positioning your call to action. Many times the call to action is placed at the top center of the page because many seem to believe it's the best place for it. But you should keep an open mind and test it on all regions of the page, even beneath the fold.
Call to action text. Your call to action text must be thoroughly tested.
Landing page copywriting. Test the results you get with long-form and short-form copy. Neil Patel found that long-form copy produced 7.6% more leads. On the other side of the spectrum, a Scandinavian gym chain got 11% more conversions with shorter copy. Find the right copy length for you and your business.
Video vs. text sales pages. For some sites, video brings a higher conversion rate than text. It's expensive to produce, but for the right industries it may be the best solution.

You might also want to test page elements such as the number of columns you use, background images and patterns, navigation links and link colors.

‎The Long Haul

John Corell, CEO & Founder at Conversion Voodoo, says that long-term testing is essential if you want to get at the facts. He advises marketers to ignore instant gratification and take the time to look at all factors that could be influencing your results.

Don't make swift assumptions about test results. There are plenty of false positives, says Corell, especially when you test multiple versions of the same page. You'll know that when you change your page to reflect the test data and the page does not perform the way you had hoped. The answer is to think, reanalyze and test again. A/B testing is a long term solution, not a quick fix.

Benny Blum, vice president of Performance Marketing & Analytics at sellpoints, says that test design means everything when it comes to gathering useful information. It can also reduce the chance of false positives, he maintains. Blum champions the work of British statistician R.A. Fisher, who pioneered design of experiments in the 1920 and '30s.

Fisher advocated testing "against a control," otherwise known as A/B testing. He also said that participants in test and control groups should be assigned randomly, and that accuracy and consistency can be better ensured by repeated testing.

One of the most important parts of a well-designed experiment is a large sample size, Fisher said.

Blum maintains that you must collect lots of results before your testing will be statistically significant, so if you're using a modest sampling of results you're probably operating from a position of weakness. However, you don't have to guess whether or not you are doing enough testing. Statistical power is a test that can determine whether or not your sample is large enough to be significant.

Tools to Get the Job Done

There are a number of different software options for those who want to delve into A/B testing. In an article titled "How to Choose the Right Testing Software for Your Business," six A/B testing packages are evaluated for their effectiveness and cost. They are as follows:

• Google Analytics
• Optimizely
• Visual Website Optimizer
• Unbounce
• KISSmetrics
• Crazy Egg

Prices range from free (Google Analytics) to $359 a month (Optimizely).

Most A/B testing software charges by the impression, so there's an economic disincentive to go for the high numbers. But on this, Blum offers a rule of thumb: "If you think your test is done, test a bit longer."

Material Disclosure: The content contained herein is provided for informational purposes only and should not be considered as solicitation, advertisement or expert advice. The operator of this website assumes no responsibility or liability for any actions taken as a result of using this website, and/or for errors or omissions in content. When you link to any third party website and/or article provided via this website, you are leaving our website. We make no representations as to the information provided at third party websites and/or third party privacy practices. Please consult with your own independent legal advisor before relying or acting on any information provided on this website and/or a third party.