Being suc­cess­ful at A/B test­ing means hypoth­e­siz­ing poten­tial prob­lems on your site and test­ing these ideas. With the right tools, you’ll be able to bet­ter under­stand who your audi­ence is and increase con­ver­sions across your site. Remem­ber: the more hypothe­ses test­ed, the more suc­cess­es you will have.

A/B test­ing is a way to test mul­ti­ple ideas on your web­site and see which ideas increase con­ver­sion rates. A/B test­ing ideas can come from insights formed from track­ing users’ activ­i­ty across your site or sur­vey respons­es from your tar­get mar­ket.

In order to be suc­cess­ful at A/B test­ing, you must first find the fric­tion-caus­ing ele­ments. Then you must define the vis­i­tors’ con­cerns that are keep­ing your cus­tomers from con­vert­ing.

After defin­ing top site issues, cre­ate and test new hypothe­ses based on the data you uncov­er. You can use a num­ber of research meth­ods to pin­point prob­lem areas and test those areas to increase con­ver­sion rates.

Below is a sam­ple of help­ful tools.

1. Usability Studies

A usabil­i­ty study involves:

• Recruit­ing study par­tic­i­pants to vis­it your web­site.
• Pro­vid­ing par­tic­i­pants with a list of tasks.
• Ask­ing for feed­back as the par­tic­i­pants nav­i­gate your web­site.
• Record­ing the whole process for analy­sis.

Brands have been con­duct­ing tra­di­tion­al usabil­i­ty stud­ies since the 1990s and focus groups since the 1950s. How­ev­er, in recent years usabil­i­ty stud­ies have become more pop­u­lar through tech­nolo­gies like UserTesting.com.

UserTesting.com has tak­en the tra­di­tion­al mod­el of usabil­i­ty stud­ies and brought it to life in an online plat­form. The site allows you to recruit par­tic­i­pants and man­age your usabil­i­ty study videos; all while promis­ing to deliv­er your study results in 24 hours.

When con­duct­ing a usabil­i­ty study you should test the effec­tive­ness of your site by ask­ing the right ques­tions to the par­tic­i­pants on your site. You must be care­ful in how you struc­ture ques­tions and be sure you are keep­ing an eye on the behav­ior of your par­tic­i­pants.

To avoid lead­ing the user into an answer, the ques­tions must be writ­ten in a way that can’t be mis­in­ter­pret­ed as direc­tion. It’s also impor­tant not to hint at the answer you’re look­ing for since this, too, will skew the results.

2. Cart Abandonment Surveys

Cart aban­don­ment sur­veys allow you to tar­get site vis­i­tors who are show­ing intent to leave the page where they cur­rent­ly are. If you want to know why the exit rate is so high on your cart or check­out pages, the cart aban­don­ment sur­vey is the eas­i­est way to gain cus­tomer feed­back.

The answer for­mat can be mul­ti­ple choice or open answer. With this infor­ma­tion, you can grasp an under­stand­ing of the site usage, the types of vis­i­tors, the visitor’s mentality/mindset or the rea­sons for aban­don­ing the site dur­ing a cru­cial step.

Here’s an exam­ple of what a site sur­vey ques­tion looks like when a user is exposed to a sur­vey:

And the results for that spe­cif­ic ques­tion:

3. Consumer Surveys

Mass con­sumer sur­vey tools, such as Google Con­sumer Sur­veys, allow you to tar­get the audi­ence by age, loca­tion, income, etc. This tool helps you gath­er data quick­ly by seg­ment­ing your sur­vey tar­get audi­ence and cre­at­ing a wide range of sur­veys that can be con­duct­ed though A/B tests, brand track­ing stud­ies, mar­ket trend analy­sis, con­cept devel­op­ment, prod­uct devel­op­ment, mar­ket design, and cam­paign mea­sure­ment.

As soon as you receive your results you also have mul­ti­ple ways to seg­ment sur­vey respons­es so you can quick­ly glean insights down to a seg­ment­ed demo­graph­ics lev­el.

For exam­ple, you may notice that con­sumers who live in rur­al areas have a high­er chance of own­ing a pet or that younger audi­ences are more open to mobile apps than your old­er audi­ence. The chart below shows Google Con­sumer Sur­veys’ var­i­ous options to fil­ter your sur­vey results.

4. Heat Maps

Heat maps reveal where your web­site vis­i­tors are most fre­quent­ly look­ing and click­ing. This infor­ma­tion will help you under­stand the eye path of the user and what attracts the most atten­tion.

Heat maps can be used in nav­i­ga­tion flow tests to ver­i­fy that the user is click­ing in the cor­rect spot on the page to con­tin­ue their path to pur­chase.

5. Scroll Maps

Scroll maps are sim­i­lar to heat maps in that they indi­cate where the cus­tomers are scrolling to and spend­ing most of their time. This data usu­al­ly indi­cate what stands out to the cus­tomer and where infor­ma­tion needs to be moved.

You Can Learn From Losing A/B Tests, Too

Not all test results from these stud­ies are going to be pos­i­tive. Yet, a neg­a­tive result doesn’t have to stay neg­a­tive.

Some of your best per­form­ing A/B tests could come from the insights gained from a los­ing test vari­a­tion. For exam­ple, you might notice that one of your head­lines does well while the oth­er one fails. The fail­ing head­line will teach you just as much about your audi­ence as the suc­cess­ful one.