This is related to the reports feature. For a better understanding, please read: "How to create an indicator?"
A result is statistically significant when it's almost impossible to get it by chance.
More specifically, we compare the results of questions related to different segments of population.
Please keep in mind that 2 types of data are concerned with this test: Average & Percentage. The test is available when your segments of population are over 30.
Use case:
We cross a question with another one. For this example, we'll study the visitor satisfaction of an e-commerce website.
#1 question - Are you satisfied? // #2 question - Reason of your visit?
As you can see, the satisfaction score is really different depending on the reason for the visit. Also, it's important to keep in mind that as this is an average, different quantities of people answered each question (1000 people could answer "Buy a product" and 5000 "Buy a service"). This is the perfect case where a test of significance is required, as you want to know if those results are logical i.e. significant, or meaningless. As you crossed 2 dimensions and have a population over 30, you have nothing specific to do. The test of significance triggers automatically.
The test is represented by colored cells, meaning those cells have been tested with the significance test. If there is no colored cell in your table, your results didn't have to be tested. From a statistical view, the average (or the percentage) of the cell which refers to another cell is significantly superior to the average (or the percentage) of the referenced cell (5% margin of error). After the test, the answers you need to study are "B - Sign In" and "C - Other" to validate the answers are significant. The letter(s) in the cell indicate(s) that the cell which contains the letter(s) is significantly superior to the cell corresponding to the letter.
As for our example, the result of the significance test validates the fact that people who came to sign in are more satisfied than people who visited the website for another reason. It helps identify and statistically validate areas of improvements for the website!
You can also find this data in the HTML (or PDF) export of your statistical report.