You have a doubt about the meaning of a data? You are not yet a professional of rates? You will become one! Find on this page all the key words to understand and analyze the results of your experiments.

*Do not hesitate to use the navigation menu on the right to directly find the word that keeps you awake at night.*

# Reliability (confidence rate)

This is a statistical measure that determines the probability of a variation to beat the original. For example, if a variation obtains a 95% confidence rate, it means that there is a 95% chance that this variation will beat the reference. At the beginning of a test, this rate can vary from one day to another: the results of a test are considered significant when this rate is stabilized. For a variation to be successful, the reliability rate must be very high (at least 90%) and constant.

This information is very easy to visualize thanks to the evolution curve of the rate. In the tables of results, a visual indicator reflects the stabilization of the reliability: if the 3 boxes are illuminated, it means that your rate is stable.

# Number of pages viewed

Number of pages visited on your site, per visit or per visitor. The landing page (first page on which the visitor arrives on your website) is included in this count.

# Number of visits

Total number of visits. There can be more visits than visitors, because the same visitor can visit the page multiple times.

# Number of visitors

Total number of visitors. If the same visitor visits the page multiple times, he is counted only once.

# Number of converted visits

Total number of visits during which visitors converted your goal.

This indicator is proposed in the table of results if you have chosen to display the visits (and not the visitors).

- If the “Converted Visits” option is enabled, then the number displayed is the number of visits in which a visitor converted the goal at least once;
- If the “Total conversions” option is enabled, then the number displayed is the total number of times the goal has been converted.

# Goal

This is the conversion goal you want to improve through your A/B test.

Several types of objectives are available:

## Access to a page

This goal is converted if the visitor accesses a page whose URL you have defined.

## Retention rate

This goal is converted if the visitor has visited other pages after the landing page.

## Number of pages viewed

This goal is converted if the visitor has visited the number of pages you defined.

## Number of converted visitors

Total number of visitors who converted the goal at least once.

This indicator is proposed in the table of results if you have chosen to display the visitors (and not the visits).

- If the “Converted Visitors” option is enabled, then the number displayed is the number of visitors who converted the goal at least once;
- If the “Total conversions” option is enabled, then the number displayed is the total number of times the goal has been converted.

## Personalized goal

You can configure a goal yourself. For example, you can decide to track the transactions made on your site, and define a number of transactions or an average basket value to reach to validate the goal.

## Time spent

This goal is achieved if the visitor has spent a certain amount of time that you have defined.

## Tracking clicks

This goal is reached if the visitor clicked on a defined item.

# Average basket

The average basket is the average value of purchases made by a customer. To calculate it, we divide the turnover by the number of orders.

# Conversion rate

It is this measure that you seek to optimize during an A/B test.

- If you are in “Visits” type of view: percentage of visits during which your goal is converted at least once (we divide the number of visits by the number of visits converted).
- If you are in “Visitors” type of view: percentage of visitors who have reached a goal (we divide the number of visitors by the number of visitors who converted).

# Rate of improvement

It is the conversion rate improvement compared to the original. This rate is calculated for each variation towards the original, so the original has an improvement rate of 0. To obtain an estimation of this data, subtract the conversion rate of the original to the conversion rate of the variation, then divide the result by the conversion rate of the original. At Kameleoon, we use an even more reliable calculation method, using the raw numbers. If you do the above calculation, you may notice a slight difference with the result we provide: it is a rounding error! Trust in our result.

On hover, two rates appear: this is the reliability interval. It indicates the interval in which your actual improvement rate may be. The improvement rate has a 95% chance of being between the minimum and the maximum displayed.

# Breakdown

The breakdown is a sorting tool that allows you to display more precise results: thus, you can display the details of your results by browser, operating system, days of the week, etc.

It is also possible to filter the results. The filter, unlike ventilation, only displays the results for the chosen filter.

Example:

- If you allocate your results by device type, your data will all be displayed but divided into three groups (computer/smartphone/tablet);
- If you filter your results by device type and choose the “smartphone” filter, only data about visits/visitors on smartphone will appear; the others will not be displayed.