APDEX or how to measure user satisfaction?

APDEX or how to measure user satisfaction?

What is user satisfaction?

User satisfaction is a complex subject because it has a "fuzzy" definition. Indeed, it can depend on various factors such as for example :
LErgonomics or the ability to interact with the application
The ability to provide the expected functionalities
The availability of the application
The ability to deliver results quickly
…

The criteria associated with these aspects depend on specific requirements for a given application. In addition, these factors are highly dependent on the type of application and user expectations. It is not possible to define a priori whether a user is satisfied based only on technical metrics and without knowledge of the application or users.

What is APDEX?

The APDEX measurement standard (for Application Performance Index, see http://www.apdex.org/ ) provides a clear definition of user satisfaction in terms of response time. From now on we will only deal with the ability of an application to respond quickly and the other aspects related to user satisfaction will no longer be discussed in this article.

This measure applies with the measurement of response time of a set of requests over a given period of time. APDEX needs 2 parameters that define the response time requirements of the application:
The maximum response time beyond which the user experience is described as "tolerable". With a response time below this threshold, the user experience is described as "satisfactory". We will call this value 'T'.
The maximum response time beyond which the user experience is described as "frustrating". This response time is greater than the previous 'T' parameter. We will call this value 'F'.
The APDEX is a measurement between 0 and 1 and its calculation is based on the two above-mentioned parameters. The formula for the calculation of this measurement is as follows:
Apdex = (#Sat + (#Tol / 2)) / #Total

Where :

##Sat is the number of requests whose response time is less than 'T'.
#Tol is the number of requests whose response time is greater than 'T' and less than 'F'.
#Total is the total number of requests. If #Fru is the number of requests with a response time greater than 'F', then we would have #Total = #Fru + #Tol + #Sat.

APDEX is therefore a weighting on the number of requests according to their response time. The response time being discriminated in 3 sets: satisfactory, tolerable and frustrating.

From APDEX to user satisfaction

The 'T' and 'F' parameters are satisfaction thresholds in relation to the application response time. With these thresholds and response time measurements, we can therefore measure user satisfaction with the application's ability to respond quickly.br />
APDEX is a ratio with a value between 0 and 1, so to make the information more understandable, it is easy to convert it to a percentage (simply multiply by 100). This percentage is correlated to the number of queries through the weighting defined by APDEX. It is not exactly a percentage of satisfactory queries, but is still a percentage related to the number of queries of certain types (satisfactory / tolerable / frustrating). This can be used as a reference from which the improvement or degradation of the system can be observed.

A percentage based on a standard is a very good way to understand user satisfaction.
APDEX is a good way to measure and understand if the satisfaction (in response time) of users is there when they use your application.