Tasks are defined as a sequence of interactions that the users have to perform. Analytic tasks allow you to measure how long and how many clicks users need for the execution.
Task KPIs
Once the task flow is defined the task list will update and show several KPIs. These KPIs allow you to get a first understanding of whether the testers' behavior is inline with your expectations.
Example: You have created a new design for a shopping app, and you are interested in how the design works. You run some tests and see that the users took an average of 2:30min. This is much more than expected, so you should do a deep dive and check out the detailed statistics to find if there are only some users that have issues and skew the KPIs, or if there is a general problem.
Start
The start rate tells you how many users were able to start a given task.
Example: Let's assume you have a prototype with three screens. A, B and C. A is wired to B, and B is wired to C. Also, you have defined a task, that measures if users go from B to C. If a tester views only screen A, this test session is not counted as started. If the tester visits A and B, the session is counted as started.
A low start rate indicates that the users have issues navigating in your design. They were not able to navigate to the right screen.
Success
The success rate tells you how many users were able to complete a given task once they have started.
Example: We continue with the example outlined above. Let's assume we have three testers. Tester X visits only A, tester Y visits A and B, and tester Z visits A, B and C. In this case, the start rate would be 2 / 3, because testers Y and Z managed to visit screen B. The total number of successes is however only 1 since only tester Z managed to visit screen C. The success rate is thus 1 / 2.
In general, you want all users to understand your design and be able to complete the tasks. Hence, the higher the success rate, the better your design.
Duration
The task duration measures the average / mean time the users take to complete a task. Most of the time short durations indicate that the users understood your design.
Example: We continue with the example outlined above. Let's assume we have 4 testers that have completed the task. They took 10s, 12s, 10s and 28s. In this case, the mean duration would be (10 + 12 + 10 + 28) = 60s divided by 4, so 15s.
Please be aware that often duration times vary a lot between individual users. As you could see in the example the first three users were quite fast, whereas the last user was slow. It would, in such a case, make sense to review the screen recordings or check the user journey graphs to understand the user's problems.
One can click on the to open a detailed view of the distribution of the durations in the scatter plot.
Interactions
The task interaction measures the average / mean number of interactions the user needs to complete a task.
Example: We continue with the example outlined above. Let's assume we have 2 testers that have completed the task. They took 5 and 15 clicks. In this case, the mean interaction would be (5 + 15) = 20 divided by 2, so 10 clicks.
Like the duration, the number of interactions may vary a lot between the individual users. The mean value might hide such case outliers, that are interesting for further analysis.
One can click on the to review the distribution of the interactions and spot outliers.
Task Charts
You can get a deeper insight into the distribution of the duration and interactions by clicking on the button. In the dialog, you can switch between scatter plots and box plots. Scatter plots show the duration of a task vs the number of interactions and make it easy to spot outliers, for instance, a user that took an unusual amount of time and clicks.
Clicking on a dot in the scatter plot will open the screen recording, so you can easily understand where the testers struggled.
Box plots show also the duration and interaction but in a condensed way. They visualize the distribution and skewness of the data, by showing the average and 75% quartiles.