Aggregating Earliest Time Correctly

I'm working on a dataset where for a single user the data would look like below:

User, Activity, Time

User 1, SignOn, 7:50am

User 1, EnRouted, 8:00am

User 1, Arrived, 8:30am

User 1, Completed, 9:00am

.

.

.

User 1, SignOff, 4:00pm

And I can create a table showing this - filtering down to every user and find their earliest 'Arrived' status for a given day by using a filter on Arrived and giving me the earliest Time.

I want to take this a step further and look at multiple days/multiple users and find what the average earliest 'Arrived' status is.

For example:

User 1, Arrived, 8:00am

User 2, Arrived, 9:00am

User 3, Arrived, 9:30am

My average 'Arrived' time for the three employees should be 8:50am. 

Any help is much appreciated.