What methodology do you use when estimating field service calls durations?

Scheduling service calls for your technicians in the field can be very challenging. One of the variables to producing an accurate schedule is correctly estimating how long a repair task will take. 

In my organization, we service hundreds of different models. Some equipment could have a service duration of 30 minutes while others have over 2 hours. I'm in search of best practices used by field service delivery organizations around managing call durations. 

• What methodology do you use to estimate the duration of the service event? Do you use engineering service targets, historical duration metrics, or something else?

• What best practices do you have when a technician is close to exceeding or has exceeded the duration of the event? Do you send out alerts to the technician or manager? Do you force them to escalate to second-level support? 

• If your technician can not resolve the issue on his or her first visit and must return on the following day, how do you estimate the amount of time required for the second visit? Example: A service call is expected to take one hour and twenty-five minutes to resolve and therefore the scheduling system blocks an hour and twenty-five minutes on the technician's calendar. At 1 hour the technician closes the first call and a return visit is scheduled. How does your organization calculate the duration of the second visit? Does the technician get 25 minutes, another one hour and twenty-five minutes, or something else?

• If a service call is going to take longer than expected it is important to keep the scheduling system updated with the technician's expected finish time. This will allow additional service calls to schedule correctly. What are your best practices that allow your technicians to update their expected finish time?

Please feel free to share any other best practices. 

Thank you.