Loading
Set the resolution target you can hit in the worst week rather than the average one. An SLA you breach in December is worse than a slower one you always meet, because the exception is what customers remember.
It takes your ticket volume, staffing and cover hours and tells you whether a proposed SLA is achievable, before you commit to it in a contract.
Most breached SLAs were unachievable when they were signed, and nobody did this arithmetic.
Enter arrival volume by hour and day, how many agents are available in each window, average handling time, and the response and resolution targets you are considering.
Use real arrival patterns rather than a daily average. A day with a Monday morning spike behaves very differently from one with even arrivals.
It will not account for complexity. Handling time varies enormously by ticket type, and an average hides that.
It also cannot model absence or attrition, both of which make real performance worse than the model.
The output is the proportion of tickets you would meet the target for, given that staffing. Anything below the high nineties means the SLA will breach routinely.
Check the peak windows separately. An SLA that works on average and fails every Monday is an SLA that fails.
Whatever your staffing supports. That is what this works out.
Yes, and be modelled separately.
No.
Half an hour on your own figures, and an honest answer about the parts Treepie does not improve.