Case Study
Head-count planning using Monte-carlo simulation for IT Support Team
Use Monte Carlo simulations to build the model that predicts head-count needed for the work-load.
Features:
| Factor | Lower Bound | Upper Bound | Most Likely | Units |
| Number of new support tickets | 20 | 50 | 32 | tickets / week |
| Time to resolve new support ticket | 6 | 24 | 8 | hours / ticket |
| Number of support analysts | 15 | analysts | ||
| Hours per analyst per week | 40 | hours / analyst / week |
With the current capacity of support team, there will be a few weeks – around 2 weeks in a year when support analysts will be working full capacity and yet tickets might roll over into next week. Which I believe is ok as it is rare occurrence.
Normal Weeks
In a given week, a support analyst spends 21-33 hours to resolve support tickets. And we can say this with 80% confidence interval.
A support analyst ends up spending 3 and half days or more, working on tickets in more than half of the weeks. Which is very close to our company policy which requires 80% of resource utilization for support team – that is 4 out of 5 working days in a normal week.
Unusual Weeks – bad and good
There can be 10 weeks in a given year when all the analysts will end up spending more than 32 hours each. Which might be ok in context of maintaining sustainable pace for the team. As there will be 10 weeks in a given year when an analyst will spend 20 hours or less on support tickets.
Long range explained
It is just the unpredictable nature of the job. It depends on tickets influx and complexity of the tickets which drives resolve time for the ticket. Support team does not have direct control over both of these factors which drive its utilization levels.
Customer needs resolution in a week
We also know that as per industry standard, support team needs to resolve a ticket in a week. Customer shouldn’t be waiting more than a week for the resolution. Hence, it is best to complete the ticket in the same week it came in. We do not want tickets to roll over into the next week.
14 is too less for sustainable pace of the team
Please note that even if one support analyst leaves, the team will start losing the sustainable pace. Each analyst will be working 22-35 hours a week 80% of the time. There will be 25% of the time an analyst will end up working for more than 32 hours. We will have 13 weeks in a year when an analyst will be utilized more than 80% for support work.
14 is too less to keep customers happy
It is worth noting that if team reduces to 14 analysts, there will be 6 weeks in a year when tickets might roll over to next week. Which might cause reputation issues with customers.
To avoid the long-tail impact, keep the team count at 15.
16 is too relaxing
There will be just 3 weeks in a year when an analyst needs to work more than 32 hours.
In a normal week analysts would spend 18-30 hours to resolve support tickets.
Hence, adding one more analyst to the team does not add value – it will just make everyone’s life easier and probably boring.