Email staffing requirements

[ChrisH]

I am interested in how others are incorporating email service level objectives when determining staffing requirements. For example, if the email SLA is 100% response within 6 hours, how do I determine the requirements (assuming I know the processing time)? One thing I would like to do is evaluate the cost/benefit of an email SLA of 100% in 6 hours vs. 100% in 2, 3 or 4 hours. Any feedback or suggestions would be welcome.

Thanks!

Chris

[JoeB]

Unlike staffing for telephone calls, which involved complex queuing theory statistics, staffing calculations for email response can use a more straightforward approach because there is unlimited queuing capacity, relatively very long response time, and no concern for abandonment.

So to get a good approximation, simply calculate the email clearing rate (3600/AHT in seconds). Dividing the email volume by that number will give you the staffing requirement (donÂ’t forget to round up). This approach assumes utilization of 100%, which is not quite practicalÂ….

To get more precise results you can use an Elrang calculator that supports the range of calculations you need (e.g. www.DiagnosticStrategies.com/EasyErlang.htm), and set the target SLA to 99%. An important added value of using this type of software is that it will also calculate staff utilization, so you can select staffing level that will result in acceptable utilization.

What may be confusing is the fact that when you target 100% of the email messages there will be no (significant) difference between staffing for 100% in 1, 2 or 3 hours.

[sgordon]

We use the following calculation to determine email staffing needs:

Volume/(response time goal/average handle time)=Agents

Or =100/(480/10) where 100 emails received have to be handled in 8 hours where the average handle time is 10 minutes. The staffing requirements in this scenario would be 2.08.

I hope this helps.

[JoeB]

Sherri,

I think that this approach may be misleading. If the *total* volume of email is 100, and the call center has 8 hours to respond to them (i.e. no additional emails are received during these 8 hours), then your formula is will produce the correect result. But if 100 is actually the email *rate* (100 emails per hour), then this formula will produce the wrong result.

If the email rate is 100 per hour, the required number of agents is: 100/(3600/600)=16.667, or 17 agents.

If he total volume is 100, then effective rate is only 100/8=12.5 email per hour, and the required staff is: 12.5/(3600/600)=2.08 (3 agents).

Joe

[dlexmond]

Solving the e-mail equation is actually far more complex than the one for phone calls. Calculating staff requirements is simple, the examples in the other responses indicate that. It is creating schedules that match the resource requirements, and generate the desired service level is far more challenging.

Because phone calls abandon if not answered in a reasonable amount of time, you don't have to worry about carry over from one time interval to another, and the service level clock can be set for each interval.

With e-mails you have a legacy - if you did not complete all the work in a given interval you have to staff for it in the next, as well as for the work that comes in during the new interval. Because service levels are expressed in hours instead of seconds, it is actually not advisable to staff appropriately for each interval - if you did your service levels would be less than the length of the interval that you schedule for. So by over- and understaffing by interval you need to balance out the fluctuations of inputs and carry overs, so that you balance productivity and service level.

I'd be happy to share my experiences with you. My e-mail is dlexmond@callcpm.com.

[ricg]

Hi, im having a problem with my email scheduling. I've followed all the formula but im having a problem calculating the number of agents on a weekly basis. I have an 85% in 24 hours Response time with let say a 10min AHT... what formula should I use?

[JoeB]

A general comment to dlexmond. The reason why email staffing calculations is easier is because you are much less sensitive to load fluctuations and to the Poisson distribution of arriving emails. If your load and AHT statistics are accurate then you will be able to catch up on the backlog. The “generous” response time (e.g. 85% in 24 hours) is designed for that. Conversely, the tight response time requirement for live response is much harder to deliver and easy to miss, and I’d rather not see abandoned calls as a solution for missing the service level threshold.

If you are having hard time getting there, then the problem is likely in high variability and not in the actual calculations.

Ricg: I am not sure if you are unable to use the formula to determine staffing in general or only for weekly schedule? Unless you have significant differences from day to day, the same calculations apply. What formula have you tried?

It is impossible to assess the number of staff you need without knowing the email volume…

[ricg]

im using the following formula to compute for the staffing requirement on a 30min interval:

volume/(30mins/AHT)/.80 - - > Efficiency or utilization

But for the daily requirement, i use this formula:

Base Staff=volume * AHT / shift (7.5hrs) * efficiency of .8

or volume/(Response time - Increment or length of shift)/AHT/Efficiency

Please advise if Im doing it right? Also, how do I compute for the weekly requirement if I have a 15-16K volume with an 85% in 24 hrs TSF or response time, with a 10-12mins of AHT? Is it safe to compute based on man hours let's say, I get the max staffing per interval of the day and multiplied it by 7-8 hours and multiply it by 5 and divided it by 40hrs?

Hope you can help me on this as I have report for tomorrow :( thanks a lot!