Hi All,

Just a quick question on calculating the Email FTE estimates -
for calls we use the erlang C, and get the estimates.

But when we look at the Emails, and if we have service level like
90% < 60 mins how do we do this??

Ex - Total interactions 10,000
60% calls and 40% emails
90% < 30 secs for calls and 90% < 60 mins for emails
AHT for calls - 10 mins and AHT for emails 12 mins
support hrs - 24 * 7

Please let me know how to go bout this,
specially for estimating on the emails

Thanks in advance
Regards,

Santosh,

As the old maps used to say....

Here be Dragons.....

I'm away for a few days but I'll have a
think and get back to you.

I seem to remember a discussion about this a whille ago.

Have a dig around with the front page search function.

Regards

DaveA

Thanks for that,

Hoping for an answer, its becoming a bit challenging when we have the mix of calls and emails.....


Thanks,

Hi Santosh,

I think you need to be careful here - Erlang assumes the work "goes away" after some period of time - I.e. calls abandon after a period of time.

Trouble is emails don't do this - they back up over time and still need doing whether or not you made your service level. ~This means your FTE resources required keeps going up if you do not make your service targets - i.e. you need more people tomorrow to cover the backlog.

The other thing about emails is they do not all need a reply - so how are you measuring service level - is it a target reply time or target completion of the inbound email?

Seb

Hi,

Thanks for the reply,
I understand the point you raised,

For the emails we have 2 types of service levels

1)Response Time - Immediate response with possible solution can be provided - 90% < 60 mins
2)Resolution Time - Complete resolution of the issue - 24 hrs

Hope this answers you,

Thanks,

Hi Santosh

I presume these are weekly interractions? If so and based on a 37.5 hour agent FTE week then you need about 80agent FTE to cover the phone and e-mail interractions. 63 would be needed to handle calls and e-mails and 17 who would just handle e-mails.

Eamon

Hey Eamon,

Thanks for the reply,
Is it possible you to share how did we arrive at 63 for calls and 17 for emails?

If we have the mix of calls and emails is there a possibility increasing their utilization,

Current utilization is 59.09%

Thanks,

Sure

It takes about 63 FTE to cover the phone traffic 24/7 using an erlang model that I have in excel. If you consider that their utilization is reasonably low then you can get them to do e-mails as well up to an upper utilization level (say 85%). That takes care of about half of the email work. The other 17 are then just worked up as emails x AHT / upper utilization level / (1-shrinkage - 30%ish) / 37.5

Eamon

Hey,

Did you consider the Service level target (90% < 60mins) while estimating the Email FTE, which is the main criteria?

Since when we add this email work load to the call taking agents we should ensure that we are meeting this target as well.

Let me know if you have any clarifications on this,

Thanks,

Hi

I did, the e-mail service level only affects the risk involved, not the FTE requirement. eg, you have an e-mail team that needs 20 FTE to turnaround 1000 emails per week ensuring that 90% of e-mails are replied to within 48 hours. If you make that service level more difficult then it does not change the number of FTE you need (as the volume and AHT remains unchanged) you may want to employ more FTE in order that you guarantee service level but that is more down to managing risk rather than the new service level.

Eamon

Eamon,

So..

Can we (theortically) make the assumption that
E-Mail arrival is a Poisson process?

If so I may have something.

Regards

DaveA

We can

Training the fish and waterproofing the computers is the tricky bit

E

Hey,

If we have to estimate the FTE's based on the assumptions saying -

Agent is multi skilled and will be handling calls and emails both, how do we confirm that we are meeting the service levels for both calls and emails.

For calls we do use the erlang (SL 90% <60 secs), but for emails can we use the same? (SL 90% < 3600 secs).

Is it ok to consider the weighted average for both calls and emails, then input the same?
Ex - Total interactions 10000
Calls - 50% (5000) and SL 90% < 30secs
Emails - 50% (5000) and SL 90% < 3600secs
Can I take the ASA as 1815 secs, Hence my overall SL target would be 90% < 1815 secs

Is this the right approach??

And when you work it through a call centre modeler it comes out at 80.3 FTE so in the respect that there are two different ways of modelling it and they come out with the same FTE requirement I'd say yes.