Hi,

Just signed up. Would just like to ask your opinion on this. When computing weekly FTE requirement, we use the Erlang add-in to project our staffing assumptions. Here is how we use it:

SL - 80%
Service time - 30 secs
Weekly volume - 2500
Operation Hours - 24 by 7 = 168 hours
AHT - 500 secs
Shrinkage - 25%
Occupancy - 75%
Agents hours per week - 8 hours per day / 5 days a week = 40 hours

Agents(80%,30,(2500/168),500)*168/(40*(1-25%)*75%)= 30 HC

Your thoughts please. Also, because I plug in Occupancy in my computation, my requirements bump up, without occ requirement is 22 HC. The reason i bake in occu is that i am assumming my agents are on phones only 75% of their working hours.

Appreciate any rebuttal on this.

Thanks,
Mike

Appreciate any rebuttal on this.

MIchael,

I'm not sure why! {GRIN}

It should work, just a few points,

i) Your data isn't granualr enough for an accurate forecast, however, Erlang
will give you a good approximation. Obviously better than without" The
probability function will help with the distribution.

ii) What does the OCC figure include? Is it just Talk + Wrap? You'll need
to factor a small % for water, loo etc...

Apart from that as a weekly it's OK.

All you need to do now is break it down to the half hour.

Have a look around here, there's a few good threads on the topic.

Welcome to CCv, good luck, don't be afraid to ask more and do
let us know how you're getting on.

Regards

DaveA

HI Michael

Not sure I understand your calculation?

Agents(80%,30,(2500/168),500) * 168/(40*(1-25%)*75%) = 30 HC

If we break it down so make its easier to read I get:

Agents(80%,30,(2500/168),500)= 12 Agents
168/(40*(1-25%)*75%)= 7.47 (not sure what this is?)

12*7.47= 89.64???? (I don’t get your result from this calc)
------------------------------------------------------------------------------
I understand the first part where you work out the workload FTE. I would have then applied the OCC and shrink after that:
12 / (1-25%) / 75% = 21.33 FTE

This is how I would do it?

Do you factor schedule inefficiency into your shrinkage??
I guess this would be assuming your call arrival patterns are very flat or even spread?

Dave,

Appreciate your response, thanks. Occupancy is "like" utilization on our end, and yes, we consider talk,wrap and hold. The other breaks, e.g water, restroom, are factored in the shrinkage assumption, comprising the 25%.

I just have to revalidate with guys like you for "check and balance" of some sort.

Michael,

Agents(80%,30,(2500/168),500)= I get 4 agents, not 12.

168/(40*(1-25%)*75%)- this side of my equation spreads the erlang formula in terms of week view, since i am particularly concerned with weekly HC requirements. Since the agents erlang formula is in hourly period, i estimate my weekly requirements by multiplying it with the operation hours, which is 168 (24*7) and dividing it with the number of hours each agent is working in a week (with shrinkages and occupancy factored in)

Hi Michael

If Occupancy is 75% what are the guys doing in the other 25%?

Eamon

Hi Eamon,

The occupancy assumption covers agents actively taking calls, at 75% (or 85% if operation is not 24 by 7). The remaining percentage is considered avail/idle time. as leeway so that agents are not burnout.

Shrinkages include breaks, discretionary mins, time-offs, on other auxillary time.

I do at times rethink if i need to plug in occupancy, since it really bumps up my HC requirement especially for accounts with large FTE. Anyways, those are my considerations.

Ah, in that case Michael you are double-counting the avail/idle time.

The Agent formula dictates what the avail/idle time should be. In your one of your earlier replies you show that the agents required in any one hour is 4

Agents(80%,30,(2500/168),500)= 4

The Agents formula though does not consider that the agents will be busy 100% of their time. If we consider that in this hour the agents have to handle 2500 / 168 calls = 15 calls (approx) and that each call takes 500 seconds then we can estimate how much work has to be done in that hour

15*500 = 7,500 seconds of work to be done, that equates to 2.1 hours (7500 / 3600)

The Agents formula tells us that we need 4 agents in that hour all supplying an hour each to do 2.1 hours of work, therefore the occupancy is 52%.

In your overall calculation you inflate the HC requirement by saying that 25% of 75% of the 40 hours that an agent works is lost to availability, therefore you'd need to remove the assumption on occupancy.


As Dave mentioned earlier it would be much more accurate for you to forecast each hour individually to come up with the HC requirements. In doing this you can apply the Agents formula separately to each hour. This means that the formula can take into account when the operation can perform at different occupancy levels.

2500/168 = an average of 15 calls per hour

Agents(80%,30,15,500)= 4 agents, 2.1 hours of work required at a 52% occupancy

if we double the number of calls and get the Agents formula to work out the requirements we find that

Agents(80%,30,30,500)= 7 agents, 4.2 hours of work required at now a 60% occupancy level.

This shows that occupancy varies when the call total changes - greater calls, greater occupancy. It appears that you have noticed this yourself when planning for larger operations as you used a higher occupancy assumption (85%).

I'd advise that you use a hourly call forecast, apply the Agents formula to each hour, sum those totals for the week and uplift the number to take shrinkage into account and then divide by 40 to get you HC total.

Hope this helps.

Eamon

Eamon,

This is a good explaination. I was always considering that the agents erlang formula would result to agents 100% taking calls. I was not able to countercheck the resulting occupancy using the volume and AHT provided.

Thank you very much!
Mike

You can do this quite simply by using a straight forward workload calculation and factoring it up by a calculated wait time (the difference between and average hour erland requirement and workload requirement). Use the Erlang tool to calculate this rather than trying to use it to calculate a weeks requirement.

Also dont forget that Erlang will inherently overstaff... it doesnt take into account the fact that some calls abandon...

It would be not be totally correct to apply erlang calc on avg weekly vols considering its a 24X7 process. You will have to staff to cover your peaks anyways and also to make sure that you have anough coverage during lean period. Doing it on weekly will give you a consolidated figure which when spread across during 24 hours in different shift timings may fall short of requirement to meet SL at 80/30. To add to this your avg occupancy may be 25% for the day however for different periods it may range from 80% to 20%. So I believe its best to staff as per hourly arival pattern than avg weekly volume.

In addition to my previous post, I have added a wait (available) time calculator to my website - give it a go...

http://www.2planconsulting.com/downloads.html