Balance service level every ½ hour ?

[davidhb]

Is it possible or practical to balance your service level every ½ hour of the day? Our goal is 70% in 30 on a daily basis. We have been attaining this goal for sometime. Now we are looking at each ½ hour. In a typical day there are wide swings in the goal from a 40 to 90 depending on the day. Is this ok or do we need to be concerned and measure it each ½ hour?

Thank you in advanceÂ….

[beejay]

Absolutely, David.

Not sure what you mean by "balance" but you should track the service level every half hour. Your statistics should include:

NCO
NCH
Abandon
Abandon %
Service level %
ASA
AHT
# agents avail

And you should match it against your call forecast. These operational stats will tell you what is happening in your center on a near real time basis and also give you an indication of what's contributing to the peaks and valleys in your service level. It could be the call array is different than expected, staffing/shifts need to be tweaked, or the floor needs to do something differently.

Keep these statistics all day every day. From an historical perspective, you will be able to better utilize your teams, expect the call spikes and understand what is happening in the operation.

Hope this helps............................

[JoeB]

Ahhhh, another proof that when you average large enough groups you can pretty much hide anything... As BJ suggests, measuring key performance indicators frequently will help you understand the peaks and valleys in performance, which you can then trace to root causes and adjust accordingly. You will find that these causes may come from the natural distribution of calls (if you like statistics, the distribution of call lengths is usually log-normal, and the arrival rate is Poisson, although often there are multiple peaks, reflecting the callers work schedule), or from internal scheduling, differences between individual agents, etc. Further to BJ's point about developing a historical perspective and effecting staff/shift changes, measuring variances (instead of averages) will give you an added perspective that will help in your decision-making.

BTW, just as an interesting side point, when you sample data every ½ hour, your accuracy is really of one hour, because statistically your sampling rate may coincide with significant events in the sampled data (this is known as Nyquist's Law)

Joe Barkai
DIAGNOSTIC STRATEGIES
http://www.DiagnosticStrategies.com/
Tel. 781-433-0833

[beejay]

Nyquist Law, eh? And the good Jesuit priests doubted I would ever understand business math and statistics (heheh)

You are certainly correct about the accuracy really being an hour.

Just an aditional note to my list of stats, I forgot to include ACW as part of the measures. I utilize it and include it in the stack but forgot to put it in the list I provided David. What a goof.

Thanks