Delivering target service levels by Interval can be very difficult as it's suspectible to flucuations that you don't always have control over. The best you can really hope for is an excellent adherance team that can quickly move around resources as required.

But by the time you realize that you are experiencing a call spike and need to react it is often too late and you've lost that Interval.

To not incur additional costs for staffing you'd need to be able to exactly predict call volumes. Even then you'd need a flexible enough staff (part time/split shifts, etc) that would allow you to staff appropriately during those intervals.

Not a real easy task.

I've seen one company create a "rule" that if their predicted call volume was over 10% then that interval was an "exception" and they wouldn't total or count it. Their thinking was that a spike of over 10% wasn't predictable so shouldn't reflect poorly upon them. Not exactly customer-centric though (nor something I agreed with) as I felt it really took the pressure of them trying to be accurate.

If you really want to guarntee a service level by interval you need additional staffing which would drive up your cost.

Service level achievement by interval is achievable through a variety of methods but does service level achievement mean custmoer satisfaction? I say this because there are numerous options for automated caller responses such as optional IVR, queueBuster etc which will aid service delivery. Without automation it is still achievable but needs a greater degree of queue monitoring and intervention from trained specialists, this however can only be achieved if there are options for the intrday guys to deploy such as back office support, call blending, multi skiliing etc, all of which should be used in moderation and tactically to avoid CPOP.

As Justin says, achieving target levels sustainably is difficult when there are rapid changes in call volumes. Even if you have a 'specialist' monitoring the call traffic, primed ready to sound the alarm if they spot a 'spike'.

To have someone waiting to react to a call spike is expensive, impractical and disruptive to staff.

It is far better to employ an automated system such as QBypass to smooth out peaks and spread the load out.

You do not have to have people sat around waiting for a spike, this can be done through blending and/or multi skilling.
Employing any automated queue handling system can cause more problems than it solves and, In my opinion, is no replacement for a warm voice. Customers want dealing with at a time that suits them not the organisation they are contacting.

Depending on the service level you are aiming to achieve, and the period you measure your service over, you may also decide not to fully staff for a pronounced peak.

Personally, when I call a contact centre, I would prefer to have a call unanswered than to be promised a call back at a time that may not suit me. I would prefer to go unanswered and then I can call back at a time that suits me.