[February 25, 2005]

Shrinkage in Call Center Operations

BY BILL DURR

February being the year’s shortest month, it’s only fitting to talk “shrinkage.” Shrinkage – an odd term – is employee time you pay for but get no work. WFM professionals are keenly aware how important shrinkage is in creating a set of schedules that optimizes agent labor and still meet contact center performance requirements.

Shrinkage is comprised of familiar things like breaks, lunches, meetings and training and of things we don’t think about so often like vacation and paid holidays. There are components of shrinkage that we can schedule with no impact on center operations. And, there are components of shrinkage that are totally unexpected and random like illness, tardiness and “where did they go time” that present serious problems.

My friends at the Call Center School have a PowerPoint slide that neatly annualizes various sources of shrinkage. Using very reasonable assumptions about breaks, lunches, meetings, vacation and holiday and unexplained time, the Call Center School showed that the stark reality is that it’s fairly easy to have a shrinkage factor of 28.8%.

Suppose that we expect to get a steady stream of calls into the center all day long. No spikes or lulls, just a steady metered stream. Using Erlang tables we calculate that we require 34 staffed positions all day long to handle the traffic. I could hire 34 agents and give them each a position. But agents take breaks, lunches and have meetings with coaches. We need to have more agents scheduled than we require and that the difference between these two numbers is directly related to shrinkage. The less time on the phones, the more people I need to schedule in order to always have the required staff.

There are two ways to calculate the financial impact of shrinkage. The more popular formula is:

Scheduled Staff = Required Staff / (1 – Shrinkage factor)

Assume the required staff is 34 and the shrinkage is .288. Using the formula we determine that the scheduled staff needs to be 47.75 agents. Shrinkage costs us the difference in expended labor dollars between the 34 agents we have to have on the phone and our need to schedule 48 agents. Center management teams realize that shrinkage needs to be managed and controlled. A reduction of just 3 points in shrinkage saves two full time agents – a savings approaching $100,000 annually.

You may have already intuited from the list of shrinkage components above that thinking about controlling shrinkage is not as simple as it might seem. A complication is that shrinkage comes in flavors and not all of them are distasteful. We all agree that people need breaks, lunches, training, coaching and even a certain amount of unplanned absence. Then, one flavor of shrinkage has to do with what we expect and can plan for. It is benign with respect to impact on operations. We also know that people will show up for work late, sometimes not at all. They will leave on a break early and come back late. This flavor of shrinkage has to do with what we couldn’t expect and have to react to. It is dangerous and causes management to become reactionary. And, there are a handful of discretionary things like meetings, training (to a degree) and coaching. This flavor of shrinkage has to do with what we wanted to do but may give up to meet service level objectives. This seems relatively benign and even prudent. It is more like high blood pressure – a creeping, silent killer.

There are special flavors of shrinkage associated with agent schedule adherence. If an agent leaves their position early or late, within parameters, they trigger a schedule violation. And when they return to their position they may trigger another schedule violation. The confounding complication is that most of the time we want agents to take their time and conduct a professional interaction rather than worry about leaving their position precisely at their scheduled time. So it turns out that there can be both good and bad schedule adherence violations.

A use-case illustrates the problem and suggests the solution as well. An agent is scheduled to go on break at 10:30 a.m. At one minute before the break time, the agent got connected to a technical support call. The call lasts 10 minutes. The agent logs off for break upset about missing the break time. Caught up emotionally, the agent mis-times the break and returns to the position 22 minutes later. The agent is out of adherence 9 minutes at the beginning of the break and by 7 minutes at the back end of the break. These are two very different behaviors being flagged as schedule violations. We should not penalize the agent for properly conducting the call instead of taking the break. Coming back from a 15-minute break 22 minutes later is a very different situation. We need to be able to track and assess agent time management behaviors granularly and recognize the myriad flavors and varied impacts of shrinkage.

Armed with insight into the sources and flavors of shrinkage, center management teams are better informed in prioritizing operations plans and developing effective programs for improving center performance.

Bill Durr is Principal Solutions Consultant for Witness Systems, provider of workforce management software and services.