Tag: managing call center agents

By Peter Lyle DeHaan, PhD

In retail, the term shrinkage is euphemistically used to reference stock which “disappears” before it can be sold. In essence it is a product that the retailer bought, but can’t sell. To be direct, shrinkage is theft. While some of this occurs in the form of shoplifting, it also results from employees, both through acts of commission and acts of omission.

Regardless of the source or the motives, shrinkage hurts everyone in the form of higher consumer prices and lower company profits. This affects jobs and threatens the business’s future viability. Some retail operations take a surprisingly relaxed position about shrinkage, viewing it as an inevitable cost of doing business; whereas others see it as the theft that it is, taking aggressive steps to eliminate or at least reduce it.

Shrinkage in the retail environment has an analogous application to the call center. True, a call center does not have tangible inventory that can disappear. A call center’s inventory is human capital, that is, the call center schedule. Shrinkage in a call center, therefore, is agents who are “on the clock” but who aren’t processing calls. This could be manifested by agents who are not at their stations when they are supposed to be, not being logged in, not being “in rotation”, or who employ some “trick” to block calls.

Similarly to retail, some call centers take a surprisingly relaxed position about this shrinkage of the schedule, also viewing it as an inevitable cost of doing business. Their response to it is intentional over-staffing. This only serves to cover the problem, not resolve the underlying cause. Other call enters see shrinkage as little more than stealing – stealing time. Like their retail counterparts, they too take aggressive steps to eliminate or at least reduce it. Call center shrinkage likewise hurts everyone: a lower service level offered to the client, reduced profits to the call center, decreased morale, and even less compensation for the call center agents.

There are three factors that help track, explain, and counter call center shrinkage. They are adherence, availability, and occupancy.

Adherence: Adherence is a measurement of the time agents are scheduled to work compared to the time they actually work. Why is adherence important? Quite simply, it is because the schedule was developed to match the traffic projection and when the schedule is not fully worked, the result is understaffing. In an ideal situation, staff should adhere 100% to their schedules. Unfortunately, this is not the case.

Adherence can be best tracked by comparing logged in time to scheduled time. Most call center managers are shocked the first time they look at this. It can represent a huge unnecessary cost to the call center, as well as contribute to lower service levels.

Several factors can account for differences between the schedule and the time worked. The first area is scheduled breaks, lunches, and training. This is the only acceptable contributor to adherence discrepancy. Depending on the length of breaks, the best resulting adherence will be around 90%. Forty-five minutes of breaks in an eight-hour shift will result in an adherence of 90.6 % (7.25 hours / 8 hours). The second consideration is absences, late arrivals, and early departures. Unless these openings are filled, the result is a disparity between the schedule and the fulfillment of that schedule. If this missed work is paid time off, such as paid sick time, then there is both a dollar cost and service impact that results. The third area is unscheduled breaks or any other distraction that causes agents to leave their positions. When factoring all of these items together, it is not uncommon for call centers to have adherence rates around 75%, although well-run centers will be in the low 90s (as determined by their established break schedule.)

Adherence is the first of three related scheduling metrics. The next is availability.

Availability: A second, and related, staffing metric is availability. Availability is a subset of adherence. Of the time that staff is adhering to their schedule, availability measures how much of that time they are ready (that is, available) to answer calls. It can be easily calculated by comparing available time (also called, “on time,” “in rotation,” or “ready”) to logged in time. Specifically, it is the percentage that results from dividing available time by logged in time. Although the ideal goal of 100% availability is achievable (that is, ready to process calls all of the time agents are logged in), 98% to 99% is more realistic.

Agent availability is strictly within the control of agents. It is determined by each agent’s willingness to keep his or her station in a state of readiness to be assigned calls. Simply put, it is whether the agent is available to take calls all of the time.

Availability is the second scheduling related metric. The third is occupancy.

Occupancy: Occupancy is the amount of time agents spend talking to callers compared to the time they are turned on or are available. Although it is possible to have 100% occupancy, the corresponding service level would be poor and generally unacceptable. One hundred percent occupancy means that agents are talking to callers the entire time they are logged in. It also means that there are calls continuously in queue, waiting to be assigned as soon as an agent completes a call. The resulting efficiency is great, but callers can end up waiting in queue for several minutes. Therefore, 100% occupancy does not produce quality service and can lead to agent burnout and fatigue.

Interestingly, ideal occupancy rates vary greatly with the size of the call center. Smaller centers can only achieve a low occupancy rate (perhaps around 25%) while maintaining an acceptable service level. Conversely, large call centers can realize a much higher occupancy rate (90% and higher) and still maintain that same service level. This dynamic relationship between occupancy rates and call center size is the underlying impetus for mergers and acquisitions among outsource call centers; it is a profound example of economies of scale. Call centers in the 10 to 20 seat range typically see occupancy rates around 50%.

To calculate occupancy, divide the total agent time (that is, talk time plus wrap-up time) by agent “on” time. This should be determined for each agent as well as for the entire call center.

Two Case Studies: Now, let’s consider all three of these metrics together and apply them to two call centers. The first, a well run call center and the second, an under-managed one. We will assume that they are the same size and both have a realized occupancy rate of 50%.

Call Center A has an adherence rate of 90% and an availability rate of 95% (along with the aforementioned 50% occupancy rate. For each 8 hour shift there is 3.42 hours of on-line time or actual work (8 hours x 90% x 95% x 50%).

Call Center B has an adherence rate of 75% and an available rate of 65% (with an occupancy rate of 50%). For each 8 hour shift there is only 1.8 hours of on-line time or actual work (8 hours x 75% x 60% x 50%).

Although the results for call center A, a well run operation, may be surprising, the corresponding number for call center B is shocking. In fact, to maintain the same service level, Call Center B would need to schedule almost twice (1.9 times) as many hours as Call Center A. Consider what a significant impact this would have on the bottom line.

Lest you think that these are unrealistic numbers, both are real situations describing call centers I have visited. It takes a concerted and ongoing management effort to be like Call Center A, while all too many operations are more like Call Center B. I challenge you to run your numbers to see how you compare – and then take steps to improve them.

Don’t let call center shrinkage lead to profitability shrinkage.

Read more in Peter’s Sticky Series books: Sticky Leadership and Management, Sticky Sales and Marketing, and Sticky Customer Service featuring his compelling story-driven insights and tips.

Peter Lyle DeHaan, PhD, is the publisher and editor-in-chief of Connections Magazine, covering the call center teleservices industry. Read his latest book, Healthcare Call Center Essentials.

By Peter Lyle DeHaan, PhD

For years, starting when the teleservices industry was in its infancy, call center managers have likely wish that their staff could work from remote locations.

The reasons for this vary, from tapping into a different labor market to providing a local presence and gaining greater efficiency by tying together multiple call centers.

As technology advances and becomes less costly, and as more options emerge from telephone carriers and the Internet, it has become possible to have a reliable, remotely located agent station.

The most common use for remote agent stations is connecting multiple locations. Aaron Boatin, vice president of Ambs Message Center, has connected three offices together using dedicated T1 circuits. A Startel 5700, located in Jackson, Mich., handles all call traffic and client information for the three locations. LAN traffic is TCP/IP and sent over one channel of each T1.

Agent audio is compressed using Adtran multiplexers for one location, while the other site uses a separate T1 channel for each station’s audio path. The company’s main office has ten stations, as does one of its satellite offices, with the third location having five positions.

Each location taps into a separate labor market and all agents are cross-trained on all accounts.

Medcom, in Columbus, S.C., also connects three offices. While Ambs Message Center’s three offices are all in the same state, Medcom’s locations are in three different states.

Robbie Parnell, system developer, said Medcom also cross-trains its agents on all accounts. Medcom uses a Telescan Earthnet system for its call processing and a dedicated T1 circuit, with Adtran channel banks, to connect its offices.

“The only problem,” Parnell said, “has been the backhoes” – which have twice cut the fiber optic cable carrying one of his T1 circuits. He advises knowing the path that a dedicated T1 will take and using that as part of the evaluation criteria, as opposed to price alone. He is currently looking for an alternate provider for the problematic T1 circuit.

Michigan Message Center has been using remote agent stations to connect offices together since 1989 and added home-based agents a few years ago. Its primary CTI platform, Amtelco’s Infinity v5.1, lends itself to both home-based agents and remote office installations.

Kurt VanderScheer, vice president of engineering, said, “The technology employed for remote offices was primarily determined by the desire to keep the agent experience at remote offices as close as possible as to that of the host office.”

To attain this goal, offices are interconnected via point-to-point T-1 circuits. Network data is routed as a typical dedicated WAN (wide area network) using Cisco routers.

Agent audio is a dedicated connection that rides on individual channels on the T-1 and terminates directly on the Infinity ACD. Remote maintenance is easily accomplished with VNC (virtual network computing) desktop sharing.

Michigan Message Center’s home-based agents connect to the office network via the Internet over an encrypted VPN (virtual private network) tunnel.

“To be effective, home agents are required to obtain high-speed Internet service meeting minimum bandwidth requirements of 256 Kbps,” VanderScheer said.

For audio, home-based agents use a standard phone line and simply dial into an Infinity account configured for agent audio. For security, VanderScheer requires home-based agents to have their computers protected with anti-virus software and a properly configured firewall.

MedCom Professional Services, Inc. in Levittown, Pa., also uses remote stations to permit agents to work from home. Chris Bell, president of MedCom Professional Services, indicated that remote agents have allowed him to implement “just-in-time” staffing.

Although multiple means are available to connect to the Internet for the data portion of the remote station, MedCom’s vice president, Tom Sheridon, is opposed to dial-up because of speed and reliability problems.

Instead he prefers DSL, which provides good bandwidth, to extend the network to the remote station. Some remote agents connect using cable modems, although they have experienced unacceptable service interruptions.

At the call center they have a fractional T1 with a dedicated IP address and firewall. Network address translations (NAT) use port mapping through the firewall to provide for highly secure connections.

So far Sheridan has used the Internet only for CSR (customer service representative) data screens and for messaging. The voice path is connected via a preprogrammed DID number into the company’s Startel 5700 switch.

He reports that this method has had no significant audio problems. “Even though we could layer in voice over IP (VoIP), standard phone lines are cheap and reliable. When managing remote agents, simple works best,” he concluded.

“One agent,” Sheridon noted, “who relocated 600 miles away, uses an unlimited long distance plan” for her audio connection. He believes these unlimited long-distance plans will become more common in the future, meaning that MedCom will never have to lose an agent to relocation.

MASCO Services Inc., in Boston, has two remote agent stations off its Avaya switch; they were added to support an initiative to reduce traffic in congested Boston. According to Gary DuPont, director of telecommunications, one remote station uses the Teltone Office Link product.

It is a flexible, affordable solution that allows a remote agent to dial up and log into their Avaya ACD. Once logged in, the agent uses a 2500 compatible telephone with 10 to 20 programmable speed dial buttons for ACD access codes.

A call whisper feature on the ACD identifies the call type to the agents. The client database is available via high-speed Internet connection and VPN.  

Their other station uses the Avaya Definity Xtender allowing remote agents to dial up and log into the ACD over standard analog lines, using Avaya proprietary digital ACD telephone sets, which interface with the PC. Again, high-speed Internet access is required and as well as a VPN.

Betsy Petty, owner of Always In Touch in Rapid City, S.D., also employs a home-based agent. In her case, however, it wasn’t to retain an existing agent, but rather for a new hire, who was employed as a home-based worker.

After a few weeks of AccuCall training in the call center, Petty felt that her new hire was ready to begin working at home. The remote station was “very easy to set up, with a little bit of assistance from CadCom,” Petty said, adding, “It only took about 15 minutes.”

Petty also has a remote position in her home, which she uses for administration, programming, monitoring, and answering calls when needed.

She has been using remote agent stations for about a year. In her configuration, the remotes connect to the AccuCall system through the Internet for the data component and use dial-up for the audio. The remote stations are fully functional, including the AccuCall voice logger.

Kevin Bachelder, director of IT for Ansaphone, Inc. in Quincy, Mass., has implemented remote agent technology for his company’s Alston Tascom Evolution system.

Currently, several members of the Ansaphone management team are able to take calls from their homes during unplanned traffic peaks or unexpected call volumes, such as during snowstorms.

Their Tascom digital phone switch allows them to use any outside telephone number as a call taking position and then the end-user connects to a PC in the office using software, such as PC Anywhere, for the data portion of the call. Bachelder is currently doing some internal testing of the Tascom software to get it running well under Citrix, which “is a way of delivering a remote desktop without having to install remote control software or needing a high-end PC on the receiving end.”

Because this kind of “connectivity can be delivered through a Web browser it makes it a very easy and fast way to have an outside agent or agents, because all they need is a phone and a basic PC with decent speed Internet access such as DSL,” Bachelder added.

John Detrich has also used the Citrix server for connecting remote agent stations. His implementation was with an Amtelco Infinity telemessaging system and PI 2000 order-taking system from Professional Teledata.

“An advantage of using Citrix,” Detrich said, “was that if the connection is lost, the agent can log back in within a couple of seconds and continue on the call where it left off.” A disadvantage of Citrix, he said, was the cost of equipment and software.

Detrich mentioned that a second method of provisioning remote agent stations is to use VPN (virtual private networks) and go through a firewall. A clear advantage of VPN is that it is highly reliable.

However, when a VPN connection is lost, all accounts assigned to that agent are then transferred to other agents or sent back into the system queue. Another disadvantage with VPN, Detrich added, “is that it was only rock solid when used with Windows XP.”

At the remote location, Detrich had agents use high-speed connections, such as DSL, cable, or dish. He found DSL to be the most stable; the dish was the second most stable (although only one person used it). The last choice was cable modems; outages of four to six hours were not uncommon.

None of the preceding call centers are using VoIP for agent audio, although several locations are considering it or watching technology developments for future deployments.

(With VoIP, all that is needed is a stable Internet connection at each end; audio signals are sent from one location to the other over the Internet. As such, VoIP eliminates dial-up audio connections and additional phone lines, and has no usage charges.)

Joe Miller, president of Checkpoint Communications Co. in Greenville, N.C., has been successfully using VoIP for more than a year. In his implementation he has two remotely located call centers. At each one, he has a fractional T1 circuit installed and connected to the Internet.

This Internet connection handles not only the data for the stations, but the agent audio as well, along with incoming DID traffic and outgoing calls. Connected to each fractional T1, Miller has a Tenor VoIP MultiPath Switch from Quintum Technologies installed.

At the main call center, data is split out via a standard network port and connected to a hub on his Amtelco Infinity system. Agent audio, incoming DID, and dial-out lines are each connected to their respective ports on Infinity. Corresponding connections are made at the remote office.

The network port is connected to a hub on the local network for the agent stations. The agent audio goes to headset boxes at each station, while the DID and dialout ports are connected to the DID trunks and phone lines provided by the phone company.

Miller tested the service for close to a year before running serious traffic through it. Cable modems and various versions of DSL did not produce the stability and audio quality that he needed. Eventually he migrated to the more reliable but more costly fractional T1.

“VoIP doesn’t require much bandwidth, but it does need to always be there,” he said. He is also quick to stress that when VoIP is being used, all non-call center traffic (such as email and Web access) must be routed through an alternate Internet connection in order to maintain optimum audio quality.

These are the keys to a successful VoIP implementation.

Miller is sold on the quality and reliability of his VoIP service and the Quintum switch behind it. The Quintum switch provides for a dial-up back up in the event that there is a problem with the T1, but his company has used it infrequently.

In his current configuration, Miller runs four remote agent stations and has the capacity to go to eight. The implementation is scalable, so he can go beyond eight by adding more bandwidth and expanding his Quintum units.

When asked about the audio quality, Miller says it is as good as or better than normal phone conversion and that “no one has complained about quality.”

What Miller has built using Infinity and Quintum will be “a big benefit for acquisitions” as a company can retain the staff at the acquired call center, but still obtain economies of scale – regardless of where the call center may be located.

This is “very, very big for tying offices together,” Miller concluded, and may be the wave of the future.

Using the Internet for Remote Agent Stations

For sites considering remote agent capability using the Internet for the station network connection, Tom Sheridon offers the following advice:

• You should have a static IP address for the call center’s Internet access. Otherwise, your remotes will have difficulty connecting with your system. There are some ways to get around this (that go beyond the scope of this article), but having your own address is the simplest approach.
• Next, get the technical side working reliably.
• Automate as much of the connection process as possible.
• Thoroughly test and then train a couple of key people. (You can even set up a “remote” position in the call center and use it there until you have all of the bugs worked out.)

The goals are to make your remote position work as well as it does in the call center and to guard against unauthorized access to your system. There is no single approach to accomplish these things and it will take some time and effort for you to figure out the best approach for your company.

Read more in Peter’s Sticky Series books: Sticky Leadership and Management, Sticky Sales and Marketing, and Sticky Customer Service featuring his compelling story-driven insights and tips.

Peter Lyle DeHaan, PhD, is the publisher and editor-in-chief of Connections Magazine, covering the call center teleservices industry. Read his latest book, Healthcare Call Center Essentials.