1a)How many chat sessions must an agent conduct simultaneously in order to be cost-efficient?
At least three…
1b) What are the recommendations to increase agent's productivity?
A list of standard answers to customer questions in the form of ready-to-go sentences and paragraphs.
2) What are the most commonly uses of chat? e.g. assist customer during ordering process.
Assist Web-based customer service. Moving from self-service attempts to agent-assisted Web-based service.
3) What is the average length of chat?
It may differ by type of chat. Definitely different by type of chat, but in our benchmark, across industries, the average “length” of a chat session is 23 minutes. However, your agent is not bog down by this “seemingly” lengthy process as they are handling other customers in parallel.
4) What is the percentage of chat sessions which are followed-up with a call to the contact center?
Only 14% of chat sessions are follow-up with a call to the center. Chat is proving to be a very efficient and useful channel for both the company and its customers.
Often there is a discrepancy between these two metrics. The most plausible explanation is the “utilization” factor, namely what percent of the time that an agent is in their seat “ready” to handle phone calls, are they actually on the phone.”
Though a delightful concept, this is VERY rare. Instead, in some centers, employees are allowed to select certain other employees that the prefer to have lunch with so that the workforce management system can arrange for this luncheon timing.
Great question…..from our experience with call centers, the target for adherence to assigned schedule is 95%. This key performance indicator (KPI) is mission critical to a call center that wishes to maintain an acceptable service level.
There are large variances in our collected numbers when the groups are less specific in terms of the type of call that they handle. In a Peer Group, this variability is more limited.
Great question. And, yes, there are now companies that supply these virtual reps as a product. These same companies have also done the research you are looking for.
One such company is called "Finali." They are located in Boulder, Colorado. I do not have their coordinates with me as I am on the road at the moment. Check out their Web site for the details.
Good luck.
Dr. Jon
Peter, the ideal/best practice for the “staff to supervisor” ratio in contact centers across all industries is 17 to 1.
--Dr. Jon Anton
Answered 01/17/05
Good question, and yes we do have some best practices benchmarks in the retail industry for you to consider.
Since cost depends on the type of retail store, I would suggest a “unit” cost approach in our comparison:
1. Brick and mortar cost per item sold = 1,000 units
2. Telephone catalog channel cost per item sold = 100 units
3. Web site self-service cost per item sold = 1 unit
Fulfillment of these orders and the items sold is, of course, another story.
--Dr. Jon
Tough question since agent absenteeism can be a major hurdle in delivering acceptable service levels.
If I study our database of over 10,000 call centers, we see an average absenteeism of less than 5 percent, or about 10 days per year. This number does vary quite a bit between industries and regions of the country.
If you have not already done so, I suggest you join our benchmarking community, and enter your performance numbers. This way you can get a more exact comparison with your peer group of call centers.
Absenteeism is a challenge for all call centers.
Dr. Jon
Steve, this is a tough question and would require some intense cuts of our database. We have the details, but I suggest you contact MichaelFeinberg@BenchmarkPortal.com for more details on how to proceed.
But, I can give you a quick answer from our database for the US call center after a definition of the word “utilization”:
Utilization is defined as “the percent of time that an agent is on the telephone with a caller as compared to the total time they are in their seat, connected to the ACD, and ready for a live caller.
”Utilization, across industries, for captive, in-house, in-bound customer service call centers in the US of A:
72% utilization
Utilization, across industries, for outsourced in-bound customer service call centers in the US of A:
87% utilization
Dr. Jon Anton
12/20/2004
Thanks for your question Norm,
The most commonly used formula for calculating call center staffing is called the Erlang-C formula. This formula is used to estimate the number of agents required at an inbound call center for given call volumes and assumes that all incoming calls stay in the system where, if the calls cannot be immediately answered, they enter a queue. Most vendors of ACD systems provide their customers with a software package based on an 'Erlang C' calculator, which will calculate the number of staff required to sustain a particular service level for incoming telephone calls (e.g. 80% of the calls answered in 20 seconds or less) when given the number of calls and average handle time.
If you don’t have an Erlang-C calculator, you can obtain an Erlang-C software calculator on the Internet (just Google Erlang-C) that lets you use Erlang-C painlessly by simply plugging in the variables. Fortunately there are a number of Erlang-C calculators available online to choose from, some of them free.
-Dr. Jon
Answered 12/09/05
FTE stands for Full Time Equivalent.
Contact centers usually have a mix of full time agents and part time agents. A full time agent equals one full time equivalent. A part time agent working 30 hours per week equals 75% of a full time equivalent. A part time agent working 20 hours per week equals 50% of a full time equivalent.
The number of FTEs can not be determined from total paid hours, as total paid hours would eliminate absenteeism, leaves of absence and other non-paid activities.
FTE is a very important thing for contact center managers to understand and know, particularly when building staffing models and budgets. It is one of the data points captured in Purdue's benchmarking survey.
Dave,
We have found a statistical significant relationship between turnover and the supervisor:agent ratio. The relationship is non-linear as indicated by the following:
1. between a ratio of 1:5 and 1:12, there is not much impact on agent turnover.
2. between a ratio of 1:12 and 1:20, there is a gradual increase in agent turnover.
3. between a ratio of 1:20 and above, there is a ever increasing impact on agent turnover.
-Dr. Jon
Great question. Our data shows the following:
1. Calls per shift for a CSR in a top performing telecom call center is 94 calls.
2. Typical call is 3 minutes.
From this you can calculate the "utilization" factor, and this can be impacted by "occupancy." There is a lot to know about each of these important call center management metrics.
Hope this helps.
Dr. Jon
In another recent BenchmarkPortal One-Minute Survey we investigated how call centers handle their seasonal fluctuations in call volume. The results showed that the majority of call centers experience a volume increase that is between 20 and 30 percent above normal. We then asked how the centers handled their seasonal fluctuations. The responses indicated that 38 percent of the respondents have their agents work overtime to handle the seasonal fluctuations. Next in line were "other" and "hire seasonal agents," followed by "convert part time to full time," "outsource," and "bring on at home agents."
We asked for a recommendation of something that they had done that had worked well in handling their seasonal increase in call volume. After analyzing the open-ended responses we saw the following popular responses.
Every missing agent that you had counted on being at work when you did your manpower loading, will decrease SL by 2%.
-Dr. Jon
According to our benchmark research, the optimum span of control is one supervisor to 16 agents. Supervisors need time to monitor, coach, field questions, take escalated calls, track attendance and plan improvements. You might want to enter all your performance data into our Web site, and receive a report detailing your performance as compared to your Peer Group. This would bring out important performance gaps for your new boss to consider.
Human factors design has become more popular as management focuses on the seemingly “little” things that can often make a big difference in productivity. Human factors is an engineering discipline in which the prime thesis is that the “human-to-machine” interface is critical in applying effective technology solutions, and maximizing ROI. The call center agent, or knowledge worker, must use complex technology to consistently deliver excellent service. A few examples of factors that can be researched for design changes include the following: the keyboard (size, shape, layout, etc.), the screen (color, size, height, angle, etc.), screen layout (location of fields, color-coding of fields, pop-up windows, etc.) and many, many more. The human factors specialist focuses on those elements of the human physiology and psychology that can impact performance. Major improvements in agent effectiveness and efficiency have been realized by human factors re-engineering call centers technology.
On average 60% of the operational cost of a call center is tied up in human resources, and just finding great agents that can read, write and speak English articulately is becoming tougher and tougher. Countries that were part of the English Commonwealth, like India, British Guyana, and others are stepping in to take up the slack.
Many citizens of these countries benefit from having an excellent education that includes reading, writing, and speaking English fluently, but they cannot find jobs opportunities due to the lack of industrialization of their countries. With the globalization of telephone networks with the resulting low cost of long distance calls, call centers in these remote areas have become financially quite viable.
Handling e-mail, especially technical support email, is a great way to begin.
On average, 60% of the operational cost of a call center is tied up in human resources, and just finding great agents who can read, write and speak English articulately is becoming tougher and tougher. Countries that were part of the English Commonwealth, like India, British Guyana and others, are stepping in to take up the slack. Many citizens of these countries are receiving an excellent education. With the globalization of telephone networks resulting in lower long-distance costs, call centers in these remote areas have become quite viable financially. Handling e-mail from these countries, especially technical support e-mail, is a great way to begin.
Ritesh, the average number of workstations for an in-house call center in the US is 78. The average for outsourced centers in the US is 250. There are about 100,000 call centers in the US if you include inbound plus outbound centers and internal help desks.
--Dr Jon
In an outbound, tele-sales centre it is much easier to let agents take breaks together. The socialization is important to them and can motivate and energize them when they return. I pick number 2 from my experience.
Dr. Jon
I would begin by asking each CSR to document the "type" of activities required after the call, for example reasons like, a) claims correspondence, b) claims paperwork completion, c) claims investigation, etc.
I would then create a frequency report of these activities to determine the major time "hogs." I would then determine if a "less skilled" (i.e., less expensive) clerk could handle some or all of these activities. I would investigate if there is software available to minimize the agent's time spent on these post call activities.
And finally, I suggest you benchmark your call center's post call work time against a peer group of insurance call centers through the Purdue performance database. By benchmarking, you will better "see" if your amount of post call work time is normal, or is it really too big, and therefore needs your attention.
I am not aware of a specific article that addresses overtime issues in call centers. However, from our work in benchmarking call centers, I would suggest an average budget allotment of 10% to 15% for overtime. Almost all centers encourage overtime for handling unanticipated peak call volumes. Too much overtime may be detrimental, causing burnout and increased turnover.
Most companies in our database report the turnover of full-time agents and part-time agents and part-time agents separately. Mixing them could be confusing.
We have polled call center managers and Web masters, and have averaged their results across industries as follows: a) telephone=$3 to $4 per inquiry, b) e-mail = $15 to $20 per inquiry, c) fax = $25 to $35 per inquiry, d) regular mail = $25 to $35 per inquiry, e) Web site = $0.50 to $1 per inquiry.
This is a very common situation, that is never easy implement without some "pain." In a recent benchmark study on this issue, we found that over 50% of call centers in our database do have desk sharing. Frequently, this is limited to part timers and/or agents that are new to the job. Having "your own" desk is sometime a way to add benefits to keep agents longer.
HR and Resource Planning are both right. It takes strong leadership to convince agents that the "meeting us halfway" is the right thing to do. I think if other HR issues are "on target", i.e., compensation, hours of work, great frontline supervisors, then desk sharing becomes less of an issue.
Dr. Jon
In most situations, scheduled time off is considered discretionary, and an agent can decided that he wants to work right through the scheduled break without being scored negatively by the workforce management system. Of course, the main focus should be on adherence to the scheduled time “on” the phone, since this time “in your seat” is so critical to the overall center’s ability to make service-level goals. So, as long as the agent adheres to his on phone assigned schedule, he can make his adherence to schedule goals. Time on the phone above and beyond this assigned schedule should always be appreciated and recognized by management.
In most call centers, one would schedule for daily business using some type of scheduling tool. Over ten vendors supply such ingenious scheduling tools. Outages are completely unplanned and can not be statistically forecasted. Therefore, standard manpower scheduling tools do not work for such random events. Instead, most call centers have some type of “business recovery” plan in place for business outages caused by weather, fire, theft, and other natural disasters. Usually, this does not include extra staffing, but instead includes contracts with companies that provide business or disaster recovery through the availability of stand-by agents at alternate locations.
-Dr. Jon (answered 05/18/05)
Our benchmark research indicates that the typical team leader would mentor between 15 and 20 people at most. You’re definitely much too lean. The best way to convince your boss that you are not creating a bureaucracy is to benchmark all of your performance metrics with other utilities of the same size (peer group). If you are performing substantially below average compared to your peer group of utilities, you may find it much easier to justify more team leaders.
--Dr. Jon
There is not one single company in our database that reports closing during the lunch hour. Customers need access to information at all times, especially during the business day, when working customers often call during their lunch hours.
Todd,
Home-based agents are a very realistic and exciting new source of low-cost "people power" for today's call, chat, and e-mail channels. I am in the processing of completing a white paper that focuses on enabling physically disabled Americans to work from home. This paper will be available by the end of July and then posted on our Web site.
-Dr. Jon
Answered 7/11/2005
In our database of call centers, the ratio of agents to supervisors that produces the most effective service in the most efficient way is 14 agents to one supervisor. It looks like you have room for improvement.
In our database of call centers, the ratio of agents to supervisors that produces the most effective service in the most efficient way is 14 agents to one supervisor. It looks like you have room for improvement.
When we compare our multi-view call center performance index (MPI) with the size of call centers, we find that the most efficient and effective call centers (i.e., those with the highest MPI), are between 250 and 500 agents. The MPI dips at both the very small call center and the very large call center.
Average calls per loan application is about three to one.
Daily = 96% accurate
Monthly = 83%
We have noticed an ever-increasing interest by contact center managers in best practices related to e-mail management. For that reason, we have launched major initiatives in e-mail benchmarking. The majority of e-mail is responded to in less than 6 hours. This is in spite of the fact that when we survey consumers, they are quite happy to have their e-mail answered in 24 hours, or even more. Regarding e-mail quality, the majority of managers seem to sample the e-mail that was sent and perform their quality check there. We would encourage a greater focus on ensuring that each e-mail is answered in one e-mail, i.e., "first time final," as we all strive to do in our telephone calls.
First of all, you have a wonderful array of enabling technologies to run a world class contact center...congratulations. In answering your question, let’s not confuse adherence to schedule (ATS) with Agent Occupancy, and/or Agent Utilization.
ATS is simply the individual Agent’s adherence to exactly the schedule that you give them for a particular shift. It might include phone time, classroom time, back office time, and the like...ATS for the telephone time is very easy to track with your Aspect system...you are essentially tracking to see that the agent is in their seat at the exact time required by the Aspect WFM scheduler...namely, do they arrive on time, do they take their breaks on time, do they return from breaks on time, and the like...union shop or non-union shop, you should be able to achieve a better than 95% ATS.
Occupancy, by contrast is the percent of time that an Agent is in their seat and connected to the Aspect ACD and ready to handle inbound phone calls (availability), as a percentage of their total time at work (total paid-for shift time). If their only job is telephone work, you should be able to achieve an occupancy percent of 85 to 90%...much more than 90%, and you will begin to see Agent burnout and turnover.
A third key performance indicator (KPI) is called Utilization, (again available from your switch) and it indicates the percent of time an Agent is on the telephone with a customer as a percentage of their being in their seat connected to the ACD (available). This key performance indicator is controlled by proper management in forecasting calls and scheduling Agents. You should strive to achieve 95% utilization of your phone agents.
Hope this helps…
