Business performance assessment and management is a daunting task that managers face on a daily basis. How are my employees performing? Are we improving the customer experience? Where can we improve?
These questions can be tough to answer, especially for a small or medium-sized company that can’t afford business intelligence applications with all of the bells and whistles or the luxury of bringing in a consulting firm to improve their operations.
Luckily, OpsDog is here to help.
This article will describe the basics of analyzing and benchmarking the performance of a Call Center, or Contact Center.
NOTE: The steps detailed below are tailored specifically to Call Centers, but the general methodology can be applied to any business unit or department.
Step 1: Scope Your Project
Don’t you love it when a good plan comes together? Start your assessment efforts off on the right foot by creating a solid project plan before diving head first into the abyss. Outline the areas of focus, identify key stakeholders, outline team member responsibilities and create a day-by-day project plan to keep things moving in the right direction. For a Call Center, some common areas of focus are:
- Call Processing & Issue Resolution
- Call Center Training & Coaching
- Call Center Technical Support
- IVR/VRU Management & Maintenance
- Workforce Management
You will need to take a Venn diagram style approach to improving operations for a Call Center. Each one of the areas above is tied in some way to the others (Check out our Call Center organization chart for some details on these areas.). Operations should be assessed holistically, then you can dive down into more granular sub-functions once you have a data-driven view of Call Center performance.
Other considerations at this step: What locations will I be focusing on (onshore/offshore?)? Which job roles will we be focusing on? Who is my go-to source for data and operational information?
Step 2: Identify & Gather Relevant Metrics
Most Call Centers are relatively data rich environments. If this is the case for you, there should be multiple systems that you can export data from to start putting together your data sources. Produce a list that breaks down which metrics you ‘need to have’ and which metrics are ‘nice to have.’ Take that list to your ‘go-to-guy’ for data and see which systems house each piece of data (IT may need to get involved here, so be ready to explain what data you need to the tech gurus). Base data points (e.g., call volume, total handle time) are more valuable than derived data points (e.g., average handle time, first call resolution rate), but you will have to work with what is available.
Next, figure out how to pull data from the identified systems and get ready to start your data analysis. Lay out a basic framework for the data that you need to get started with your analysis. You should attempt to gather data going back AT LEAST 6 months (preferably 16-36 months) in order to look at some trends. Each of these metrics should be gathered at the most granular time unit available (e.g., hour would be better than day, day would be better than month, etc.):
- Organizational Metrics: data on how many people are staffed for each job role (e.g., agent, manager, technical specialist, etc.) and how the organization is structured (call center headcount ratio, span of control: call centers, etc.)
- Volume Metrics: calls handled, calls offered, calls transferred, volume by call reason/product, total handle time, etc.
- Cost Metrics:overhead expense, personnel expense, total expense, etc.
- Quality Metrics: occupancy rate, quality score, IVR containment rate, etc.
- Service Metrics: first call resolution rate, abandonment rate, hold time, customer satisfaction and survey results, etc.
- Productivity Metrics: how much work each individual employee is producing (average handle time, average talk time, occupancy rate, etc.).
- Revenue Metrics: total revenue generated through call center channel, revenue per call center representative, total revenue: IVR/VRU system, etc.
In addition, it might be helpful to listen to some calls (recording or live) to produce some anecdotes that can back up your data. For example, if you find that first call resolution rate is very low, you can refer to some specific calls and say: “Hey, your first call resolution rate sucks — but it is obvious why that is, listen to this call.” Ok, don’t say that exactly, but you get the gist. Examples like this help to tell a story and add to the significance of your data analysis. You can also listen to a sample of recorded calls to gather some information if other available data is sparse.
Once you have collected some, or all, of this data you are ready to get your hands dirty in Excel, or in your data manipulation tool of choice. Getting ALL of this data is probably a pipe dream, so try to collect everything you can and get rolling as soon as possible. If you wait until 100% of the data that you want is gathered — well, just don’t do that. Your project will struggle to get off of the ground.
Step 3: Analyze Your Data
If you have your base data collected, you can start doing some simple math, running trends and checking correlation for different metrics to start producing some insights.
Here are some good examples:
- Calls Handled per Call Center Representatives: How many calls is each representative handling per hour/day/week/month? Which representatives are performing poorly in this regard? What are the top performers doing differently?
- Abandonment Rate: Does customer service suffer when call volumes spike? This could be an indication of poor staffing models. Excellent call forecasting and staffing is a pillar for any Call Center worth its weight in salt.
- Call Volume by Reason for Call: Are customers calling in for reasons that could be dealt with by looking at an FAQ on the company website? Reducing inbound call volume by taking care of eliminable calls before they happen can reduce stress on a Call Center.
- Occupancy Rate: How much time are representatives actually spending on the phone? You need to find a happy medium here. 70-80% is a good sweet spot. Anything less, and time is being wasted; anything more and representatives might lose their minds. Compare representative occupancy to average handle time and call volumes to look at the correlation.
- Cost per Call: How much is each call costing, on average? How much of that cost is devoted to personnel? Overhead and technology costs? What does the trend look like? The majority of Call Center expense should be devoted to salaries and benefits.
Keep charting and analyzing until you produce a set of solid, actionable insights. Prepare an executive summary for key stakeholders and present your findings. Again, it is helpful to supplement your data with anecdotes to tell a great story.
Step 4: Identify Areas for Improvement
What did the data tell you? How are you performing compared to: a) your own best/worst demonstrated performance levels, and b) your peers and competitors? At this point it may be helpful to purchase a benchmarking report to get some external data that can be used for comparison.
Are you lagging in the Calls Handled per Agent KPI? Then see what top performers are doing differently and roll out their ‘best practices’ to the rest of the staff.
Is abandonment rate spiking when call volume peaks? Figure out when call volume is the highest based on historical levels and staff accordingly. Staffing models should be very granular — thirty minute intervals.
Is average handle time higher than you would like? Then observe some calls and see how you can eliminate low-value-added time within a call. Is it taking too long to validate customers? Are agents putting callers on hold for excessive amounts of time?
These are just a couple examples of improvement opportunities that can be identified through performing your data analysis in the fashion laid out the steps described above.
- Scope Your Project
- Identify & Gather Relevant Metrics
- Analyze Your Data
- Identify Areas for Improvement
Once you produce these kinds of data-driven insights, you can work with management to outline a business case and focus on high-value areas of improvement. Many times, you can pick out ‘low-hanging fruit’ improvements that can be addressed in the near-term (within 1 month, or less). These improvements should be prioritized, while long-term objectives are being addressed in the background.
Although data alone is not usually enough to change the way that people work, it is a great first step towards improving operations for any business.