Convergys Point Of View

A POINT OF VIEW PAPER

Service levels

one size does not fit all

Using Science to Set

service levels “A center can achieve its [service level] objectives yet still be wasting resources, creating extra work and providing poor quality.” ~CallCenterMagazine.com

The majority of contact centers manage their operations by adhering to strict parameters around key performance metrics, like speed of answer, commonly referred to as service level. In fact, some argue that service level is the most widely accepted metric in the industry. Debates on the definition of service level and whether there is an industry standard have been raging for years. While some argue for a ‘gold standard’ in setting service level targets, others state that service levels must be established based on specific business needs and customer requirements. So what is the best practice for establishing a service level? Typically, service level targets are determined by contact center management. Despite cautions to the contrary, management teams often apply the old adage “one size fits all.” It is not uncommon for service levels to arbitrarily be set at 80/20 (80% or more of the calls answered in 20 seconds or less). Clearly, the gravitation of contact center management to this “industry standard” illustrates the need for guidance on exactly how to set a relevant— and achievable— service level target.

The solution requires a complex ability to answer myriad questions: • How does time in queue impact the overall customer experience? • Would changing staffing levels substantially impact customer perceptions of their experience? • How can service levels be optimized to better balance labor costs, customer satisfaction and abandon rates?

Sensitivity Analysis: How We Do It Within the call center environment, speed of answer data are potentially available at two levels. The most readily available source is aggregated reporting from the call switch for specific time intervals; for example, Center A’s average speed of answer for the day, or for a half-hour interval within a day. When Center A also has a robust customer survey program, we statistically match service level data with survey data for the same time period.

While the relations among labor costs and service levels can easily be modeled, many centers still overlook a critical clue in unlocking answers to the above questions. Only by incorporating “voice of the customer” measurement will center managers understand how, and when, service level affects the experience in their centers. Without that, we can only guess what customers truly think about their time in queue.

Historically, patterns most often emerge when service levels vary significantly between time intervals within a center. Setting appropriate rules for matching and controlling length of time, our approach reveals variations in how service level impacts the customer experience results. In other words, we statistically isolate the point at which customer satisfaction with the experience drops off.

At Convergys, we have solved this problem through the development of a scientific, data-driven approach to statistically calibrate voice of the customer feedback with service level data. This technique can be used to set any service level, from telephone speed of answer to e-mail speed of response to ideal handle time.

The second level of service level data is at the call center level. Our recommended method for service level sensitivity analysis is to collect speed of answer information for each call as part of an ongoing voice of the customer measurement program. This call-level detail then can be matched to satisfaction data collected from the same customer. While this may not be possible

“Most centers operate as if service level is a goal identical to corporate profit— more is better and less is not,” writes Bill Durr.

2

Sensitivity analysis highlights where all centers might be over-investing. with all switch configurations, the approach yields much stronger insights. We not only collect customer perceptions of the call, but also tie actual process data to that experience in a way that facilitates analysis. Regardless of matching approach, our analyses isolate the effects of varying answer times on levels of customer satisfaction. Through multivariate statistics, we are able to point to whether wait times (and more generally, call handling practices) affect satisfaction levels and if so, to what degree. This quantification of the impact of speed of answer on satisfaction is what most call centers lack. Once we know which levers have the greatest impact on satisfaction, we use sensitivity analysis to understand how customer satisfaction might be affected at a range of wait times. This layer of analysis helps us understand how service levels should be set to drive the maximum blend of satisfaction and cost efficiency. Such analyses have provided significant savings for our clients, by showing that tighter service levels do not necessarily lead to an increase in satisfaction performance.

Service Levels: Rules to Live By While every business is different, we have encountered several common themes with regard to service levels in recent years: 1. What happened on the call is more important than wait time to reach an agent; 2. An 80/20 service level is not a “gold standard” for delivering higher satisfaction; and 3. Over-investment in staffing to meet service levels is more common than you might think.

Rule 1: Call Handling Is More Important. At Convergys, our statistical analyses consistently reveal that agent skills and call resolution have much stronger impacts on satisfaction performance than does time spent in queue. This holds true even with self-reported perceptions of wait time. Across the board, customers whose issues were resolved have higher satisfaction scores—regardless of how long they waited. So when you are challenged to allocate budget to satisfaction improvements, consider placing wait time lower on the list than the above factors.

to pin down. Staffing a center to reach an 80/20 service level inevitably involves labor costs beyond what is needed at non-peak call times. But if it is not important to the customer to answer the call in 10 or 20 seconds, why spend the money to staff to a peak call volume that may only occur a few times a day?

Rule 2: The 80/20 Rule Is No Guarantee. Setting service level to 80/20 does not guarantee a positive customer experience. For example, in the shipping industry we saw that queue times of up to a minute do not detrimentally affect customer satisfaction scores. In the insurance industry, our analyses showed no strong effect on customer satisfaction at wait times of up to two minutes. In financial services, we noted some cases where queue time did not affect customer satisfaction at all.

Return on Investment

As a general statement, queue times of less than 30 seconds do not strongly impact the customer experience. What are the exceptions? Trends vary by customer segment and by wide variations in daily service level performance. In one group of centers where queue time varied substantially day by day, customers were highly sensitive to wait time; perhaps these daily changes resulted in widely varying customer expectations. Similarly, different call types or segments of customers have different sensitivity to speed of answer; customers calling with a technical problem are much more willing to wait in queue than are customers who need to verify their mailing address. Rule 3: Over-Investment in Staffing Is Common. With increased pressures to meet satisfaction metrics, many centers overinvest in staffing to achieve high service levels when the quantified return is difficult

It may be perfectly feasible to drop service level to a 70/30 or a 60/40 average speed of answer, without negatively impacting satisfaction scores. We know cost savings in these decisions can be substantial.

The first step in setting service level targets should be to conduct calibration analyses for each major call type or customer segment. This is particularly true if the goal is to implement a prioritized queuing strategy. When you understand how different wait times affect customer satisfaction scores, you are in a position to make objective, educated changes to service levels. How do you incorporate this knowledge into running the business? We believe that, as with other business decisions, the ROI of various service level combinations must be taken into consideration. It is important to not only avoid the tipping point where the customer experience degrades, but to maximize cost reductions as well. What do we mean? The results of sensitivity calibration uncover where slower speeds of answer cause drops in customer satisfaction; clearly, you should steer clear of that threshold. Just as importantly, calibration analysis highlights where call centers might be overinvesting. If there is no difference between customer satisfaction scores in a 70/20 or 80/20 environment, why pay to staff the incrementally higher service level? We approach this cost reduction / satisfaction benefit evaluation through simulation and business case analysis.

3

Across a range of alternative scenarios, we evaluate the likely achievement in cost reduction or cost increment balanced against the likely impact in customer satisfaction scores. Once our clients have this information in hand, they can make the right decisions about where their service level should be set. The payoff to this type of disciplined approach to service level analysis can be significant to the bottom line. For example, a major shipping provider was concerned about preventing slippage in satisfaction during an upcoming holiday season. To prevent this deterioration, the client decided to tighten service level from 80/20 to 90/20 resulting in significantly increased labor costs. Convergys’ sensitivity analysis predicted changes in satisfaction scores at various service levels and showed there would be no gain in customer satisfaction by moving to the 90/20 model. In fact, the client could loosen service level without adversely impacting satisfaction. Dollars saved from this avoided investment exceeded $1.2 million.

Conclusion Contact center managers can only understand how sensitive customers are to queue time by incorporating their “voice” into operational analyses. Through our experiences, we have drawn the following conclusions about service levels:

• Queue time generally has a small relationship on the experience except at higher than ‘normal’ queue lengths; • Satisfaction with the agent and first contact resolution (FCR) are both stronger drivers than queue time; • Sensitivity to queue time differs by industry and customer type; • Where queue times are less controlled, they may have stronger effects on satisfaction; and • In many cases, loosening service level targets can significantly reduce labor costs without negatively impacting satisfaction or abandonment rates.

Driving the Change

About Convergys Convergys Corporation (NYSE: CVG) is a global leader in relationship management. We provide solutions that drive more value from the relationships our clients have with their customers and employees. Convergys turns these everyday interactions into a source of profit and strategic advantage for our clients. For more information, visit www.convergys.com.

While many ‘rules-of-thumb’ exist as to how to set service levels, there still is no strong set of unified findings into how various service levels drive the top line. It’s always been easy to calculate the cost of increased staffing levels at more stringent service levels, but it’s more difficult to show how high service level targets improve the business. Convergys’ approach to service level sensitivity analysis is one way to understand both sides of the speed of answer question so that call center management can strike the right balance between loyal customers and optimized costs.

• Service level targets should be based, in part, on analysis of customer expectations;

Corporate Headquarters

Regional Headquarters

North America

Europe, Middle East & Africa

Latin America

Asia Pacific

Cincinnati, Ohio USA

Cambridge UK

~ Sao Paulo Brazil

Singapore 049712

Phone +1 513 458 1300

Phone +44 1223 705000

Phone +55 11 5504 6800

Phone +65 6557 2277

US Toll-Free 800 344 3000

Fax +44 1223 705001

Fax +55 11 5504 6730

Fax +65 6557 2727

Fax +1 513 421 8624 www.convergys.com

©2008 Convergys Corporation. All rights reserved. Convergys and the Convergys logo are registered trademarks of Convergys. Convergys refers to Convergys Corporation and its wholly owned subsidiaries.

PV5-048 One Size Does Not Fit 7.11.08