Rca Diamond

A Root Cause Analysis Strategy for Improving Customer Satisfaction “What Can Jethro Tull Teach Us About Customer Service?"

Introduction In the largely agrarian 17th century economy, farmers struggled to improve centuries-old inefficient processes. Traditionally, seeds were thrown, or "broadcast," which made it difficult to weed and harvest the crop. Finally, farmers used a "dibber," a board with holes evenly spread apart, for planting crops. A stick would be pushed through the holes, and a seed placed in the hole made by the stick. This was effective and targeted but also tedious and time-consuming. It was during this time that a farmer named Jethro Tull (in addition to inspiring flute-playing rock-androllers everywhere), invented the seed drill using parts from the foot pedals of his local church organ. Tull's drill sowed seeds in uniform distance and the runners on the drill made holes that allowed the seeds to be sowed at uniform depths. It also enabled a consistent, precise method for harvesting since the seeds would grow in predictable patterns. Ultimately, his invention enabled mechanized farming and is recognized as one of the driving forces of the industrial revolution.

In many ways, customer service operations today are riddled with the same inefficiencies that were faced by 17th century farmers. This paper examines the successful, innovative and cost-effective solution one customer-centric company deployed to identify and address the causes of customer dissatisfaction.

The discussion includes: I.

Need for Strategic Root Cause Analysis in Customer Service

II. Challenges of Strategic Root Cause Analysis III. Identifying Drivers of Dissatisfaction – Crafting the Solution IV. Benefits V. Can Your Company Find the Root Causes of Customer Dissatisfaction?

© 2004 DiamondCluster International, Inc. All rights reserved.

This Viewpoint was prepared by Steve Rudolph, Amaresh Tripathy, and Michel DiCapua. Contact:

Mark Keeley Managing Partner, Telecom & High Tech Practice Strategy & Business Analytics 312.255.5642 [email protected]

OR

Steve Rudolph Principal, Telecom & High Tech Practice Strategy & Business Analytics 312.268.3752 [email protected]

viewpoint

Volume 4 Number 1

!

Why are customers dissatisfied?

!

Why are customers calling care?

!

Why do customers defect?

!

And most importantly: What can we do to increase customer satisfaction cost-effectively?

Just as Tull's seed drill mated the range afforded by the broadcasting process with the targeted precision offered by the dibber, answering customer service questions requires a strategic approach that is at once wide-ranging and targeted. This paper outlines a root cause analysis strategy developed by DiamondCluster in collaboration with a national telecommunications carrier to address customer service issues.

1

Root cause analysis reveals the underlying causes behind customer dissatisfaction and leads to targeted actions to resolve them. It is based on the understanding that strategic, rather than merely operational, methods are required if companies aim to be serious about solving the challenges of customer satisfaction. The strategy follows the design and implementation of "code-based intervention." Code-based intervention enables targeted call monitoring and analysis by employing a coding schema for inbound calls that is then matched to actual call recordings. In this way the carrier can hypothesize about root causes and subsequently validate, quantify, and tactically respond to these problems. When combined with existing primary and secondary research tactics, code-based intervention greatly increases a company’s confidence in decisions that have far-reaching impacts on customer satisfaction, retention, and the cost of service. Implemented for less than $1million, the solution has allowed the carrier to identify over $60 million in variable cost savings within the first 6 months after implementation.

Krass, Peter, "CRM: Once More, Without Reeling," CFO Magazine, March 17, 2003. Citing AMR Research, Krass notes

the following percentages of companies achieving various levels of success with CRM implementations: 12% - Failure: started but failed to go live; 47% - Implemented: went live, succeeded in the technology aspects, but business change and adoption failed; 25% - Adopted: Succeeded in both adoption and systems, but could not quantify business benefit; 16% - Improved performance: Reached the promised land. It measurably improved business performance.

2

viewpoint

Despite all of the money spent on customer satisfaction and CRM systems, many companies are no closer than they were 20 years ago in understanding the specific drivers of customer dissatisfaction. With only 16% of companies stating that CRM implementations result in "measurably improved business 1 performance," the most basic questions still remain unanswered:

I “Companies in many industries need to be able to identify drivers of dissatisfaction in order to reduce costs, increase customer satisfaction, and prevent defection.”

Need for Strategic Root Cause Analysis in Customer Service Conventional approaches to customer dissatisfaction have encountered the weaknesses of either the broadcasting technique or the dibber. Traditionally, companies have used primary research tools such as customer surveys and focus groups to understand customer dissatisfaction. They rely on third-party firms who are divorced from the day-to-day business operations. A broad analysis such as a closed-ended customer satisfaction survey may capture general feedback: customers are disgruntled with the company's customer service; or customers primarily choose products based on price; or the company suffers from an overlyconservative brand. But these types of results are often already obvious or are not specific enough to be actionable. Additionally, selfselection or other biases to which surveys and focus groups are prone may skew the results. Similar to broadcasting, there is no efficient and clear way of harvesting actionable information from this process. On the other hand, more rigorous hands-on analysis, such as live monitoring of inbound calls, suffers from the same issues as the dibber. It is very time-consuming and expensive to make the process effective and actionable. A small sample of calls may give the listener insightful detail into particular customer concerns but it may not be representative of the customer population as

3

a whole. Conversely, the call may be representative of the customer population but due to the broad range of issues covered in the calls monitored, may not be sufficiently detailed or actionable. As the shortcomings of these methodologies are exposed, the need for a strategic approach to customer service becomes even more important. Barriers to switching between products or service providers are constantly falling. The growth of the Internet has empowered customers with access to a wider range of consumer choices in an increasingly competitive business environment. Customers are becoming savvier in their choices and would not give a second thought to switching loyalties when faced with a situation where their needs are not consistently met. These current realities, coupled with regulatory changes, such as number portability in telecommunications and HIPAA in the healthcare industry, make dissatisfied customers even more likely to defect. In short, companies in many industries need to be able to identify drivers of dissatisfaction in order to reduce costs, increase customer satisfaction, and prevent defection (churn). Moreover, companies need to strike a careful balance between cost reduction and the potential loss of customer satisfaction. In most cases, this is easier said than done.

CASE STUDY Keeping Customer Service a Key Competitive Lever A wireless telecommunications operator was facing high call volume in customer service operations and declining customer satisfaction. The average call per subscriber was above the industry benchmark. With the cost between $3-$5 per call, customer service operations were a significant portion of their general and administrative expense. Realizing that customer service is a key competitive lever in an industry with declining margins and high churn, the operator decided to elevate their customer service operations from a tactical to a strategic level. The carrier charted a three-step process: 1) Creating a coding schema for inbound calls, allowing the operator to categorize the root cause prompting a customer call and to obtain a "radar view" of the issues; 2) Developing a statistically significant tentative causal model for the relationship between coded events and their value (i.e. hypothesize and prioritize); 3) Intervening when the causal relationship is established ("deep-dive" analysis) – and creating strategies/tactics to address and resolve the issues. The solution integrated statistical process analysis with off-theshelf technology already in use by the call centers. It also required staffing a specialized root cause analysis team to support the ongoing process. The solution allowed the carrier to conduct accurate, actionable, cost-effective root cause analysis for the first time. This paper explores the challenges presented by strategic root cause analysis, how the solution created by the organization met their challenges, the benefits achieved, the obstacles that were faced in implementation and execution, and the applicability of this solution to identifying drivers of dissatisfaction in other industries.

4

II

Challenges of Strategic Root Cause Analysis Do you ever wonder why you often hear the message, "Your call may be monitored for quality assurance purposes," but you seldom hear, "This call may be monitored for the purpose of improving customer satisfaction?" Most likely, the reason is that many companies perform primary analysis for purely operational reasons. The carrier in the case study (p.4), for example, was already performing call monitoring on a regular basis. Nonetheless, when it did so, the monitoring was typically targeted at the agent level,

enabling call center managers to grade their specialists. Call monitoring for root cause analysis purposes, on the other hand, represents a strategic version of primary analysis. Indeed, many companies are not accustomed to leveraging a tool such as call monitoring for strategic purposes such as understanding and improving customer satisfaction.

satisfaction, it is extremely difficult to validate and quantify the extent of customer satisfaction issues, forcing most companies to jump from hypothesis generation immediately to implementation. Often, problems are identified through anecdotal information – the "problem of the moment" that catches the attention of a customer service supervisor or department VP.

Yet even for those companies that do use primary analysis to address customer

Strategic Root Cause Analysis Overview Analyze impact of results to determine future improvements

Objective

! Summary of available data; determine level of statistical accuracy and actionability

! Current data sources and collection methods require revision to ensure accuracy and actionability

! Generate hypotheses as to the underlying causes of customer interactions based on available data sources

! Brief articulation of several key hypotheses that need to be validated or disproved

! Hypotheses often based on incomplete or anecdotal evidence (e.g. Executive complaints, one-off call monitoring)

C

! Validate hypotheses through direct observation (call monitoring, market surveys) and statistical analysis

! Quantification of detailed reasons for root causes of customer interactions

! Resources and methodology not available to perform “deep dive” of strategic root causes of dissatisfaction

D

! Identify the “quick hits” and long-term changes required to optimize the targeted customer interaction

! Detailed workplan: timeline, dependencies, and responsibilities

! Determining proper action difficult without understanding of specific, actionable causes of dissatisfaction

! Form teams to implement required actions across the organization

! Implementation of identified actions

! Implementation cannot be prioritized correctly without validating scope or scale of problem

! Track the impact of the results and determine if the root cause has been addressed

! Resolution of problem or generation of new hypotheses if not resolved

! Data sources and tracing methodology not available to track results and isolate impact of actions

B

Generate Hypotheses Validate & Quantify

Implement Actions

E

F

Track Results

Typical Issues

! A lightweight front-end to existing operational and transaction systems to allow a faster, more flexible application

A

Gather Data

Develop Action Plan

Outputs

Many companies perform some type of root cause analysis to drive out the underlying reasons of customer dissatisfaction; however, most do not validate and quantify their hypotheses prior to implementing proposed solutions.

5

There are typically two problems encountered by companies that attempt a systemic validation and quantification of hypotheses through primary analysis. The first problem is that many forms of primary analysis have basic statistical biases. An outbound survey of sources of customer satisfaction and dissatisfaction may evoke responses from only the most frustrated customers with the most reason to vent (or, it may be asymmetric on the other side, drawing responses only from the most pleased customers who are happy to contribute a short amount of time to their preferred service provider). If the first problem with primary analysis conducted through surveys and focus groups is the slippery slope of bias, the second problem is that primary analysis conducted too broadly (e.g. through generalized call monitoring) gives statistically valid and believable results but which are often too obvious and generalized to be actionable. Based on our extensive work with telecommunications carriers there are two alternatives they typically consider in executing primary analysis through call monitoring that is at once broad enough to give statistically valid, unbiased results as well as detailed enough to provide actionable conclusions. Either the company hires a large team dedicated exclusively to monitoring and analyzing large, random sets of calls, or the company tries to implement sophisticated automated tools to perform these functions.

Both solutions turn out to have drawbacks. Hiring a team of call monitors is prohibitive due to labor costs and the massive amount of data that requires analysis. Automated call monitoring tools have various strategic or technological shortcomings. For example, text-mining software (which combs through notes typed in by specialists as they listen to calls) is vulnerable to inaccuracy, sensitive to user-defined key words, and prone to human typing errors. Also, notes entered by customer service reps often contain the solution to the problem (e.g. "gave customer $5 credit"), rather than the root cause of the trouble (e.g. "customer stated that they were misinformed of service fee"). On the more technical side, voice recognition systems would be useful but may not yet be ready for implementation. It is clear that sound root cause analysis is required to strategically address customer dissatisfaction, and that such analysis is dependent on some form of primary analysis. But it is also clear that conventional call monitoring approaches either do not make economic sense or fail to provide significant and/or actionable value. Instead, a winning solution for customer satisfaction root cause analysis would deliver a complete, statistically valid picture of the issues; it would lead to actionable recommendations; and it would be capable of being implemented at a manageable cost.

6

“A winning solution for customer satisfaction root cause analysis would deliver a complete, statistically valid picture of the issues...”

Identifying Drivers of Dissatisfaction – Crafting the Solution

III

The solution to identifying accurate, actionable drivers of dissatisfaction is a comprehensive methodology and underlying system that begins at the level of reasons for customer calls and culminates in detailed customer experience-enhancing initiatives with quantifiable ROI. The solution involves statistically rigorous call reason tracking matched to highly-targeted call monitoring efforts and is structured on three components:

1) A call reason tracker which presents a running "radar view" of the reasons behind customer calls; 2) A call library of recorded customer calls that enables rapid, efficient, targeted "deepdive" call monitoring; 3) A root cause analysis team to perform "deep dives" into specific call reasons through targeted call monitoring.

A Strategic Root Cause Analysis Solution 1 – Billing education 2 – Network coverage 3 – Dispute charges 4 – Change plan 5 – Other

1 reason 2 reason 3 reason

Obtain “Radar View” of Call Drivers

Workstep

Output

Validate previous hypotheses Track impact of initiatives

Perform “Deep Dive” Analysis to Identify Call Reasons

! Select random sample of CSRs to code calls with standardized call reasons

! Store 3-6 months of recorded calls in central database

! Perform primary analysis on specific call drivers through targeted call monitoring

! Record 100% of these CSRs’ calls

! Match list of coded calls to the recordings of those calls

! Perform more complex, secondary analyses using customer information databases

! “Radar view” providing directional guidance as to call drivers

! Library of coded calls allowing primary research team to select specific call reasons to monitor

! Actionable, accurate insights into root cause reasons of customer dissatisfaction

! Technical issues such as storage space for recordings need to be resolved early on

! A mix of primary and secondary analysis required to identify and validate call drivers and quantify their financial impact

! Ability to accurately track month-tomonth changes in call reasons ! Sample of CSRs cannot contain bias

Issues

Create a Call Library

Generate new hypotheses

! Sample size of coded calls must be sufficient to track changes over time

! Validation and quantification of root causes prior to launching initiatives

Creating a call library of recorded calls coded by call reason allows a team of analysts first to obtain a “radar view” of major call drivers and then to perform a “deep dive” analysis to identify actionable underlying causes of customer dissatisfaction.

7

The combination of the "radar view" to detect the major call drivers with the "deep dive" analysis to pinpoint the underlying causes behind these calls gave the carrier an efficient, results-oriented strategy for tackling customer service issues. From this strategy emerged recommended actions the carrier could adopt to increase customer satisfaction. Examples of these actions included improvements to the customer experience for credit-challenged customers, resolution management for the most frequent callers, and evaluation of potential call savings that could be achieved from specific IT investments. To see how these types of recommendations could be driven from the strategic root cause analysis solution, let us examine each of the solution's components in more detail. The first component is a call reason tracker, which records call reason frequencies of a random representative sample of calls each month. The application is designed to detect slight but statistically significant deviations among these call reasons from month to month. Data generated from this tool proffers a "radar view" – current and historical – of call reason patterns across the enterprise, allowing the carrier to identify major reasons for customer calls, anticipate growing problems in the customer base, and quantify the costs attributable to different call reasons. For example, "10.9% of all calls in April were billing inquiry calls" or "7.2% of calls were from customers inquiring about their minute balance.” But while these results provide a statistically sound view of the big picture, real root cause insights can only be found in the details. It's good to know that many customers are calling with billing issues, but what particular aspect of the billing process can be repaired to dramatically lessen this pain point?

Here is where the strategic component of the solution fits: highly targeted call monitoring. Once executives have studied the "radar view" of call reason results and identified issues or customer segments of concern, a call monitoring team performs "deep-dives" into the issue, validating and quantifying hypotheses for root causes. This team, comprised of five full-time employees with analytical expertise and operational experience, is charged with driving root cause insights from listening to calls, carrying out statistical validation of results, and delivering actionable recommendations. These recommendations are uniquely tailored to address specific root causes and come with quantified projections of the impact that could be expected from their implementation.

“Recommendations based on the root cause analysis identified $11 million in annualized cost savings.” The third component, a call library of recorded calls, allows for targeted call monitoring and thus fills the gap between the "radar view" of the call reason tracker and the "deep dive" of the call monitoring team. While it is easy to gather a large set of previously recorded calls that the call monitoring team could access to perform the call monitoring deep-dive, providing the user with a targeted call sample is a trickier matter. Trying to locate among a large set of randomly-recorded calls a validsized sample of calls that fit the call reason or customer segment under investigation is simply too time-intensive and costly. The solution to this challenge was a "call library," a catalogued inventory of recorded calls, searchable by call reason or other target specifications. This innovation has permitted call monitoring projects with targets as 8

focused as callers requiring a second activation call or customers from a specified credit segment calling on holiday weekends. Combining these three components to produce a complete strategy for addressing customer dissatisfaction yielded significant, tangible value for the carrier. Findings generated by the call monitoring team allowed the carrier to identify critical root cause themes. Among these themes were problems in the customer experience for credit-challenged customers that were resulting in dissatisfaction and higher cash cost per user (CCPU) for this segment. Recommendations based on these findings led to the development of business cases that identified $11 million in annualized cost savings. A separate call monitoring project confirmed the hypothesis that a decisive driver behind extraneous calls related to balance inquiries was misleading or inconsistent posting times being reported by different payment channels. Sometimes the targeted call monitoring can reject an existing hypothesis. One example looked at the call reasons for the most frequent callers. It was assumed that there were specific recurring issues that drove these subscribers to call with such regularity, and that therefore they would have a substantially different call reason profile than the rest of the base. In fact, the opposite was true. While secondary analysis revealed that less than 4% of callers were responsible for more than 20% of calls, call-monitoring efforts revealed that these customers were calling for quite similar reasons as everyone else. The findings indicated that many of these frequent callers called out of personal habit rather than for a particular reason. The concomitant recommendation suggested that the carrier single out, through a review flag on the subscribers' accounts, those frequent callers with legitimate issues (a small portion

call just to chat or simply to inundate the call center) and resolve the problem, thus improving the customer experience and averting continued frequent calling in the future. Yet another project demonstrated the importance of hypothesis validation and quantification. As noted earlier, a flaw in some forms of primary analysis is that anecdotal evidence may bring to the forefront problems that seem glaring but are actually not as crucial as other root causes. Without validation and quantification, the “problem of the moment” can often lead to a knee-jerk reaction that fails to consider the full cost/benefit picture. One example of such a problem at the company was balance and billing information provided across automated systems (e.g. web, IVR). Within the company, there existed a perception that update lags or unavailable balance/billing information were prompting a high volume of calls.

attributable to problems with the synchronization of account balances across systems, a business case was constructed to weigh the savings generated from fixing the problems versus the costs of these improvements. Since IT improvements often involve large amounts of fixed and variable cost, this case serves to illustrate the importance of targeted call monitoring determining where to focus human and capital resources. Using inputs from the call monitoring effort it was determined specific fixes would generate cost savings in the range of $8M-$10M, others would likely

produce less than $25,000 due to the low frequency with which they occurred, thereby not justifying a costly IT fix. Without targeted call monitoring to validate and quantify specific issues, problems were often identified through anecdotal feedback, leading to expenditures that outweighed the resulting cost savings. The example illustrates how root cause analysis strategy not only offers a clean, exact way of locating and validating call volume drivers but also presents a method through which to quantify the extent of the calling costs.

Reasons for Account Management Calls from Frequent Callers 50% 40% 30% 20% 10%

To understand the extent of this problem, the call monitoring team first measured the amount of calls attributable to account management, and the percent of those calls that could have been avoided. The first graph shows that for frequent callers calling about account management, 76% of the calls concern “Account Balance,” “Account Information,” or “Minute Balance,” all of which represent types of calls that could have been avoided had the caller used automated systems. The second graph quantifies the root causes behind why the callers chose to place a live call rather than use the automated systems. The findings show that callers are sometimes not aware that such systems exist or have learned not to trust these systems as they sometimes provide delayed information. Once the call monitoring analysis had given an estimate of the call volume that was

0% Account Balance

Account Information

Minute Balance

Lost/Stolen Phone

Password

76% of account management calls from frequent callers could have been handled via automated systems.

Root Causes Related to Automated Options 0%

5%

10%

15%

Trust/system update issues Prefer live rep Tried but failed Unaware of self-serve option

11% of callers who could have used automated options chose to place a live call instead because they have learned not to trust the systems. 9

Implementation Challenges While the carrier's root cause analysis solution has been a quantified success and is easy to understand in theory, execution posed various challenges. From the program management perspective, establishing a standardized process that filters, prioritizes, and schedules call-monitoring projects and dictates the consistent methodology to be followed each time is critical. It is necessary to employ a system for ensuring compliance among the advocates that are responsible for tracking call reasons. Technical challenges include development of a call-matching algorithm, acquisition of sufficient storage space for the library of recorded calls (approximately 1TB), and dependency on existing infrastructure capabilities at a company or on the willingness of management to pursue requisite IT expenditures. Justification of these expenditures hinges on clear demonstrations of ROI. Finally, there are numerous "people" challenges. One of the problems the carrier encountered was the initial reluctance of executives to adopt a strategic approach to customer satisfaction. Faced with limited resources, extreme pressure to deliver results, and short timeframes, some executives balked at what they viewed as an overly cautious methodology of validating hypotheses. Others questioned the ability of primary analysis such as codebased intervention to yield concrete results and significant returns on investment. Overcoming opposition required building a "proof-of-concept" prototype. The simplicity of the technical design allowed the team to create a small-scale call library, storing a month's worth of calls using existing server space. This low-cost, small-scale solution enabled several targeted call monitoring projects before the full build-out was approved. The results of these projects allayed executives' concerns and justified full-scale production of the call library. Choosing the right personnel for call monitoring team and getting them to deliver consistently actionable analyses presented a different type of challenge. The five members of the call monitoring team brought invaluable internal know-how; they had excellent operational experience from prior positions at the carrier that could be leveraged to gain an understanding of specific problems with internal processes, systems, and service channels. On the other hand, because their background was predominantly operational, it was not second nature for them to perceive customer satisfaction issues strategically and drive action-oriented tactics to resolve the problems. Targeted training and frequent after-project reviews helped the team increase its value.

10

Benefits Code-based intervention to improve customer satisfaction has produced substantial benefits. Quantitatively, root cause analysis enables the operator to optimize spending and maximize the returns on investment. The call library was used to make specific, actionable recommendations for improvement in processes and prioritizing and quantifying the benefits of various technology projects. Code-based intervention is another powerful weapon in a company’s analytical arsenal. By improving the precision of primary analysis through code-based intervention and combining it with existing secondary research, the carrier has identified more than $60 million in variable cost savings in the first six months after implementation for a cost of less than $1 million to implement and operate. Qualitatively, the solution has added precision to primary research. The solution's methodology and tools allow for ongoing identification, verification, and measurement of actionable drivers of dissatisfaction. The methodology can be applied across the organization (such as to understand network problems and churn reasons).

IV

V

Can Your Company Find the Root Causes of Customer Dissatisfaction? Any company facing mounting customer service costs and customer dissatisfaction challenges could stand to benefit from such a strategy. Within financial services, for example, code-based intervention could be employed to understand precisely why customers are switching to other banks, or why customers feel that product offerings are not tailored to their investment needs. Cable companies could seek to understand why subscribers are unwilling to purchase premium content or why broadband users are migrating to DSL. Companies such as software distributors or appliance manufacturers that field technical support questions could use this combination of "radar view" and "deep dives" to draw attention to the most frequent call drivers and determine the underlying causes behind recurring problems.

11

In fact, any organization with extensive customer service operations, including insurance companies, utilities, and some governmental agencies, could potentially decrease costs and increase customer satisfaction by applying the principles of this carrier's successful approach to customer satisfaction root cause analysis. By creating uniformity in the way that primary customer information is categorized and stored, code-based intervention can revolutionize customer service in the same way that Jethro Tull's seed drill changed the face of farming. Harvesting primary information becomes a secondary concern, allowing companies to focus on what really matters: identifying and addressing the root causes of customer dissatisfaction.

Visit www.diamondcluster.com for more information.