Business Process Management Journal, Vol. 13, No. 4, pp. 588-606
Process-oriented taxonomy of BPOs: an exploratory study Authors: T. T. Niranjan, Doctoral student, K. B. C. Saxena, Faculty Sangeeta S. Bharadwaj, Faculty Management Development Institute, Gurgaon Last revised: 12 Aug 2006 Contact author email:
[email protected]
Abstract: Purpose: To classify business process outsourcing (BPO), linking it to service level agreement (SLA) design needs. Design: Develops a framework based on prior literature to classify BPOs. Illustrates it with field research of Indian vendors. Findings: Identifies criticality and complexity as the dimensions of classification. Explicates the role of SLAs along these dimensions. Limitations: An exploratory research involving four vendors. A larger study is needed to strengthen/enrich the proposed framework, and make the findings more conclusive. Practical implications: The taxonomy aids BPO industry practitioners in understanding the characteristics of different processes and the control issues arising therein. It also helps analysts to make more qualified generalizations within the BPO industry. Value: Addresses a dearth of literature on BPOs, especially from a vendor perspective. The taxonomy serves to position future work in this fast growing field of research. Keywords: Taxonomy, Business process outsourcing (BPO), Indian vendors, Service level agreements (SLAs) Category: Research paper
Process-oriented taxonomy of BPOs: an exploratory study 1.
Introduction
Today, IT-enabled Business Process Outsourcing (BPO) is a very important and rapidly evolving business phenomenon. India alone exported BPO services worth $6.3 billion during FY 2005-06, and the sector is growing at a CAGR of 35% (Nasscom Analysis, 2006). Strangely, although the ITESBPO has been around for over a decade, it has not received due attention of academic researchers. Given that there are a wide variety of BPO services today, it is a matter of concern that a well accepted taxonomy of vendor companies is yet to be developed. Without a proper classification, much of the growth projections, industry trends etc. projected at a macro-level miss out rich intra-sector information and become far less useful. This impedes finer analyses and knowledge sharing within the BPO sector. To address this, in this exploratory research we provide a taxonomy of outsourced business processes and evaluate service level agreements (SLAs) against the same. Processes are “any activity or group of activities that takes an input, adds value to it, and provides an output to an internal or external customer” (Harrington, 1991, p. 9). “BPO is the delegation of an IT-intensive business process to an external provider who owns, administers and manages it, according to a defined set of metrics” (Gartner 2004). The reader is referred to the definitional clarity provided by Rouse and Corbitt (2004) on this. In the Indian context, we refine their ‘offshore BPO’ by highlighting the distinction between offshore captive centers (where the client owns the offshore workplace) and offshore third party vendors. It is only the latter that is truly BPO, a distinction that is often missed out in industry reports. In this article, we consider only the BPO (third party) vendors as shown in Fig. 1.
BPO
In-shore BPO
Offshore BP work Offshore Third party vendors
Captive Centers
Complex IT supported processes
Fig. 1. Types of Business Process Outsourcing
The contributions of this paper are the following:
Acknowledgements: 1. An earlier version of this paper was presented at the First Int. Conf. on Mgt of GDW (Niranjan et al., 2005). The authors gratefully acknowledge the valuable feedback from the anonymous reviewers of the conference and this journal. 2. The first author is grateful to Naga for insightful comments and support, and to him and Prof. Subhash Datta, for providing access to two of the case companies. 3. The second and third authors acknowledge the research support of the European Union under EUIndia Small Project Facility (SPF) Programme (Contract No. IND SPF/191002/965/95-830).
•
It identifies the need for a new classification of vendors based on the nature of their services, and develops a framework for the same. This framework is applied on Indian vendor companies. • The use of this framework is tested by mapping the varying roles of SLAs across four Indian vendor companies representing the different types as per the framework. The paper is organized as follows: In the Introduction, the term BPO was discussed, and the scope of this research defined. In the subsequent sections, we first establish the need for proper classification. The framework is then presented. This is followed by a discussion on SLAs. We then argue that SLAs capture several important aspects of outsourcing relationships, and hence provide a good starting point for studying our proposed framework. The next section describes the research methodology. We then present the field research involving four Indian vendor companies, for studying our proposed framework. In the last section we discuss our findings and conclude.
2.
BPO and the need for classification
Ramachandran and Voleti (2004) analyze the current and future scenarios of BPO in India, and present an optimistic picture of growth. In line with this, we expect academic research on BPO that is so scanty today, to also grow exponentially. By way of illustration, a search of Proquest database covering the period 1980 to 2004 with “business process outsourcing” included in the title or abstract revealed only 4 ‘scholarly and peer reviewed’ articles (Rouse and Corbitt, 2004). A similar search in Aug 2006 revealed over 120 articles. At this stage, when research is burgeoning, and definitions are neither well established, nor commonly agreed upon, it is important that every research clarify what type of BPO it addresses. This is because BPO is too broad a term, and its subsets too varied, for consideration under one basket. It is on this basis that we make a case for classifying BPOs. A study of websites of Indian vendors (a directory is available at www.bpo.nasscom.org) reveals that they classify their services somewhat loosely. Industry vertical is perhaps the only unambiguous basis for classification (e.g. Finance BPO, HR BPO, Admin BPO). Other bases such as high-end vs. low-end, and strategic vs. non-core services, are either misunderstood or misused. This impression was strengthened during our preliminary interviews with company executives. Given such a situation, proper taxonomy is likely to facilitate progression of this important field of research. The importance of taxonomy can be gauged by observing that in well developed fields like Services management, considerable research has been undertaken to classify and understand Services. Verma (2000) provides a review of literature on service taxonomies/typologies. “The diversity of the service sector makes it difficult to make useful generalizations concerning the management of service organizations. Therefore, a considerable amount of research has been focused on developing service classification schemes”. Lovelock (1983) and several subsequent authors have argued that unlike managers in the manufacturing sector, the service sector managers are bound in thinking that each sector is unique and lessons cannot be transferred from industry segment to another. “But some of these services do share certain relevant characteristics, and stage can be set for cross fertilization of concepts and strategies through proper taxonomy” (Lovelock, 1983). In a seminal paper exhorting taxonomy of services, Schmenner (1986) echoes this sentiment by noting that “over the years, manufacturers have been unified by their acceptance of certain terminology to describe generic production processes: job shop, batch flow, assembly line, and continuous flow process. However, service managers all claim to have idiosyncratic operations and consequently they do not promote service operations management techniques with the same vigor”. From this, it is clear that a proper classification is an important step toward theory development in a new field. Today such classification is sorely missing in the BPO sector. To address this, we propose a classification that identifies common characteristics and facilitates cross fertilization of ideas and best practices across different industry verticals within BPO industry. As against some of the existing client centric classifications, our objective is to classify BPO vendors, and for this we need a process centric approach. Toward this end, we begin with a discussion of the popular, and as we argue, loosely used classification based on ‘strategic’ vs. ‘non-core’ nature of outsourced service. The concept was originally introduced in a classic work by Quinn and Hilmer (1994). They suggested that companies concentrate their resources on a set of core competencies where they could achieve preeminence, and strategically outsource all other activities, including those that were traditionally considered integral to the company. They noted the difficulties in identifying
what exactly are ‘core competencies’ because “most of the literature on this is tautological” and “there is little consistency on what core’ really means”, and “consequently, many executives have been understandably confused about the topic”. Similarly, in a contemporary paper, Aron and Singh (2005) note that without a standard methodology for differentiating processes, most executives find it tough to distinguish between core, critical and commodity processes, and “they endlessly debate the differences between the core and critical ones”. However, today, a study of company websites reveals that many Indian vendors claim to be strategic service providers to their clients, and take pride in calling themselves so. Obviously, this claim is questionable because of a certain degree of confusion about these terms. In fact, as Kakabadse and Kakabadse (2002) argue, certain authors consider what is core and what is peripheral to be an academic debate. They posit that administering outsourcing relationships (the contracts and informational exchanges) effectively becomes a core competence in itself. Quelin and Duhamel (2003) bring in further clarity on what constitutes strategic outsourcing. “Historically, outsourcing was seen as a cost cutting device and only the non-core activities were outsourced. It is only recently that more core activities have been outsourced. Alexander and Young (1996) challenge the conventional wisdom that core activities should never be outsourced. Activities critical to performance are distinguished from activities that create a competitive advantage. The first type refers to activities that support the core businesses. The second type refers to activities that create a competitive advantage for the firm. Strategic outsourcing refers to outsourcing of both/either of these types of activities with the objective of achieving competitive advantage from doing so.” Thus, the distinctive feature of strategic outsourcing is the objective behind outsourcing, and not the apparent strategic nature of the function being outsourced, per se. Strategic Outsourcing is thus a heavily client centric term, not well understood by vendors.. From the vendor’s perspective, the concept of strategic outsourcing is not very relevant, because of the following reasons. The vendor is not privy to the reasoning and objectives (strategic or cost cutting) behind a client’s decision to outsource, something that is necessary before one can talk of the term ‘strategic outsourcing’. Strategic outsourcing is a management decision of the client, the nuances of which exist only in the minds of its top management. To be able to appreciate the objective of outsourcing, the vendor would require deep understanding of the client’s business environment: a tall order, given the type of transcontinental, arms-length outsourcing relationships typically serviced from India. Therefore, a classification based on the apparent sophistication or ‘strategic nature’ of some of the services provided by the vendor would likely be superfluous or flawed. It is for this reason that we propose a process-oriented taxonomy that is vendor centric, easy-to-understand, and unambiguous.
3.
Proposed classification
In the previous section, we identified a pressing need for classification in the BPO context. To address this, we propose a classification based on prior literature. In a widely quoted paper providing “grouping of services other than by industry classifications”, Lovelock (1983) concluded that combining classification schemes in two dimensional matrix yields better insights than along one variable at a time. We base our classification along two dimensions: the degree of criticality and complexity of the outsourced service. We shall first relate it to the client centric classification schemes available in literature: those proposed by Aron and Singh (2005) and Mani et al. (2005). Aron and Singh (2005) identify the traits of processes that are good candidates for outsourcing. They provide a framework for evaluating operational risk and identify codifiability of work and the precision of metrics used to measure process quality, as the dimensions. Codifiability refers to the ability to document the work the employees do, describe different situations they face, and stipulate what the employee responses should be in each scenario. The difficult-to-codify work coupled with imprecise/subjective involves the highest risk. Similarly, the easy-to-codify work along with precise/objective metrics suggests lowest risk. Codifiability is closely related to the term complexity used in Mani et al. (2005) and in our present work. Mani et al. (2005) found from their empirical research of American outsourcing companies (clients) that the success of BPO relationships is largely dependent on a fit between the nature of the process and the contracts/SLAs used to monitor them (Johnson 2006). They found three particularly relevant process dimensions:
Complexity: The extent to which the people performing the process must adopt different methods or procedures to do their work, and the degree to which these individuals lack immediate, established solutions to process problems. Independence: The degree to which a process can be analyzed and changed without affecting other processes in the organisation Strategic importance: The extent to which the process affords the organisation a competitive advantage. As discussed previously, ‘strategic importance’ becomes a somewhat nebulous concept from the vendor’s perspective. Hence we build upon previous frameworks, explain our choice of the dimensions and propose the Complexity-Criticality (C-C) framework. Complexity and criticality are two features of the outsourced activities which are easily understood and readily observable, and hence especially meaningful from offshore vendor-perspective. Some authors have conceptualized processes as belonging to the following three categories, from the client perspective: Strategic processes, Core processes and Support processes (Click and Deuning, 2005; Saxena and Bharadwaj, 2006). Criticality, as used here, is the degree of essentiality of the activity without which the very existence of the client as a functioning business unit would immediately cease, and can refer to any of the above three processes. Criticality includes strategic importance of an activity as well as its importance to the business processes in general (Pandey and Bansal, 2003). Critical functions are not necessarily a source of competitive advantage (Hinks, 2002), and when outsourced, it is often only to achieve cost savings rather than for strategic reasons. An example of a critical process, from the manufacturing sector, is the task of maintaining power supply to the machines. In this, when the vendor fails, the damage to the client is immediate, severe and certain. Thus, the measures of criticality are, (i) the quickness (ii) magnitude and (iii) certainty, of adverse impact on the client’s business, when the vendor fails to provide the service. Complexity (of processes) is a concept that has been extensively analyzed by researchers in organisation design/theory. Woodard (1958) defined technical complexity of a production process as the extent to which it can be programmed so that it can be controlled and made predictable. Perrow (1967) introduced two dimensions of complexity: task variability i.e. the number of exceptions encountered, and task analyzability i.e. the degree to which search activity is needed, while performing a task. Thus, a complex outsourced service is one characterized by low technical complexity, and high task variability and high task analyzability. Such tasks depend primarily on people, and their skills, knowledge and judgment, and not on machines. The employees/vendors tend to command higher wages/billing rates because of the requirement of superior mental abilities for such tasks. Complex tasks are frequently, but not necessarily, a source of competitive advantage, or critical in the sense used above. An example from the manufacturing sector, of a complex process that is also a source of competitive advantage, is the process of designing the engine. A contrasting example is the process of budgeting in an auto manufacturer: although quite complex, it is seldom a source of competitive advantage to the firm. And neither of the two examples of complex processes is very critical in the sense we use. These two dimensions are thus orthogonal i.e. an increase along one dimension is not necessarily associated with increase in the other dimension. A vendor may view its service on a continuum of criticality (of the process to the client) and complexity on the C-C framework. Lovelock (1986) argued that merely classifying is not enough: it must offer managerial insights. The value of the C-C framework is manifold. Its use would bring in clarity on the scope of much of current research/media reporting. For example, till now, the entire BPO industry used to be grouped as one while making generalizations such as “attrition is the biggest problem faced by the BPO industry”. Now, with the C-C framework which transcends narrow industry boundaries and identifies meaningful commonalities and patterns, it would be interesting to note that it is only the low complexity business process vendors that face problems of high employee turnover. In fact, BPOs such as the financial service provider presented later in this paper, enjoy much lower attrition rates than most of corporate India! Thus, if industry analyses and comparisons of wage rates, attrition rates etc. across companies would be more meaningful and useful if they consider the complexity of the processes involved. Similarly, not all relationships involve the same level of security concerns. Damage caused by stray instances of data theft may not impact all clients equally. It may also not be justifiable to invest
in the same levels of security management or imitate successful practices across all companies. Best practices from one successful BPO vendor can be transferred easily to another only when one can establish that they share similar positions on the Criticality dimension. It is in this way that the C-C framework helps practitioners in the sector.
4.
Service level agreements
We found limited literature on SLAs in outsourcing, mainly of practitioners exhorting proper use of SLAs. However, there was little empirical research available. Yallof and Morgan (2003) identified the evolution of partnership as against contracts as the distinguishing feature of successful outsourcing relationships in the past. In order to match this trend, they underscored the importance of the use of appropriate metrics and SLAs for the success of BPO relationships: “Contracts don’t build partnerships, create SLAs that do!” They also recommend agreements to be made on how SLAs would be changed in response to changes in market conditions as and when they happen. Quelin and Duhamel (2003) observe that contracts are becoming denser, as agreements become more sophisticated. Aron and Singh (2005) argue that as SLA metrics become imprecise/subjective, the operational and structural risk increase, reducing attractiveness of those processes for outsourcing. The purpose of an SLA is to identify service commitments of both service supplier and the client at the boundary of their responsibilities (Larson, 1998). SLAs can be broadly divided into two basic types: direct services and indirect (or support services). Direct services are a set of specific deliverables which are the primary reason for outsourcing. They are supported by indirect services like review meetings, performance reporting, disaster recovery procedures etc. Feeny et al. (2005) cite service levels, end-customer opinion, speed of conflict resolution, innovation achieved and number of times that penalty clauses were invoked, as the measures of client satisfaction. In our research, we also wished to see the extent to which the SLAs covered these aspects. Misra (2004) presents a strong case for the use of metrics in establishing the level of success of BPO engagement and proposes a comprehensive framework for outsourcing metrics. Below, we build upon Misra (2004) and present the functional categories of outsourcing metrics in a good SLA in a BPO: 1. Output related Metrics: This refers to the absolute quantity of the work deliverable by the vendor. Examples are volume of work such as number of calls processed, number of successful leads generated, etc. 2. Quality related Metrics: This refers to the quality of the output, both in terms of the process (i.e. how well the calls were handled; how closely they matched the scripted responses, subjective evaluation on politeness, cheerfulness, etc.) as well as the final outcome (the end customer’s evaluation/perceptions of the way his calls were handled). 3. Responsiveness related Metrics: How quickly the vendor accomplishes the work after its receipt. 4. Efficiency related Metrics: These are floor level aggregates, measured by the metrics such as average handling time, number of calls handled per agent, etc. While these may not seem to directly impact the client, these are measures of the management competencies of the vendor and are thus important. For, vendors have to make a reasonable margin to stay in business. The worse the vendor’s business gets, the worse will the client’s business become (Rottman and Lacity, 2006). In our field studies, these were often considered by vendors as ‘internal process metrics’. 5. Risk management related Metrics: Include security (number of physical security staff, electronic security), back-up and continuity plans, network infrastructure robustness, disaster recovery scenario metrics, etc. 6. Reporting related Metrics: After appropriate metrics from the above categories have been selected, both parties must agree on to how these metrics will be reported: in what form, frequency, and to what detail. 7. Incentives related Metrics: To motivate and sustain desired behavior, it is necessary that the achievement of the desired metrics be rewarded and metrics are collected on them. 8. In addition to the above metrics, Yallof and Morgan (2003) include in SLAs, the expected demand on the process (because an important reason for outsourcing is scalability, and for the vendor to anticipate, rather than react to, demand spikes, open sharing of demand forecasts is essential).
Thus, we see that SLAs cover several aspects of the client-vendor relationship. Further, after the decision to outsource has been made, SLAs cover much of the ‘control and monitoring’ aspect which is proposed by transaction cost theory (TCE) as being central to the success of outsourcing (Williamson, 1979; Levina and Ross, 2003; Ramachandran and Voleti, 2004). “TCE established that the transaction costs-the costs associated with arranging to have work done, rather than the cost of doing work itself-offered the best explanation for the existence of separate firms” and the TCE has been readily applied by IS researchers to identify the most economically efficient governance structure in outsourcing (Aron et al., 2005). For these reasons, SLAs are perhaps the most important aspect of outsourcing and we proposed to test our classification with reference to them.
5.
Research Methodology
5.1
Research design:
The research synthesizes available outsourcing literature, trade reports, and company websites to make a case for classifying BPO work from vendor perspective, and proposes a framework for the same. The applicability of the framework is demonstrated using four Indian case companies. SLAs governing their operations are analyzed with reference to the framework to explore how they vary across the proposed dimensions. The unit of analysis is a specific process outsourced to a vendor.
5.2
Data collection
Data for this study is drawn from extensive ongoing field studies undertaken by the first author as part of his doctoral work and by the second and third authors as part of their EU-SPF funded research work (Saxena and Bharadwaj, 2005; 2006). To demonstrate the C-C framework we chose 4 companies that we knew (from a study of company websites and our field studies) were doing work representing the different types as per our classification. Selection within a category was thus based on convenience/accessibility. In the initial phase (Jun 2005), semi structured interviews were conducted, duration between 30 and 40 minutes. In each company, the respondent was the senior-most manager directly dealing with SLAs; often from the Quality department. The interviews began with a detailed description of the service and the business of the client. This was followed by questions on the nature and role of SLAs in those relationships. For this, we heavily relied on the frameworks of Misra (2004) and Larson (1998). The procedure we adopted was to first elicit all functional SLAs pertaining to the process. This was followed by questions on which SLAs were prominent and received more importance, and anything special and noteworthy about the relationship. The same procedure was followed in all companies, to maintain uniformity. After processing the data obtained, the portions pertaining to the respective companies were emailed to them. Brief follow-up calls were made to verify that the data as presented by us matched reality. In the second phase, interviews were repeated with additional respondents (including client representatives, where available) six to twelve months later, to ensure temporal stability and reduce respondent bias in the data. At the company representatives’ request, their identities have been concealed and the four vendors have been referred to by pseudonyms TeleService, Online, LegalEye and FinWhiz in this paper. It may be noted that three of these vendors are in the top 10 in size in India, and offered a wide range of processes, while we focused only on one specific process.
6.
Case Studies
6.1
TeleService
This represents the low criticality and low complexity business process in the C-C framework. We studied the call handling service provided by TeleService to a satellite based internet service provider to its US household and business consumers. Incoming calls were in response to downtime at the user end; it was often due to incorrect setting up of hardware, hardware failure, and due to weather outages. Inbound calls handling involved troubleshooting and problem resolution. The agents needed to be abreast of latest product launches/revised procedures in order to correctly handle the queries. They relied heavily on procedure manuals that listed step by step (if-then conditions), how to proceed with diagnosis and resolution. After successfully resolving the issues, the details pertaining to the caller, and the call details are updated on to the CRM software for future use. The function outsourced is low in criticality because in the event of a disaster that renders the vendor inoperable, the entire business of the client does not come to a standstill; instead, only the service related work is affected. In the long term, TeleService can easily be substituted by another, as
the skills required are not very specific to this process. It is low in complexity because of the high level of technology infrastructure used, and routine and highly programmed nature of tasks. The SLAs attempt to trade-off high customer satisfaction with the vendor’s objectives. As is common with SLAs in most call centre businesses, time and accuracy of response were the prime parameters monitored here. Metrics were highly quantified and were measured and recorded automatically. The external metrics could be categorized as: 1. Output related Metrics: number of calls handled per shift and per month, number of agents on duty at various times, etc. 2. Quality related Metrics: Accuracy of call handlings: was measured primarily as the frequency and magnitude of deviations of the agents’ response from the scripted response. Agents’ compensation is partly linked to accuracy. End customer satisfaction surveys are conducted every three months at client’s country with random samples, to capture the qualitative aspects of the service. The targeted levels of survey scores were part of the SLAs. A detailed case study by Niranjan and Srivastava (2006) provides the process metrics chart for auditing agents’ performance on quality. 3. Responsiveness related Metrics: Closely related with Quality metrics, these form the bulk of the SLAs. They are: (a) 80% of inbound calls to be answered (picked up) within 15 seconds and 95% of calls to be answered (picked up) within 30 seconds. These times translate into floor level metric called the Average Speed of Answer (ASA) in the industry parlance. (b) 80% of calls to be resolved (closed successfully) within 9 minutes and 95% of calls to be resolved (closed successfully) within 13 minutes. This is called the Average Handling Time (AHT) in the industry parlance. (c) Calls abandonment to be under 3%. During heavy load periods, when the queue length exceeds the system capacity, the calls are kept pending, without even providing an automated request to wait. Many of such callers hang up, and this is called abandonment. 4. Risk management related Metrics: these were agreements on safety measures like physical security, data backup frequency, backup plans, etc. These were not measured on a day to day basis; rather, they were more like one-time measures. 5. Reporting related Metrics: data on responsiveness and quality was collected through several mechanisms: in the first, the team leader simply stands behind the agent while the call is in progress; in the second, the supervisor barges into the system and overhears the calls real-time, but remotely; in the third, details recorded in the recording software is analyzed after the calls are over. These data are consolidated and floor level summaries reported to the client on (i) 8 hr shift basis, (ii) weekly basis and (iii) monthly/quarterly basis. 6. Incentives related Metrics: Penalties/rewards to the vendor in proportion to under/overachievement of targets of performance were present. These were in addition to a constant base figure of payment, agreed to at the time of contracting. Agents’ compensation was in turn linked to their own performance on responsiveness and quality of calls parameters. 7. Efficiency related Metrics: (related to number of calls handled per agent; agent ratings) were not monitored by the client. However, these formed part of the internal metrics and were closely monitored by the vendor management. Metrics had a central role in this relationship and were used as a tool for computing the incentives to be paid. Real time feedback from end customers was not be available as there was a gap of three months between successive customer surveys. Sophisticated equipment enabled easy and cheap archiving of entire conversations and all other call details. Metrics were highly objective, and well defined. Even subjective aspects like Customer satisfaction was also made measurable in terms of proportion of respondents scoring certain values in the customer satisfaction index.
6.2
OnLine
Email based customer service for Retail company: This represents the critical but low complexity type of outsourced business process in the C-C framework. The client here is a major European retail company whose major part of revenue is through online sales. The process outsourced is the email call centre back-end work to support the business. The services include processing of electronically submitted orders, follow-ups, answering
queries, consolidating multiple products orders by the same client on different days, tracking and status monitoring of orders, etc. The service is critical because the client’s business ceases to operate immediately if the vendor fails. Although the OnLine does not possess any highly client specific skills, and a replacement to it could be searched and contracted with relatively easily, it would be impossible for the client to run a major part of its business in the intervening period. The tasks are not complex because of the high level of technology infrastructure used, and the routine and highly programmed nature of tasks. The SLAs were conceptually similar to those in the call centre case explained in 6.1. except that the mode of communication here is email. Further, the call centre described in the previous case required real time response, with the maximum available buffer was limited to the call holding time. However, the email contact centre can afford to inventory the customer demand, because they can be answered after a few hours, thus improving efficiencies. The major SLAs are: 1. Number of mails handled per period. The periods range from shifts to weekly and monthly periods. This metric thus includes responsiveness as well as output related dimensions. 2. Maximum times allowed (illustrative figures) for responding to end customer mails: Different slabs depending on the type of emails: • Query replies: 80 % within 6 hours and 95% within 12 hrs. • Order processing: 80% within 18 hours and 95% within 36 hrs. • Dispute resolution: 80% within 24 hrs and 95% within 48 hours, and other time slabs. 3. SLAs relating to accuracy, adequacy of details etc. These pertain to the frequency of occurrence of errors/deviations relating to query processing. E.g. the number of faulty coding and misidentification not to exceed 3 % of the transacted volumes. Missing of details such as product returns not to exceed 2 %, etc. 4. Efficiency related metrics: The number of mails handled per agent per period was monitored. The contract with the client entailed not only stipulations on service levels, but also on minimum staffing levels. In order to ensure that the available agents achieve satisfactory levels of performance, their efficiency is monitored by the client. 5. Risk management related Metrics: these were agreements on safety measures like physical security, data backup frequency, disaster recovery plans, etc.
6.3
LegalEye
Legal process outsourcing vendor. This represents the complex and non- critical type of outsourced business process, in the C-C framework. LegalEye offers a range of legal services to clients in the form of reports on past cases, latest laws, consultancy, recommendations, IPR management etc. For our study, we considered the ‘Multi jurisdictional survey’ service offered to its US clients. The typical client for this service is either an individual lawyer or a law firm based in the US, and specializing in corporate law. The function outsourced is the collection and presentation of various laws and statutes requested by the client. In the following discussion, the law firm is the Client, who in turn represents the end client. LegalEye has a team of qualified lawyers at its Indian site. In addition to their basic qualifications (frequently, post graduation in Law), these lawyers are trained and certified in US laws. The highly complex nature of the lawyer agents’ services is quite obvious. The task is low in criticality, because, if LegalEye is suddenly cut-off, the client can still continue its business. In such an event, although it would take considerable time and effort to search and develop a replacement to LegalEye, the client could continue to operate its business in the intervening period albeit less efficiently. Once a request is received at LegalEye’s site, data is collected from various sources: paid American law databases, legal encyclopedias, law newsletters and journals. Ready to use templates which simply need filling are available there. Usually, a major part of the current request is filled in with previously collected data (for past requests by other clients). The remaining specific data is compiled, passed through Quality control (that deals mostly with adherence of presentation to the standards), and replied to the Client. Specific queries at short notice are replied in raw data form, while others are encapsulated in a memorandum.
SLAs are considered important in this relationship, but they are at a broader level, and do not go into the finer details of operations. 1. Output related metrics: Contracts were for fixed periods. The quantity of work in terms of number of states to be covered, and the periodicity and scope of reports are agreed while drawing up the contract. It is next monitored at the time of payment or re-contracting, and is not a significant SLA metric. 2. Quality: Accuracy, relevance and good presentation of the data sought, are the attributes of quality. 3. Responsiveness related metrics: Perhaps the only true metric, measurable frequently. Refers to the quickness of responses. Except in cases where information was sought urgently, in which cases responses were in fact usually sent very fast, these measures were not considered important. They were essentially only on paper. 4. Risk management related metrics: These include • Precautions taken to protect sensitive case details and end client data: these include coding of Clients to protect identity, Information security etc • Limited liability of LegalEye i.e. liability in case of damage caused by inaccuracy/wrong usage of its reports 5. Reporting related metrics: Closely linked with 2 and 3 above, these SLAs list the presentation standards, frequency of reporting and mode of communication. The clarity of the requests/reports made by the client/vendor; their adherence to prescribed formats, brevity, use of appropriate legal language etc. are laid down for mutual compliance. The Client is also required to freely share its documentation policies and the final outcomes of their cases. This is to enable LegalEye to appreciate how data provided by it is actually used by its client. 6. Incentives related metrics: Were absent. The incentive for LegalEye for good performance was the goodwill and word of mouth spread of its service, among the US legal fraternity, leading to better business for him. 7. Efficiency related metrics were not in use. This outsourcing context is characterized by high complexity, and difficulty of laying down metrics. Here SLAs have the role of providing a broad framework of ground rules under which to operate, rather than to providing a basis for payments/incentives. Also, the client and vendor are in good and frequent communication (through phone, email). Perhaps because of the complexity and inability to fix the quality of service, there are no performance related incentives/penalties. The vendor still notes the importance of SLAs in leading it towards desired behavior. For example, requests conforming to specified formats leads to efficiency and faster responses for the vendor. The resulting reputation and increased clientele by word of mouth, rather than the monetary reward, is the motivation for better performance. The few concrete metrics i.e. reporting formats, timeliness and legalese were within control of vendors. However, even here violations are rarely viewed seriously due to existence of good communication. Most metrics used were highly subjective.
6.4
Vendor FinWhiz
Financial Process outsourcing vendor. The service outsourced in this case is a very core function and is both critical and complex according to the C-C framework. The client here is a financial services market association, ‘WatchDog’ of a European country. All financial services players there need to comply with the country’s financial services regulations. The front end agents, either as individuals or representing financial service companies, sell financial solutions to end customers (the public, in general). These services range from advising end customers about various insurance schemes, mortgage solutions, payment options, etc. and offering/selling customized solutions to suit their risk profiles. These agents and other financial intermediaries are members of the client WatchDog. WatchDog exists primarily to ensure better compliance with regulations, and to promote growth of the country’s financial services industry in general. This core function has now been outsourced to FinWhiz. WatchDog now employs a skeletal staff solely to manage this outsourcing relationship. Fig 3 describes the relationship. FinWhiz employs a team of people who are given extensive training on the client’s financial service industry and possess a deep understanding of the regulations. WatchDog randomly picks up cases and routes them by email to FinWhiz for compliance check. This sample is the input to
FinWhiz, and is a measure of its workload. Within about a day, its executives identify any deviations, and communicate the suggested improvements directly to the agents, with a copy to WatchDog. To monitor the performance of FinWhiz, WatchDog in turn picks out a sub sample of the sample and subjects it to analyses. Any deviation from the ideal is discussed with FinWhiz. The criticality of the function is due to fact that the core function (the primary reason of existence of WatchDog) has been outsourced. After outsourcing, WatchDog’s original compliance team has been disbanded and now reduced to a core team of only a handful of managers involved in managing the outsourcing relationship. Hence, WatchDog’s operations would cease immediately if FinWhiz fails to deliver. In other words, FinWhiz’s failure would have the same business impact as that of WatchDog itself failing. Also, it would take a considerable effort and time for WatchDog to muster up enough people, trained sufficiently to meet the requirements of substituting FinWhiz. The complexity of the function is quite obvious, and is due to the high degree of training, skill and judgment required of the agents in performing the tasks. Client’s financial services environment Public
agents
Europe
Membership
WatchDog
Compliance service
India
Outsourcing
FinWhiz
Fig. 2. Business model of FinWhiz Here, SLAs have the objective of meeting WatchDog’s objectives of enhancing compliance and fostering development of the financial services industry in general. The SLAs in use here seem to target both output as well as final outcome to the client’s business. They can be classified as follows: 1. Output related Metrics: The number of cases handled, and the speed of processing them. Forecasts provided by WatchDog also form part of this metric. The workload (sample size of cases directed for processing) varies with the growth of financial service industry in the client’s country. WatchDog shares its forecasts of the industry growth, as well as its estimates of the vendor agent strength needed to support compliance. 2. Quality related metrics: The compliance feedback provided by FinWhiz has to achieve a fine balance of meeting conflicting needs: the regulating body’s objective of protecting the public; the members’ profit making objective and WatchDog’s objective of improving the financial services market in general. WatchDog and FinWhiz were both aware of this. Hence, under-picking as well as over-picking is penalized. These refer to too lenient or too strict monitoring respectively, of cases with reference to the standards set by WatchDog. During periodic meetings with the client representative, calibration of compliance strictness is continually done, to match the requirements of WatchDog. This is very subjective, and hence, many of the agents are made to actively involve in the process. The proceedings are recorded in an electronic repository of calibration history. 3. Responsiveness related metrics: timeliness of response, as measured by Turn-around time (TAT): the time within which cases received are to be processed and fed back into the system.
4. Efficiency related metrics: to encourage improvement of vendor’s efficiency. This aspect is covered by measuring the number of cases handled per employee. There is also an SLA on minimum staffing levels. 5. Risk management related metrics: Although direct measures like volume, TAT and quality are available, WatchDog closely regulates the processes at FinWhiz: quality of hires, their training and certification, number of agents etc. This is because the function outsourced is of a very core and critical nature, and any failure would simply bring WatchDog’s operations to a standstill. There is a high degree of mutual dependency. The training imparted to the agents is highly client specific, and asset specificity is very high for the vendor. On the positive side, high employee attrition, which has been the biggest problem in the Indian BPO landscape, is virtually nonexistent in this case. 6. Reporting related metrics are covered under agreements on the periodicity of meetings with client representatives, mode of feedback etc. This is closely linked with Quality metrics. The relationship is characterized by very good communications. In fact, one manager from WatchDog is permanently stationed at FinWhiz’s centre. During our interviews, he explained his role as “facilitating good communications and actively involving in the Calibration process. This role is crucial not just for measuring quality of past work, but in fostering better performance in the future.” 7. Incentives related metrics: These primarily include agreements on penalties for deviations from desired performances. Penalties are categorized by color codes (10 levels), and graded in terms of seriousness. For example Red implied fatal errors: violations of some regulatory clauses, and attracts severe penalties. Green implied a low impact error such as sub-optimal advice to the end user, and attracted low penalties. Compensation of FinWhiz’s agents is directly dependent on these. It is clear in this case that there is a good understanding of the end customer environment, by the vendor. Barring some of the less important ones, the SLAs are subjective. The process of performance evaluation is mutual (to the extent of both parties literally sitting together and agreeing on it after discussions). Another feature is that many of the SLAs are intertwined i.e. achievement of some is dependent on that of some others (e.g. reporting and quality metrics). They cannot be analyzed separately.
7.
Findings
In the case studies, SLAs covered most of the measures cited in Feeny et al. (2005) as antecedent to client satisfaction: service levels, end customer opinion, speed of conflict resolution and penalties. The results also suggest substantial differences in the SLAs across the different types of processes as classified by the C-C framework. Our findings are consistent with Aron and Singh’s (2005) argument that with increasing complexity (decreasing codifiability) of the outsourced process, SLAs tend to become more subjective in nature. With increasing complexity, different functional SLAs were found to be closely intertwined with each other. Communication tended to be good (in terms of both quality and frequency), making operational SLAs (pertaining to day to day work) almost redundant. Formal relationship roles, joint workshops to understand the common process goal and performances become important for the very fact that SLAs cannot capture the complete measurement of complex processes. Then, incentives/penalties related metrics do not play such a major role in the relationships. On the contrary, the less critical outsourced processes involved huge amounts of data used in SLAs, possibly because it was possible to reduce the processes to detailed steps of operations and measurement was relatively easy. In the low complexity- low criticality process, with time, the value proposition of cost efficiency is not sufficient and the vendors have to talk other value propositions like process improvement, process transformation etc. which are difficult to lay down in SLAs. Feeny et al. (2005, p. 42) refer to this as the client’s expectation of transformational competence and relationship competency of the vendor, operationalised by innovation. In our case studies, SLAs did not cover these aspects. The variation of SLAs across the criticality dimension was less pronounced. In fact, there were no differences between the SLAs of TeleService and OnLine (where complexity was low). However, when complexity was high, increased criticality was associated with SLAs getting more
specific about service quality; penalties were explicitly provisioned; forecast sharing was included in the SLAs. It is possible that with high criticality, measures other than SLAs become more important for control. We conjecture that to safeguard its business continuity, the client takes recourse to better risk management measures and relationship management. Factors such as due diligence, length of relationship, the vendor’s brand image and reputation, and trust would become central in managing critical functions (Sabherwal, 1999; Kern and Wilcocks, 2001). This may explain the low level of differences in SLAs across criticality. Table I summarizes the SLAs across the different cases. Online
High
FinWhiz
- Highly objective, moderate amount of data - Efficiency of vendor is monitored - The rest are similar to ISP Service Call centre
Criticality Low
- Focused on service SLAs and output quality metrics - Few process related metrics; manpower details more for risk management than for day to day operations - Functional SLAs deeply intertwined - Provision for penalties for deviations
TeleServe
LegalEye
- Objective and extensive use of data - Process focused: manning levels, time records - Metrics are considered central to SLAs - Little scope for qualitative evaluation of agent performance - Outcome measured only by end customer satisfaction surveys; weakly linked to motivating desired behaviour - Penalties often enforced
- No process related metrics - Responsiveness recorded, but seldom acted upon - Excellent communications resolves most of the issues on the spot - Complexity and subjectivity of quality necessitate frequent feedback. Client shares end-product of the outsourced service - No explicitly stated penalties
Low
Complexity
High
Table I: Summary of field research
8.
Directions and conclusions
This paper provides definitional clarity on some BPO related terminology. It provides a vendor centric taxonomy of outsourced processes. It thus sets a context for future research in the area so that, by positioning their subject companies/processes on the C-C framework, future researchers can specify which type of outsourcing work they are referring to, and enable their results to be transferred to other BPOs more easily. One potentially fruitful direction for future research is to empirically study the performance of outsourcing across different positions on the C-C framework i.e. which types of outsourcing are more profitable/have higher margins, which require close monitoring and which facilitate closer relationships, which tend to continue for longer durations, etc. Further, in this exploratory research, the varying nature of SLAs across the dimensions was indicated. Rigorous study of a larger number of companies is required to enrich and confirm these results. Our field study leads us to our conjecture that dyads in the upper right quadrant (high complex – high critical) are characterized by closer relationships, high inter-dependency, high operating margins, and low substitutability, while the ones in the lower-left quadrants are characterized by arms length relations, typically involving low risk and low margins, and low duration of engagement. These issues need further explication and provide a fertile area for formal investigation. Such studies would help new companies entering outsourcing by informing how to design their relations/contracts and what to expect from them.
Industry level studies based on the framework would reveal meaningful trends and the direction in which the sector is moving, so that the industry could gear up to meet the manpower and infrastructure needs. One major trend is the growing importance of knowledge process outsourcing (KPO), as BPOs move up the value chain and foray into highly complex processes such as clinical research, engine design, marketics, teleradiology etc. (Business India, 2005). To what extent SLAs help in governing such processes is not clear, and the control issues are worth further examination. Even with the simpler BPOs, the industry is unanimous that in future, quality, and not cost is the key to growth/survival: “Clients are now expecting an augmented product and not just the core product” (BPO Summit 2005) and in established relationships, the vendor executives are even expected to play advisory roles (Business Standard, 2006). Traditional SLAs do not capture these aspects and further research is needed to identify or evolve alternative mechanisms to address them. To summarize, this paper identified a pressing need for a new classification of Indian BPO vendor companies. The Criticality-Complexity framework is a step toward addressing this need. The framework was derived from an observation of industry reports, company websites, and from prior research literature. The application of this framework was demonstrated using a case study approach. We also analyzed the SLA design needs using this framework. The results indicated substantial differences in the nature of SLAs, especially across the complexity dimension. Directions for future research were also provided.
References: Alexander, M., and Young, D. (1996), “Strategic Outsourcing”, Long Range Planning, Vol. 29, No. 1, pp. 116-119. Aron, R., Clemons, E. K. and Reddi, S. (2005), “Just Right Outsourcing: Understanding and Managing Risk”, Proceedings of the 38th Hawaii International Conference on System Sciences, Hawaii. Aron, R. and Singh, J. V. (2005), “Getting Offshoring Right”, Harvard Business Review, Vol. 83, No. 12, pp. 135-143. BPO SUMMIT 2005 (3rd) under aegis of NASSCOM, India, “Customer Delight Rules”, BPOrbit section of VOICE&DATA, Aug 2005, India. Available at http://www.voicendata.com Business India, Aug 29- Sep 11, 2005, p. 46-52. Business Standard (2006, 8 Aug), “India, not just calling” Available http://www.nasscom.in/Nasscom/templates/NormalPage.aspx?id=49809 Click, R. L. and Deuning, T. L. (2005), “Business Process Outsourcing: The Competitive Advantage”, John Wiley, Hoboken, New Jersey, USA. Feeny, D., Lacity, M. and Willcocks, P. L. (2005), “Taking the Measure of Outsourcing Providers”, MIT Sloan Management Review, Vol. 46, No. 3, pp. 41-48. Hinks, J. (2002), “The Transition to Virtual: The Impact of Changes in Business Operations on Facilities Management”, Journal of Facilities Management, Vol. 1, No. 3, pp. 272-283. Gartner Group “Vendors seek clear role in SMB market,” Gartner Dataquest Report ITSMNA-MT-0108. January, 2004. Harrington, H. J. (1991), “Business Process Improvement: The Breakthrough Strategy for Total Quality, Productivity, and Competitiveness”, New York: Mc Graw Hill. Johnson, L. K. (2006), “Successful Business Process Outsourcing”, MIT Sloan Management Review, Vol. 47, No.2, pp. 4-6. Kakabadse, A and Kakabadse, N. (2002), “Trends in Outsourcing: Contrasting USA and Europe”, European Management Journal, Vol. 20, No. 2, pp. 189-198. Kern, T. and Wilcocks, L. (2001), “The Relationship Advantage: Information Technologies, Sourcing and Management”, Oxford University Press, Oxford. Larson, K.D. (1998), “The Role of Service Level Agreements in IT Service Delivery”, Information Management & Computer Security, Vol. 6, No. 3, pp. 128-132. Levina, N. and Ross, J. W. (2003), “From the Vendor’s Perspective: Exploring the Value Proposition in Information Technology Outsourcing”, MIS Quarterly, Vol. 27, No. 3, pp. 331364. Lovelock, C. H. (1983), “Classifying Services to Gain Strategic Marketing Insights”, Journal of Marketing, Vol. 47, No. 3, pp. 9-20.
Mani, D., Barua, A. and Whinston, A. B. (2005), “Of Process Analysis and Alignment: A Model of Governance in BPO Relationships” (Univ. of Texas Working Paper). Misra, R. (2004), “Global Outsourcing: Metrics for Success of All Parties”, Journal of Information Technology Cases and Application Vol. 6, No.3, pp. 21-34. Nasscom Analysis (2006), “Indian ITES-BPO Industry-Fact Sheet”, National Association of Software and Service Companies. Available http://www.nasscom.in/Nasscom/templates/NormalPage.aspx?id=2374 Niranjan, T. T., Saxena, K. B. C. and Bharadwaj, S. S. (2005), “Towards Taxonomy of BPOs: An Exploratory Study of Indian Vendors and SLAs”, Proceedings of the First International Conference on Management of Globally Distributed Work, Bangalore, India 28-30 Dec 2005. Niranjan, T. T. and Srivastava, S. S. (forthcoming), “Managing Capacity at Sparsh Call Centre”, Asian Case Research Journal. Pandey, V. and Bansal,V. (2004), “A Decision-Making Framework for IT Outsourcing using the Analytic Hierarchy Process”, Proceedings of the International Conference on Systemics, Cybernetics and Informatics, Hyderabad, India, pp. 528-533. Perrow, C. (1970), “Organizational Analysis: A Sociological View”, Belmont, CA: Wadsworth. Quelin, B. and Duhamel, F. (2003), “Bringing Together Strategic Outsourcing and Corporate Strategy: Outsourcing Motives and Risks”, European Management Journal, Vol. 21, No.5, pp. 647-661. Quinn, J. B. and Hilmer, F. G. (1994), “Strategic Outsourcing”, MIT Sloan Management Review, Vol. 35, No. 4, pp. 43-55. Ramachandran, K. and Voleti, S. (2004), “Business Process Outsourcing (BPO): Emerging Scenario and Strategic Options for IT-enabled Services”, Vikalpa, Vol. 29, No. 1, pp. 49-62. Rottman, J. W. and Lacity, M. C. (2006), “Proven Practices for Effectively Offshoring Outsourcing Work”, MIT Sloan Management Review, Vol. 47, No. 3, pp. 56-63. Rouse, A.C. and Corbitt, B. (2004), “IT-Supported Business Process Outsourcing (BPO): The Good, the Bad and the Ugly”, Proceeding of the 8th PACIS conference, in Shanghai. Sabherwal, R. (1999), “The Role of Trust in Outsourced IS Development”, Communications of the ACM, Vol. 42, No. 2, pp. 80-86. Saxena, K. B. C. and Bharadwaj, S. S. (2005 Dec), “Business Process Outsourcing: Indian Service Providers and European Clients”, Working Paper No- SPF-2005-3. Saxena, K. B. C. and Bharadwaj, S. S. (2006), “Business Process Outsourcing as Strategic Partnering: Building Win-Win Relationship between Client and Service Providers”, Proceedings of the 39th Hawaii International Conference on System Sciences, Hawaii. Schmenner, R. W. (1986), “How Can Service Businesses Survive and Prosper?”, MIT Sloan Management Review, Vol. 27, No. 3, pp. 21-32. Verma, R. (2000), “An Empirical Analysis of Management Challenges in Service Factories, Service Shops, Mass Services and Professional Services”, International Journal of Service Industry Management, Vol. 11, No. 1, pp. 8-25. Williamson, O. E. (1979), “Transaction Cost Economics: The Governance of Contractual Relations”, Journal of Law and Economics, Vol. 22, No. 2, pp. 232-262. Woodard, J. (1958), “Management and Technology”, London: Her Majesty’s Stationery Office. Yallof, J. and Morgan, C. (2003), “Beyond Performance Standards: How to Get the Most From Your Outsourcing Relationship”, Benefits Quarterly, Vol. 19, No. 3, pp. 17-22.
