MASTER THESIS MATCHING ORGANIZATIONAL INFORMATION PROCESSING CAPACITY WITH INFORMATION PROCESSING REQUIREMENTS

(1)

CONFIDENTIAL

This report contains confidential and privileged material. You are notified that without permission no part of this report may be disclosed, copied or distributed, and that any other action related to this report is strictly prohibited, and may be unlawful. Koninklijke Luchtvaart Maatschappij N.V. (also known as KLM Royal Dutch Airlines) is registered in Amstelveen, The Netherlands, with registered number 33014286.

MASTER THESIS

MATCHING ORGANIZATIONAL INFORMATION PROCESSING CAPACITY WITH INFORMATION PROCESSING REQUIREMENTS

„A TASK CONTINGENCY ANALYSIS WITHIN KLM REVENUE INTEGRITY‟

By

VINCENT ANTON HOEKSTRA

University of Groningen Faculty of Economics and Business

MSc Thesis for

MSc International Business and Management Supervisor: drs. H.C. Stek

Co-assessor: dr. K. van Veen

May, 2011

Koningin Wilhelminaplein 218 1062 KS Amsterdam v.a.hoekstra@student.rug.nl

(2)

V.A. Hoekstra / MSc Thesis

2

MATCHING ORGANIZATIONAL INFORMATION PROCESSING CAPACITY

WITH INFORMATION PROCESSING REQUIREMENTS:

A TASK CONTINGENCY ANALYSIS WITHIN KLM REVENUE INTEGRITY

Author: Vincent Hoekstra Student number: S1351974

Study: MSc International Business & Management Supervisors: Drs. H.C. Stek, University of Groningen

Dr. K. van Veen, University of Groningen Mr. J. Zandstra, KLM Procurement Marketing, Communication – Chartering

Mr. M. Starrenburg, KLM Procurement Marketing, Communication – Chartering

Mr. W. Vollenberg, KLM Central Processes and Support

(3)

3

PREFACE

In September 2009 I started on a graduate internship at KLM Procurement, domain Marketing, Chartering & Communication. KLM Procurement facilitates the procurement processes of KLM‟s business units. Besides it has an advisory role regarding the spending on third parties. In this way they are also involved with Revenue Integrity.

From the management of Revenue Integrity the need arose to gather, analyze and present management information in an easy clarifying way on one spot. Hence my assignment was to develop a dashboard with management information. The aim is to show management the key metrics and trends at a glance and with that where the focus of action needs to be. The challenge was to establish a dashboard in such a way that it can be updated each month by just adding data without harming the presentation of information.

In addition to my assignment I have researched the fit between KLM Revenue Integrity‟s information processing capacity and information processing requirements. The aim is to improve structural design in order to decrease the burden on information processing capacity, which might result in higher revenue auditing performance. Moreover, the results of this research might contribute to academic knowledge about information processing fit at subunit level. This paper describes the final results of my study.

In the end, I would like to thank Huib Stek for his patience and extensive constructive feedback, Mark Starrenburg for his positive support and enthusiasm, Johan Zandstra for providing me the opportunity of this internship and sharing interesting background information about KLM, Wim Vollenberg and his team for all their support and providing me the actual assignment, data and facilities. Last but not least, I would also like to thank my family and friends for helping me through this writing process.

I hope that Wim and his team will benefit from the dashboard and from the conclusions of this study. Also, I hope trust that the dashboard and this paper provide more insight to KLM Procurement into what Revenue Integrity involves and how it contributes to maximizing passenger revenue.

Amsterdam, May 2011

(4)

4

ABSTRACT

The initial motive for this research is an information processing capacity problem observed within KLM Revenue Integrity. A literature study is used to explore the information processing theory (Galbraith, 1973), the information processing model (Tushman and Nadler, 1978), and contingent factors. Most previous research is conducted at MNC level and used the information processing theory as an intervening concept to examine the fit between organizational structure and strategy. However, less research is conducted that examines information processing fit at subunit level. This paper examines the fit between information processing capacity and information processing requirements from a task contingency perspective. The subject of analysis is a subunit that acts within the complex and dynamic environment of a major global airline. Based on the findings from the literature study and quantitative data a hybrid structural approach is proposed, which explains that information processing fit is less straightforward and mutual exclusive than literature suggests.

Keywords: information processing, uncertainty, ambiguity, task contingency, task context,

subunit structure, aviation, KLM, revenue auditing.

(5)

V.A. Hoekstra / MSc Thesis 5

1 INTRODUCTION

This study examines the fit of information processing capacity and information processing requirements within the complex and highly competitive environment of KLM Royal Dutch Airlines. First, this paper introduces KLM to provide some background information on the company. Then, it introduces Revenue Management, Central Processes & Support, and Revenue Integrity to explain the context in which this research is conducted. Among other things key functions and processes are discussed. Chapter 5 of this paper describes the impact of this context on information processing design.

1.1 KLM

On the 7th of October in 1919 Albert Plesman founded KLM Royal Dutch Airlines to serve

The Netherlands and her former colonies. Currently, KLM is a worldwide airline company based in the Netherlands. Its core businesses are passenger transport, cargo shipment, and aircraft maintenance. Moreover, it comprises the core of the KLM Group, which further includes KLM Cityhopper (100%), Transavia.com (100%), Martinair (100%), and Kenya

Airways (26%). KLM Cityhopper operates a large part of KLM‟s European network.

Transavia serves the European “low cost, low fare” market for scheduled as well as chartered flights. Martin Air carries cargo and recently abolished passenger flights. Figure 1 depicts an organizational chart of the KLM group.

(8)

8

1.2 Air France - KLM

Since their merger in 2004, KLM works closely with Air France within the AIR FRANCE KLM holding company. AIR FRANCE KLM‟s strategy can be expressed as: one Group, two Airlines, and three Core Businesses. KLM and Air France each operate their own networks with their own identity and brand within a Dual-Hub Strategy. The two hubs are Air France‟s home base, Paris Charles de Gaulle and KLM‟s home base, Amsterdam Schiphol Airport. Besides, KLM and Air France serve the North-Atlantic market in unique joint venture cooperation with Delta Airlines resulting in four additional main hubs: Atlanta, Detroit, Minneapolis, and New York-JFK. Furthermore, the AIR FRANCE KLM Group has a 25% interest in Alitalia.

In terms of financial turnover, AIR FRANCE KLM is the world's largest airline partnership; it also transports the most passengers and is the world's second-largest cargo transporter (corporate.klm.com, 2010). Together, Air France and KLM carry almost 71 million passengers on an annual basis. They employ around 107.000 people and possess 594 aircrafts that operate 2500 daily flights to more than 236 destinations all over the world. As a result, the total revenue in the annual year 2009-2010 amounted to 20,9 billion Euros (corporate.klm.com, 2010).

1.3 SkyTeam

Both KLM and Air France are members of SkyTeam, one of the global airline alliances, that offers a network of 898 destinations in 169 countries. With thirteen member airlines as of 2010, it is the world‟s second-largest alliance in terms of market share (corporate.klm.com, 2010). In 2011 and 2012 SkyTeam extends its alliances even further by welcoming China Eastern, Shanghai Airliines, Garuda Indonesia and Aerolineas Argentinas as new members.

1.4 Pricing & Revenue Management

One of KLM‟s critical businesses within passenger transport is Pricing & Revenue Management. In order to understand Revenue Integrity‟s task context it is first necessary to explain KLM‟s pricing and revenue management system. KLM‟s commercial organization consists of Marketing, Pricing & Revenue Management, Network, and Sales International & Netherlands. Together, these units are constantly involved to improve, market and sell KLM‟s passenger product in the most optimal way.

(9)

9 each member has to report to an overall senior market manager, deputy senior market manager, and two inventory managers (inventory and yield). In the end these managers are to report to the vice president of Pricing and Revenue Management. Besides, in nearly each point of sale (country) is a local station (sales agency) situated that is responsible for local price watching (by a local pricing analyst), sales, and communication with travel agents.

1.4.1 Fencing

The aim of Revenue Management is to sell the right seat, at the right moment, for the right price to the right customer in order to maximize flight revenue. They maximize flight revenue by fencing, which means setting certain rules and conditions (minimum stay, rebooking and refund flexibility) on the availability of fares.

1.4.2 Multiple fares

In fact, multiple fares are used to target different market segments and to capture a wide range of the demand curve (see appendix 2). Fares differ for each origin and destination (complete journey), sub class (9 in business class, 12 in economy class), and point of sale (country). The availability of fares depends on the related rules and conditions, days before departure, booking activity, customer contribution (willingness to pay), and point of sale (market segmentation).

1.4.3 Availability steering

Besides fencing and multiple fares, Revenue Management uses availability steering and overbooking to maximize revenue. Availability steering aims to sell the maximum number of seats at the highest possible price in order to fill an aircraft with passengers paying the maximum price that can be extracted from them. The level of booking for each flight is forecasted per point of sale per period (customer behavior is tracked based on historical data) and closely monitored. Also the demand and yield (revenue) per sub class are forecasted and monitored. Based on this data and the booking activity, the Revenue Management system determines the maximum number of seats to sell per sub class and decides whether or not to offer a certain fare to a certain point of sale. Availability steering and control occurs 340 days until 0 days before departure.

1.4.4 Overbooking

(10)

10 cancellations and no-shows, and changeability of reservations. In fact, it optimizes the balance between minimizing denied boarding while maximizing load factor (relative occupancy of an aircraft) and revenue.

1.4.5 Fare distribution

The Airline Tariff Publishing Company (ATPCO) publishes the fares and their availability of over 500 airlines worldwide. The fares are updated multiple times per day and distributed in an electronic format which can be used directly by airlines or GDSs (Global Distribution Systems). Major GDSs are Amadeus, Galileo, Worldspan or Sabre. An airline is linked to a GDSs by a Passenger Service System. Travel agents use GDSs to request a journey after which an airline‟s inventory system responds with a certain availability. This process occurs as follows:

 A travel agent proposes an availability or booking request by a GDS.

 An airline‟s inventory system:

o receives the journey data; o creates origin and destination; o calculates customer contribution;

o determines bucket (sub class) availability.

 The travel agent receives a decision to the availability (available, non-available) or

booking (accept, deny) request.

1.5 Central Processes & Support

Central Processes & Support is one of the many staff units of Pricing & Revenue Management. The main responsibilities of this staff unit are project management and Revenue Integrity (see Appendix 1 for an organizational chart). An example of project management is the introduction of a new baggage policy, which highly impacts the distribution of fare rules and conditions.

1.6 Revenue Integrity

Revenue Integrity is a subunit (subdivision of a larger unit) of Central Processes & Support

(see Appendix 1 for an organizational chart) and is responsible for KLM‟s pricing and

(11)

11 maximum period that travel agents can hold a booking that is not yet ticketed. Non-automated audits are off shored to firms in India. All exceptions and other audits out of the scope are handled manually by Revenue Integrity at KLM‟s head office. Besides, Revenue Integrity conducts physical audits at Amsterdam Schiphol Airport. In principle, revenue leakage is caused by:

 System or filing errors in the inventory system or GDS; travel agents get access to

lower fares, which is seen as violation when it is not reported by a travel agent.

 Fraud by travel agents; since travel agents expanded their knowledge of ticketing

processes, they can book creatively to achieve better access to lower fares, which is seen as fraud.

In case of fraud, Revenue Integrity cancels the ticket (if not ticketed) and/or settles the dispute by proposing a fine, which is called ADM (Agent Debit Memo), to an abusive travel agent. Such an agent debit memo can be proposed for the following reasons:

 Sales Audit / Fare Control

All fare rules and conditions are subject to audit, including commissions/discounts, misplating (mismatch between actual carrier and carrier depicted on ticket), and the correct application of baggage allowance and (fuel) surcharges.

 Subclass Abuse

All tickets sold are checked on correct application of sub class.

 Breaking of Married Segments (audit of tickets sold as origin and destination)

All reservations consisting of married segments (indirect flights consisting of multiple legs) are checked on broken segments.

 Physical typology checks

All KLM passengers holding tickets with special fares, like Tour Operator tickets, are randomly checked on eligibility at Amsterdam Schiphol Airport.

(12)

12

2 RESEARCH PLAN

This study examines the fit between information processing capacity and requirements at

subunit level. It proposes changes in structural design to improve information processing fit,

whereby it is assumed that fit increases effectiveness and performance.

This chapter describes the research approach, strategy and methods that are used to answer the research questions. First, the initial motive provides background information about why this research is important to the unit of analysis. Second, the purpose as well as the main research question and sub-questions are formulated. Next, the academic relevance is clarified by explaining why this research is worthwhile. Then, the research methods for gathering data are justified. Finally, the contents show the further structure of this paper.

2.1 Initial motive

After the liberalization of the aviation sector in the early 1990s competition among airlines increased. New challengers entered the market and introduced the concept LCC (low cost carrier). Currently, the aviation sector suffers from the global economic crisis, which puts further pressure on revenues. According to the International Air Travel Association (IATA) airlines have lost between two and three years of growth.

Also KLM was impacted by the global economic crisis. During the last fiscal year

2009/10 passenger traffic fell by 3% to 74 billion passenger kilometers and passenger traffic revenues fell by 14%. To limit the impact of the crisis, KLM initiated a financial strategy

based on cost management and strategic reconsideration of investments. Also generating

cash was highly important. This was communicated throughout the organization by the motto „Cash is King‟.

(13)

13

separately by e-mail and remains separately stored in an e-mail box. In short, these problems suggest that Revenue Integrity faces an information processing capacity problem.

2.2 Purpose

The aim of this paper is to examine how the fit between information processing capacity and requirements can be improved. It is assumed that improved information processing fit leads to increased effectiveness and performance (Tushman and Nadler, 1978; Tushman, 1979; Daft and Lengel, 1986; Kim and Umanath, 1993; Keller, 1994). To enhance fit information processing capacity can be expanded and/or information processing requirements can be reduced. However, the first leads to higher costs and the second to increased uncertainty (as a result of less available information). Consequently, this paper considers additional solutions to improve fit. The results might help to achieve Revenue Integrity‟s goals of 100% checks and 80% of cashed agent debit memos. In addition, revenue auditing units of other airlines and of firms in hospitality, rail, cruise and car rental might benefit from this research.

2.3 Research questions

As stated before KLM Revenue Integrity faces a capacity problem regarding information processing. The desired outcome is an improved „fit‟ between information processing capacity and requirements. Since it is not yet known how this exactly can be achieved, this paper uses a question rather than an outcome or propositions.

The following main research question is of particular interest in this paper:

How can the fit between KLM Revenue Integrity’s information processing capacity (IPC) and requirements (IPR) be improvedwithout expanding IPC or harming IPR?

In order to answer the main research question it is necessary to examine the following sub questions:

 What is known in literature about information processing fit and how it can be improved?

Literature helps to examine what relationships and contingent factors have been observed in the past regarding information processing fit.

(14)

14

A description of Revenue Integrity‟s tasks context and current structural and information processing design are necessary to examine which change is necessary to improve fit.

 What change is necessary to improve information processing fit?

Based on the findings from literature and the description of Revenue Integrity‟s task context and structural design it is predicted and explained which changes in design are necessary to improve information processing fit.

2.4 Academic relevance

The importance of designing capacity for handling information processing requirements has long been recognized. In the early 1970s Galbraith developed the information processing theory. In short, it considers that an increase of uncertainty leads to an increase in the amount of information that has to be processed (Galbraith, 1973; 1977).

After the 1970s and 1980s Galbraith‟s information processing theory is barely retested or updated. In the meantime, significant change has occurred in firm strategies as well as organizational designs (Wolf and Egelhoff, 2002). Organizations increasingly have to deal with the coordination of subsidiaries operating with different products in a wide area of different national and cultural environments (Manev, 2003). Although recent studies show that Galbraith‟s information processing theory is still relevant (Citrin et al., 2007; Fairbank et al., 2006; Wald, 2009; Wang, 2003; Wolf and Egelhoff, 2002), further research is necessary that examines the fit between information processing capacity and requirements (Wang, 2003). Furthermore, since most research is conducted in the manufacturing industry, Rogers et al. (1999) and Citrin et al. (2007) call for research in various industries facing different levels of technological, environmental, and market turbulence. Moreover, neither earlier nor more recent studies have directly examined information processing at subunit level (Wald, 2009).

This research aims to fill some gaps by examining the fit between information processing capacity and requirements at subunit level. In fact, it considers what consequences the features of an airline have on the basic assumptions of information processing fit.

2.5 Research methodology

(15)

15 book shares experience and knowledge of the field in order to explain various research methods. It is observed that, compared to other social sciences, management studies are characterized by a diversity in research styles. This paragraph mentions the research style of this study. It first describes the research philosophy, then the research approach and strategy, and finally the data collection and analysis.

2.5.1 Research philosophy

Research proceeds on the basis of a certain philosophy called epistemology. This refers to the nature of its subject matter and the ways in which it can be known. Within social sciences three major epistemologies can be distinguished, which are positivism, constructionism and postmodernism.

The unit of analysis of this study consists of a social phenomenon, which in this case is a subunit of a multinational organization. Therefore, this study takes a approach that is based on constructionism. This approach assumes that humans are different from other entities as people construct meaning and social reality. As humans have the unique capacity to communicate meanings, simply observe how they process information is insufficient for understanding and predicting it. More appropriate methods to conduct such research are participation, interaction, and the interpretation of written documents.

2.5.2 Research approach

Two basic approaches that research can take are induction and deduction. The first refers to theory building and the second refers to theory testing. This study can be characterized as inductive research as it aims:

 to apply the information processing theory (Galbraith, 1973, 1977), the information

processing model (Tushman and Nadler, 1978; Daft and Lengel, 1986), and contingent factors observed in literature;

 to describe the task context (contingent factors) and structural design of KLM

Revenue Integrity through participation;

 to explain the gaps observed in literature (see paragraph 3.1) and the issues

(regarding information processing capacity) observed at the subject of analysis;

 to construct an information processing model that fills the gaps observed in

literature and fits KLM Revenue Integrity.

(16)

16 laws are integrated into theories that explain observations or that predict future events. In fact, the main purpose is to explain and predict rather than to test explanatory and predictive adequacy.

2.5.3 Research strategy

The research strategy starts with asking a research question that is related to a certain phenomenon. This study uses an open explanatory research questions about how information processing fit can be improved at sub unit level. The aim is to identify contingent factors (that influence information processing requirements) in literature and to gain more in-depth knowledge about the fit between these factors and structural design within the airline industry. Hence, this study uses a qualitative research, which is further explained in the next paragraph. The unit of analysis is a revenue auditing unit of a MNC that handles dynamic pricing. Such units are mainly observed within transportation and travel industry. A drawback of qualitative research is the ability to generalize findings. However, it helps to improve understanding of a certain phenomenon.

2.5.4 Data collection and analysis

Most research in the field of information processing is based on surveys. In this study the purpose of data collection is to gain more in-depth knowledge. In particular, of the current information processing capacity and the contingent factors that determine information processing requirements. Therefore, this research is based on qualitative data.

In this study the researcher participated in an assignment to improve information processing capacity. The data is constructed by the researcher‟s own observations and experiences, by informal interviews with managers and employees of Revenue Management, Central Processes & Support and Revenue Integrity, and by attending meetings. The interview questions are related to topics as organizational goals, tasks and responsibilities, and internal as well as external environment. Besides, documentary research is used as a resource for qualitative data whereby training material, business case presentations, audit data files, and the current management information are analyzed.

(17)

17

2.6 Contents

(18)

18

3 LITERATURE REVIEW

The previous chapter mentioned that KLM faces an information processing capacity problem. This chapter examines what is known about the relationships between information processing capacity, which is reflected by subunit structure, and information processing requirements, which is reflected by task context. Moreover, it aims to identify possible task contingency factors that might moderate this relationship. The goal is to predict from theory how information processing fit can be improved at KLM Revenue Integrity.

Figure 3. Basic information processing model

This literature review considers studies within the disciplines of contingency theory and organizational theory and design. It starts first with background information on information processing research and an explanation of the information processing theory. Then, the current state of knowledge and key contributions are discussed. As less contribution are available that consider subunit level, also literature that focuses on MNC level is reviewed in order to provide additional support and findings. Finally a research model is proposed, which helps to describe KLM Revenue Integrity‟s current task context and subunit structure in order to explain how information processing fit can be improved.

3.1 Background

(19)

19 should match with the characteristics of certain variables inside as well as outside the organization (Kim and Umanath, 1993). In fact, an organization‟s information processing capacity should fit with its information processing requirements to gain optimum organizational performance (Galbraith, 1973).

However, a central problem in organizational research has been to make valid predictions about the relation of organizational context to organizational structure and performance (Kim and Umanath, 1993). This might be explained by the fact that information processing requirements vary considerably across the different subunits of an organization (Wald, 2009). This supports the idea that there is no one best organizational structure that fits a whole set of different subunits. Moreover, prior research practically ignored task contingencies in their research models (Kim and Umanath, 1993). These are important to explain why certain structural designs fit a specific context or not. Hence, this literature review aims to explore information processing fit at sub-unit level and to identify contingent factors.

3.2 Information processing theory

Galbraith (1973) developed the information processing theory. The author argued that information processing has the ability to predict what structural design best fits an organization‟s context. Tushman and Nadler (1978) extended the theory by proposing a model that is related to information processing at subunit level.

3.2.1 Information processing

(20)

20

3.2.2 Galbriath’s information processing theory

Galbraith‟s (1973) information processing theory states that an organization should match its information processing capacity to its information processing requirements. Whereby, capacity refers to structural design and requirements to perceived uncertainty. In fact, information processing requirements refer to a firm‟s dependence on up-to-date, reliable and accurate information (Wang, 2003). The higher the perceived uncertainty, the greater the requirements of information and consequently the need of information processing capacity (Galbraith, 1973, 1977; Tushman and Nadler, 1978). When the required amount of information is high and the capacity to process information is low, managers perceive high uncertainty due to a lack of information.

Information processing focuses attention on uncertainty (Simon, 1990). Uncertainty is in organizational theory defined as the absence of information (Tushman and Nadler, 1978), which is reflected in the absence of answers to explicit questions. Galbraith (1977) defined uncertainty as “the difference between the amount of information required to perform a task and the amount of information already possessed by the organization”. Uncertainty differs from risk as uncertainty defines situations in which probabilities cannot be attached and may not be predictable (Rogers et al., 1999).

An organization is assumed to be effective if there is a fit between its information processing capacity and the information processing requirements of its technology, interdependence and environment (Galbraith, 1973, 1977, Daft and Lengel, 1986). In such a case, an organization has the ability to handle the required amounts of information while gathering the appropriate information, moving it in a timely fashion, and transmitting it without distortion (Tushman and Nadler, 1978). Keller (1994) conceptualized fit as the match between information processing and technology following Daft and Lengel (1986) that the amount of information processing should fit the forces of non-routineness (uncertainty) and un-analyzability (ambiguity) of a subunit‟s technology. The required amount of information and the way it is processed forms a key determinant of organizational structure.

(21)

21 In an unstable environment with non routine tasks these systems are less effective. In such a context, organizations can either reduce the need for information or increase the capacity to process information. A reduction in the need for information can be obtained by creation of slack resources and creation of self contained tasks. The first involves lowering the likelihood of a target being missed by increasing time, reducing performance levels or increasing inventories. The second involves the creation of self-managing groups and teams that have all resources to complete a task.

An increase in the capacity to process information can be obtained by investment in vertical information systems or the creation of lateral relations. The first involves sophisticated computer systems or application of advanced operations management techniques. The second involves stimulating communication and coordination among different groups and teams by creating coordinating positions (project manager) or computer systems (ERP).

However, Galbraith (1973) did not mention a specific amount of information processing or degree of uncertainty that would lead to a certain structural design. Moreover, the different design options are not mutual exclusive, which means that they can all exists in an organization at the same time. Also he did not mention how the options evolve over time. As organizations grow or environments become more turbulent, the complexity and hence the need for information increases. Consequently, organizations have to evolve from one option to (a set of) (an)other option(s). Nevertheless, Galbraith provides an interesting theoretical

foundation for the fact that an organization‟s capacity should match its information

requirements to reduce perceived uncertainty and to increase performance.

3.3 Information processing at subunit level

(22)

22 lead to either excessive costs or to delayed and inappropriate decision-making (Kim and Umanath, 1993).

A common variable when considering task context is task related uncertainty. High task related uncertainty requires a subunit structure that facilitates high amounts of (new) information (Galbraith, 1973; 1977) and continual monitoring and frequent feedback and adjustment (Chenhall, 2003). In addition, it requires rich information, which usually involves personal mechanisms as direct contact and group meetings (Daft and Lengel, 1986).

3.3.1 Tushman and Nadler’s information processing model

Tushman and Nadler (1978) examined the structure of communication in R&D laboratories and collected data on ego-centered networks. Similar to Galbraith (1973), they view organizations as information processing systems operating under uncertainty. The authors state that the tasks of an organizational subunit vary in their degree of uncertainty. Also it is proposed that different organizational structures have different capacities for effective information processing. The authors developed and tested an information processing model that supports structural design choices. They found that the higher the uncertainty faced by a subunit, the greater the information processing requirements and consequently the need for information processing capacity. The more uncertainty faced by a set of subunits, the greater the information processing requirements and capacity of the whole organization. Furthermore, Tushman and Nadler (1978) identified several sources of task related uncertainty:

 Task characteristics: tasks differ in complexity and interdependence and hence

their amount of predictability and uncertainty.

 Inter-unit interdependence: subunits differ in the degree to which they are

dependent upon other subunits, which is related to the need of coordination and joint problem solving.

 Sub environment: environments perceived by subunits can differ in stability and

degree of change and hence perceived environmental uncertainty.

As sources of uncertainty can be internal as well as external, organizations have to design a subunit structure that fits both of them.

3.3.2 Task context

(23)

23 predictability and analyzability. As tasks differ in predictability also the degree of task related uncertainty, which a subunit has to deal with, differs (Galbraith, 1973, 1977). According to Tushman and Nadler (1978) predictability is the thread that links various task contingency studies together. However, Robbins and Barnwell (2006) state in their book that task routineness determines the degree of task related uncertainty. The following sub paragraphs further elaborate on possible contingent factors.

3.3.2.1 Task complexity

Literature often speaks of technology when considering tasks. Technology refers to the processes and methods (with or without the aid of tools and mechanical devices) in the organization that transform inputs into outputs (Keller, 1994). Two key contributions related to technology are made by Woodward (1965) and Perrow (1967). According to these authors complexity refers to the degree that technology differs in routineness and variability and consequently in predictability and analyzability:

 Routineness refers to the repetitiveness and predictability of a task (Keller, 1994).

Or to tasks with a combination of few exceptions (unexpected events) and analyzable problems (Perrow, 1967). Non-rountine tasks that are not well understood or involve many exceptions, cannot easily be pre-planned and lead high task related uncertainty (Tushman and Nadler, 1979).

 Task variability concerns the variety of a task and the number of exceptions

encountered (Perrow, 1967).

 Task predictability is the frequency of unexpected events that occur while

performing a task (Perrow, 1967).

 Problem analyzability is defined as the degree to which unexpected events of a

task can be analyzed (Perrow, 1967). When a task is characterized by low analyzability, it is difficult to develop rules and standard operating procedures. In such a case, employees rely more on judgment, experience and heuristic methods (Kim and Umanath, 1993).

(24)

24 3.3.2.2 Inter-unit interdependence

Task interdependence is defined as the extent to which subunits depend upon each other to accomplish their tasks effectively (Thomson, 1967). It exists within subunits (intra-unit interdependence) as well as between subunits (inter-unit interdependence).

Lawrence and Lorsche (1967) state that the degree of interdependence is influenced by the extend of differentiation and that it determines the level of necessary coordination. They perceived both the organization and its environment as having subsets, which means that parts of the organization deal with parts of the environment. As organizations grow, they differentiate and specialize to achieve economies of scale and scope. Consequently, the organizational goals and priorities of each subunit could vary. However, different subunits are often still linked to each other. The more differentiated subunits are and the more dependent they are upon each other, the greater the task related uncertainty and the need

for coordination and integration.

3.3.2.3 Sub environment

Lawrence and Lorsch (1967) and Duncan (1972) conceptualized the task environment as either being simple and static or complex and dynamic. The task environment refers to the variability and stability of areas external to a subunit. As these areas are outside the subunit‟s control they are potentially unstable (Thompson, 1967). Other elements to distinguish environments are: turbulence (risky, unpredictable, fluctuating, ambiguous), hostility (stressful, dominating, restrictive), diversity (variety in products, inputs, customers), complexity (rapidly developing technologies) (Khandwalla, 1977).

(25)

25 However, Robbins and Barnwell (2006) argue that care should be taken when considering the environment since actual environmental uncertainty and complexity can differ from what is perceived by an organization. Furthermore, the authors state that environments can be partly stable and partly changing.

3.3.3 Subunit structure

Subunit structure represents the allocation of tasks and responsibilities and the design of internal systems to ensure effective communication and integration. In fact, subunit structure facilitates the interactions and communications for the coordination and control of organizational activities (Daft and Lengel, 1986). Galbraith (1973), Tushman and Nadler (1978) and Daft and Lengel (1986) showed that subunits can tailor their structural design to facilitate the required amount and richness of information, which is necessary to reduce task related uncertainty and to achieve a desired level of performance.

Burns and Stalker (1961) classified subunit structure in terms of mechanistic and organic approaches. Perrow (1970) distinguished bureaucratic and non-bureaucratic subunit structures. In general, mechanistic structures are associated to hierarchy and

centralized decision-making. In contrast, organic structures are related to decentralized decision-making with a high degree of participation, open communication, and joint problem solving. A literature review of Chenhall (2003) shows that mechanistic structures rely on formal rules, standardized operating procedures, and routines while organic structures are more flexible, responsive, and involve fewer rules and standardized operating procedures. In addition, organic structures tend to be richer in information than mechanistic structures.

In task contexts associated with high task related uncertainty information processing assists in taking and implementing decisions to deal with unpredictable and unexpected events. Generally, it is believed that organic structures have a greater ability to deal with uncertain conditions than mechanistic structures (Tushman and Nadler, 1978). However, organic structures offer less control to decision-makers and are often involved with increased response time.

(26)

26

 Decentralization, which is related to authority and decision-making. It is the extent

to which decision-making autonomy is dispersed or concentrated in an organizational unit and is assessed by the degree of participation of employees in decision-making.

 Discretion, which is the extent to which employees are permitted to choose how to

complete a task. For instance, which tools and methodologies to use.

 Formalization, which is the extend of job codification and rule observation. It

describes how, when and who performs a task. In fact, it defines whether job descriptions are tight or loose, communication channels are pre-specified or open, and if operating procedures are strict or flexible.

 Coordination, which involves fitting together the various activities of a (set of)

subunit(s) to achieve its objectives. It can either emphasize a personal approach (as with the use of a project coordinator) or an impersonal one (as with the use of work plans and schedules).

According to Kim and Umanath (1993) decision-making structures are linked to

decentralization and discretion while control structures are related to formalization and coordination. They argue that mechanistic decision-making structures are represented by

centralized decision-making and minimal discretion while organic decision-making structures are reflected by decentralized decision-making and members who are mutually connected and dependent on one another. Subsequently, the authors state that control structures are distinguished by bureaucratic modes and non-bureaucratic modes. The first involves plans, schedules and forecasts, formalized rules, policies and procedures, and standardized information and communication channels while the second involves more personal coordination, and less formal modes of control for communication and feedback. The authors believe that mechanistic structures tend to use bureaucratic modes of control while organic structures tend to use more personal modes of control.

Daft and Lengel (1986) isolated the organic – mechanistic approach further in seven

structural mechanisms that fit along a continuum, which reflects the relative contribution for uncertainty reduction and ambiguity resolution (see also paragraph 3.4):

 Rules and regulations, which are established to provide a known response to

(27)

27

 Formal Information system, which consist of periodic reports and computer

databases to provide data as performance evaluations, budgets, and statistical data to managers.

 Special reports, which include one-time studies and surveys to gather data about

an issue.

 Planning, which involves managers that meet face-to-face and in groups to decide

overall targets and a general course of action.

 Direct contact, which reflects the brief discussion between two managers in case a

problem occurs.

 Integrator, which represents boundary spanning activities (to reduce task

interdependence).

 Group meetings, which include teams, task forces and committees.

3.4 Key contributions subunit level

Some contributions examined the actual information processing within individual organizations at subunit level. For instance, research by Tushman and Nadler (1978) (see paragraph 3.3.1), Tushman (1979), Daft and Lengel (1986), Kim and Umanath (1993), and Keller (1994). Based on their findings, these authors proposed models that show how subunits can be designed to fit their task context in order to reduce task related uncertainty and ambiguity and consequently to increase effectiveness and performance.

Tushman (1979) observed that high performing subunits have communication structures consistent with their information processing requirements. Employees in technical service projects seem to take changing environmental conditions not as uncertainty to be actively dealt with, but as a source of information overload or threat, which is dealt with by relying even more strongly on supervisory direction. Based on these observations, the author argues that the degree of decentralization in a communication structure should be matched with the nature of a project‟s work. The results show no task nor an environmental imperative, but a general relationship between task related uncertainty, subunit structure and performance, which supports a task contingency approach to subunit structural design.

(28)

28

impersonal media. Seven structural mechanisms for coordination and control are proposed

(see paragraph 3.3.3) that address simultaneously uncertainty reduction and ambiguity resolution. The more rich and personal media emphasize ambiguity reduction (by clarifying and reaching agreement) while less rich and impersonal media highlight uncertainty reduction (by obtaining additional data).

Keller (1994) conducted an empirical study based on questionnaires among 98 project groups. The author states that non-routine (high task related uncertainty) and un-analyzable technologies (high ambiguity) require a higher amount of information processing than routine and analyzable technologies. It is found that the fit between the non-routineness of a project group‟s task technology and the amount of information processing it engages in predicts performance. In fact, the findings extend Tushman‟s (1979) work, as Keller (1994) found that the degree of decentralization should be matched with non-routineness in task technology. Moreover, the findings add support to Daft and Lengel‟s (1986) work as the fit between non-routineness and the amount of information processing and the communication medium (measured as the amount of written information) used by a project group did predict project quality. However, the fit between task un-analyzability and information processing did not predict performance.

Umanath and Kim (1992) adopted the information processing model of Tushman and Nadler (1978) and Daft and Lengel (1986) to study the relationship between task context and subunit structure. In a follow-up study, Kim and Umanath (1993) empirically examined the impact of subunit structure on effectiveness using task as a contextual contingency influencing this relationship. Two significant interactions are observed:

 interaction of task complexity and decision-making structure;

 interaction of work flow interdependence with both decision-making structure and

control structure.

The findings support that a task contingency relationship exists between task context and

structural dimensions of a subunit in determining its effectiveness. However, the authors

(29)

29 processing density to examine if the capacity of formal structures correspond to what is assumed by literature. The results illustrate that the density in the organizations with a one-dimensional structure is lower than the density of organization with a matrix structure. In fact, it is found that information processing requirements vary considerably across the different subunits. This supports the idea that there is no one best overall organizational structure that fits a complete set of subunits. However, the subject, the richness, and the importance of information were not considered.

The previous contributions support that high performing subunits are associated with sub unit structures and communication media that are consistent with their information processing requirements. Moreover, it is observed that the fit between these concepts predicts performance. This justifies a task contingency perspective to subunit structural design. In addition, the variety of information processing requirements across different subunits justifies a subunit level approach to structural design. However, Tushman (1979) found a positive relationship between task related uncertainty and reliance on supervisors, while this is generally associated with decentralized structures. Moreover, Keller (1994) found that the fit between un-analyzability an information processing did not predict performance. This indicates that the assumptions regarding the fit between a certain approach to structural design and specific task contingent factors are not always consistent, which might bias the fit-performance relationship. Hence, the next paragraph reviews also contributions at MNC level in order to find additional support for the fit-performance relationship and observed contingent factors.

3.5 Key contributions at MNC level

Most empirical contributions on organizational information processing are conducted at MNC level. For instance, Egelhoff (1982), Habib and Victor (1991), and Wolf and Egelhoff

(2002). Inspired by Chandler‟s (1962) strategy structure imperative, these studies used

information processing theory as an abstract intervening concept to predict what organizational structure is most effective given a certain context. The authors assumed that:

 information processing requirements are implicitly part of a MNCs strategy and the

environment where this strategy acts upon;

 information processing capacity is implicitly part of organizational structure.

(30)

30 subsidiaries. Furthermore, Egelhoff (1982) states that horizontal and vertical differentiation determines information processing capacity of organizational structures. Horizontal differentiation refers to whether head offices are differentiated by functions, products or geographic regions and vertical differentiation refers to whether subunits are directly linked through organizational hierarchy. The structures that Egelhoff (1982) distinguish are worldwide functional, international division, worldwide product division, geographic region or area, and matrix/mixed. To understand the fit between strategy and organizational structure in MNCs, Egelhoff (1982) formulated hypotheses that use information processing as an abstract intervening concept as both structure and strategy have implications for information processing. The results show that an information processing model is able to explain the fit relationships. Moreover, the author found that the impact of complex phenomena can be simply predicted by describing its impact on information processing.

Habib and Victor (1991) examined next to manufacturing MNCs also service MNCs. They analyzed the strategy structure fit of U.S. manufacturing and service multinational corporations. The authors found that organizations choose their overall organizational structure in accordance with their pursued strategy and in accordance with information processing theory expectations. However, they found mixed results concerned the strategy-fit performance relationship. In fact, their results suggests that MNCs which have strategy-fit do not necessarily outperform non-fit MNCs. This might be explained by the design problem of MNCs to balance the costs of information processing capacity against the benefits of meeting information processing requirements.

Roger et al. (1999) used information-processing theory as an abstract intervening concept to understand the relationship between planning and performance within the banking industry. The author developed a model with an information processing view of planning that represents the fit between strategic planning process and strategic content. The findings show that strategy moderates the relationship between planning and performance. In fact, banks implementing different strategies (represented by Miles and Snow‟s (1978) typologies: defender, analyzer, or prospector), require different kinds and amounts of information from their planning systems.

(31)

31 relevant and applicable. This indicates that, despite significant changes in strategy and

organizational design, Galbraith‟s (1973; 1977) information processing theory is still a

relevant and useful theory to explain fit.

Also recent work of Wang (2003), Tihanyi and Thomas (2005) and Fairbank et al. (2006) shows that Galbraith‟s information processing theory is relevant. For instance, Wang (2003) empirically examined the effect of information processing fit on organizational performance. The author considered the fit between centralization, formalization and the role of information systems at the one hand and information processing requirements at the other hand. The information processing requirements were conceptualized as organizational complexity, uncertainty and interdependence. The results indicate that centralization and formalization are positively related to the importance of information systems. Moreover, it is found that firms which face high information processing requirements caused by environmental uncertainty tend to maintain a more mechanistic structure rather than to adopt an organic structure, which, combined with a greater role of information systems, produce a significant positive effect on performance.

Tihanyi and Thomas (2005) argue that an increase in information processing requirements can also be the result of foreign operations. They examined size, performance and intangible assets as moderating factors and found that these factors tend to influence needs of information. Moreover, the findings show that the information processing requirements of top managers in MNCs increased during expansion due to the complexity of managing foreign operations. This is caused by the fact that smaller MNCs tend to have higher information processing requirements as they may lack available resources and skills (information processing capacity) to fulfill the increased demand of the international environment.

Fairbank et al. (2006) show that organizations should align their IT systems with their strategy to improve performance. The authors examined the relationship between information technology and risk management performance by applying the information processing theory with a strategic contingency perspective. Similar to Roger et al. (1999), Fairbank et al. (2006) found that the relationship between information processing design choices and performance is moderated by organizational strategy.

(32)

32 performance levels. In fact, the work of Wang (2003) provides additional support for fit-performance relationship regarding complexity and interdependence. It also confirms that structural mechanisms (information systems) should match organizational structure and context.

Moreover, the work of Victor and Habib (1991) and Wang (2003) could help to explain why sub-unit level contributions did not always find consistent results. First, Victor and Habib (1991) suggest that organizations balance the costs of information processing capacity against the benefits of meeting information processing requirements. In fact, Wang (2003) suggests the fit-performance relationship may be biased by the lower cost of mechanistic structures and the presence of effective information systems. Second, Wang (2003) suggest that the degree of centralization and formalization are choices of top management. These choices may to a great extend depend on cultures (Hofstede, 1991), which might bias results regarding the fit between structural design and task contingent factors. Thus, the few inconsistencies of sub-unit level contributions might be the result of cost preferences and cultural differences rather than inconsistencies in the information processing model. The next paragraph illustrates the task contingent factors, subunit structural design features, and the research model, which are derived from this literature review and used to analyze the information processing fit at KLM Revenue Integrity.

3.6 Research model

(33)

33 Task context

Low task related uncertainty High task related uncertainty Author(s)

Task complexity  Routine task  Low task variability  Predictable task  Analyzable task

Task complexity

 Non-routine task  High task variability  Unpredictable task  Un-analyzable task Woodward (1965), Perrow (1967) Inter-unit interdependence  Low differentiation  Low interdependence Inter-unit interdependence  High differentiation  High interdependence

Thomson (1967), Lawrence and Lorsch (1967), Thomson (1967) Sub environment  Stable  Low uncertainty Sub environment  Unstable  High uncertainty

Lawrence and Lorsch (1967), Duncan (1972), Galbraith (1973)

Table1. Task context

Subunit structure

Mechanistic structure Organic structure Author(s)

Decision-making structure  High centralization  Low discretion Decision-making structure  Low centralization  High discretion Perrow (1967) Control structure  Formal (bureaucratic)  Impersonal coordination Control structure  Informal (non-bureaucratic)  Personal coordination Thomson (1967)

Table 2. Subunit structure

Required approach to structural design

Low task related uncertainty

High task related uncertainty

Author(s)

Subunit structure Mechanistic Organic Tushman and Nadler (1978), Kim and

umanath (1993), Chenhall (2003) Structural mechanism Rules and regulations Formal information system Special reports Planning Planning Direct contact Integrator Group meetings

Daft and Lengel (1986), Kim and Umanath (1993), Chenhall (2003)

(34)

34 Each type of subunit structure facilitates certain types of information processing within the subunit and between the subunits of an organization, while at the same time restricting other types of information processing. Organizations are assumed to be effective if there is a fit between their information processing capacity and their information processing requirements. Therefore, structural design should be tailored to the information processing requirements in order to attain an acceptable level of performance. It is suggested that certain matches between task context and subunit structure lead to higher organizational performance than others. On the other hand, certain mismatches could lead to less effectiveness and performance resulting in excessive costs or delayed and/or inappropriate decision-making. Information processing fit at subunit level is conceptualized as the match between:

 Information processing requirements; the content and amount of information

required to perform tasks and to make decisions.

 Information processing capacity; the content and amount of information that can be

delivered using a certain structural design

By describing task contingencies in terms of the amount and content of information processing required, a framework can be created that illustrates the fit between subunit structure and certain task contingencies (see figure 2). This model is used to analyze what structural design KLM Revenue Integrity should adopt to meet the information processing requirements of its task characteristics, task interdependency and task environment and to reduce both uncertainty and ambiguity.

(35)

35

4 RESULTS

The literature review illustrated what constitutes information processing fit. After a short introduction, this chapter describes the task context within Revenue Integrity and the current observed structural and information processing design. Based on this, the analysis chapter describes the fit between both of them and examines whether there are possibilities to improve this fit.

4.1 Introduction

Since KLM is a global carrier (70% are transfer passengers), Revenue Integrity has to deal with travel agents, KLM stations and partner airlines from all over the world. Moreover, a significant part of Revenue Integrity‟s tasks are off shored to firms in India.

Within Revenue Integrity the aim of information processing is to perform audits and to support decision-making. The processed information involves among other things planning, objective setting, monitoring, feedback and corrective action to ensure that outcomes correspond with business drivers. In fact, Revenue Integrity needs information about the quality of its audits (amount of agent debit memos) and the conditions of each point of sale

(trends in frauds that show which audit category needs more attention). This information is

important as the low margins in the airline industry make the loss of a single seat to fraud painful to recover from.

Currently, a few capacity problems are observed to process such information. First, it is not completely clear what data (that lead to management information) is available and where it is exactly stored. Second, it is time consuming to transform the current available data into management information. In fact, it takes employees away from performing audits. Third, it is hard to compare and analyze the current available management information over time, since it is distributed separately by e-mail and remains separately stored in an e-mail box. In short, these problems suggest that Revenue Integrity faces an information processing capacity problem. Consequently, Revenue Integrity decided to establish a dashboard with management information in order to help to support decision-making more effectively.

4.2 Task context

(36)

36

4.2.1 Task complexity

Revenue Integrity performs audits to ensure that the fare rules and conditions and other agreements between KLM and travel agents (and passengers) are respected. It faces high task complexity due to a complex pricing and revenue management system and partnerships with other airlines. Although its tasks seem to be routine and analyzable, it has to deal with high variability and many exceptions.

The performed tasks seem to be routine as the checks itself (see paragraph 1.5 for an overview) are repetitive over time. In fact, it is mainly the rules and conditions and the way how fraud occurs that differs. Furthermore, system errors and fraud by travel agents seem to be analyzable as Revenue Integrity has sophisticated IT systems and appropriate data available. It also has the ability to track and log how travel agents have booked a certain flight in order to reveal fraud.

However, the task variability is high as the pricing and revenue management system is characterized by multiple fares, many points of sale (markets), different sub classes (9 in business class, 12 in economy class), and changing availability of fares (availability steering, overbooking). In fact, the multiple fares are frequently adjusted as result of intense price competition and promotions (temporary low fares). Moreover, the ability of travel agents to change, cancel, or hold flights and the partnerships with other airlines further increases the complexity of audits performed.

Moreover, Revenue Integrity at KLMs head office has to deal with many exceptions (the way how fraud occurs). In fact, all questionable bookings that cannot be handled by India and whereby it is not clear whether a reservation or booking can be classified as violation or not) are handled by Revenue Integrity at KLMs head office. In addition, as travel agents continually expand their knowledge of the ticketing process, they constantly find new ways to abuse KLM‟s reservation and booking system.

4.2.2 Task interdependence

(37)

37 proposed agent debit memo can be waived or not. Consequently, Revenue Integrity is highly dependent on other Revenue Management units to ensure that such changes are communicated accurately and timely.

Although the dependency between Revenue Integrity and other Revenue Management units is high, the differentiation (which results from specialization) is low. Both are focused on maximizing revenue and speak each other‟s “language”. However, coordination and integration is complex (mainly by written information) and time consuming. Moreover, as planning and priorities vary between different units it occurs that changes and system errors are not communicated timely. Next to audit errors, this has large financial consequences as audits are in the scope that in fact should be out of the scope.

4.2.3 Task environment

Revenue Integrity operates within a global carrier that faces high uncertainty from the environment in which it competes. Within passenger transport, KLM serves a small home market with a focus on leisure. To remain in business, it relies mainly on foreign markets and connection traffic in order to fill its intercontinental flights. As a result it faces intense price competition from other airlines and airports. Moreover, the transparency of fares offered by GDS‟ (Global Distribution Systems) increases competition even further and as a consequence the uncertainty regarding revenues. In addition, KLM is increasingly confronted with altered market circumstances and changed customer needs. For instance, it had to introduce a new economy comfort class to compensate a significant drop in business class bookings. Hence, it is important to remain customer focused and to watch and match foreign markets closely.

(38)