For this assignment, you will write a case study analysis that focuses on the communication strategy of an organization of your choice.

Information Flow Parameters for Managing ^

Organizational Processes Developing a framework for enhancing the design of systems and

improving management control of complex relationships.

I N T H E contemporary digital economy, intangible assets, of which information is a critical com- ponent, fuel a dominant share of growth and prosperity. This is in contrast to the value added by physical assets in the erstwhile traditional business model [3]. hiformation has typically been ana- lyzed as a product, with the focus primarily derived from a snapshot view taken at a particular rime. However, emphasis on the product view falls short of a precise measurement due to the nonquan- tifiable nature of the characteristics

(such as relevance and reliability) of By Ravindra Krovi, Akhilesh C h a n d r a , information. A systematic and con- ^^ld Balaji RajagOpalan scious effort to influence and control the flow of information will lead to efficiencies in organizational processes. Therefore, it is impera- tive to manage information flow (and not just information) to improve business process efficien- cies, especially in organizational environments.

Numerous studies done in the business process understanding ot the dynamics of such flows. We redesign realm have articulated the need for ratio- propose a parameter-based guiding framework oi nalizing organizational processes [4]. If process effi- information flow to manage organizational ciencies are to be realized, it is critical to take another processes. Ir establishes a foundation to assist organi- look at the infrastructure based on the parameters zations in measuring and reporting information by affecting the flow of information. Few studies, how- better managing their flow, ever, have speciBcally addressed how flow irregulari- ties can affect the process. The framework proposed Information Flow Dynamics here adopts a process view of information, which In order to comprehend the process view, we draw requires an understanding of information character- from an analogy of information flow with fluid How. istics during its flow through communication chan- During its flow, a fluid is known to change its prop- nels, and its processing by organizational agents, erties (such as velocity and viscosity) with respect to Understanding the process view should help man- space and time llO]. Fine-tuning its measurable agers in measuring the impact of flow parameter dimensions can meaningfully alter the nature of variations on information quality. fluid flow. Knowledge of the relationship between

Companies, however, are often poorly organized properties of fluid and its flow is used in engineering and underprepared to manage such complex infor- to design efficient fluid conduits (such as pipes) and mation flows [1]. The existing state of underpre- altering fluid flow mechanisms (such as dams), paredness may partly be attributed to a lack of Could there be a conceptual equivalent of infor-

COMMUNICATIONSOFTHE ACM February 2003/Vol 46, No 2 77

d e s i g n provides the necessary infrastructure for information processing functions that influence the quality of the resulting output. The architecture of the design should, therefore,

incorporate the flow parameters and their changes in real time.

marion flow that pafallcis fluid flow dynatnics? For exatiipie, the speed with which iiiFofniation flows in at! organizatiotial process depetids oti the tiutiiher oF intermediaries chat belotig to that process. Further- more, subtle changes cati result from flow irregular- ities due to localized delays and biases. In the current busiticss environment, an understanding of flow parameters is essential For enhancing the value of complex busitiess processes and designitig systetiis

Suppliers ^ Manufacturer *

Trad iti on a.

Suppliers *

^Distributors*

Model in PC

^ Uell *

Dell Model

Industry

. Customers

Figure 1. Reduced node ^ 1 ^ . ^ ^an more effectively density and the resuttmg , ^ ^ information flow. tnanage these flows.

Alteration in the values ot any one or combitianon of these parameters should help achieve a desirable influence on the tisability of information.

Node Density A node is used to describe an entity or a group of entities capable of altering the properties of informa- tion flow. The node density is then defined by the nutnber ot intermediate nodes in the infortnation processitig chatinel. The complexity of information flow is directly related to node density. Studies have suggested the itnportance of managing coordination gaps that arise due to the lack of useful information or the presetice of incorrect or unusable information [9]. Specifically, the strategic role of IT design in reducing coordination gaps in the form of time, space, and information distances between nodes in information flow ha.s been emphasized.

The number of intermediate nodes appears to be an impottant (Victor For two reasons. First, iFdecision making at each node depends on information from other nodes, then the presence of a large number of nodes along the processing channel should result in

an increase in uncertainty. Second, a large number of riodes may impede the speed of infortiiation Fiow. If the extreme case of tnanual processing (human node) is assumed, then an increase in the number of intermediate nodes would also negatively afFect the processing efficiency oFthe entire system. Organiza- tions can manage internal and external flows by altering the number of intermediate nodes. Two broad strategies for managing such informatioti flows include supply chain integration and efficient procurement processes.

Virtual integration across supply chains. Sup- ply chain integration generates efficient information flows For participating entities by focusing on value- added components. For example, Dell Computer Corp. adopted an aggressive strategy of revamping its supply chain by pruning the non-value-added nodes and information flows (see Figure 1). By shar- ing inFormation about its detiiand forecasts across the supply chain, Dell does not need to carry inven- tory until it is needed during production. Also, sup- pliers maintain convenient shipping points to satisfy demand on a real-time basis. InFortiiation sharing and supplier accessibility help Dell manage to carry inventory on a just-in-time basis. Dell has been suc- cessHil becattse it could effectively elitiiinate the dis- tribution rtodes from its supply chain resulting in leaner order-to-delivery times [7].

However, replication of the Dell tiiodel requires a careful accomtnodation of situation-speciflc variables. For example, in cases where product development requires several components and complex configura- tions, there is tnore material flow and consequently more inFormation flow. Implementing the Dell model can be difficult for some companies (such as Ford) because their supply chains are inherently tnore complex with many layers {tier-1, tier-2, tier-3...) and intermediary companies. While tier-1 suppliers may have a well-developed IT infrastructure, suppli- ers toward the end of the chain have neither the tech- nological sophistication nor a justiflable business case to a.ssist in the sharitig of information.

Procurement process efficiencies. Reducing the nutnber of nodes can also simplify internal workflow

7 8 February ?OOJ/Vol 46, No 2 COMMUNICATIONSOFTHEACM

processes {such as procurement) and consequently decrease the typically high cost associated with pur- chases of MRO {maintenance, repair, or operating) or indirect supplies (see Figure 2). In such situations, all employees within the purchasing organization have access to a proprietary master catalog (created from various supplier catalogs), ln order to control access, the system contains procurement rules that enforce purchasing privileges. When an employee selects a product, the purchase request is routed to the super- visor (or other intermediaries) for further approval. 1 his order is eventually sent to an exchange (hosted by either the purchasing organization or by the pro- curement software vendor). The order is decomposed into suborders and then routed to individual suppli- ers. The immediate value proposition of procurement process efficiencies for most suppliers is that being connected to an exchange reduces their costs of find- ing potential buyers. Additional value may depend on the complexity of items being purchased. Most MRO purchases do not present a problem because they are standardized functions. Companies arc, therefore, able to streamline their procurement activities.

The value of a node depends on the extent of reduction in information content or decision- making quality if that node is removed. Nei- ther of these conse- quences occurred in Dell's situation by the removal of the distribu- tor node. While fewer nodes may result in a smoother transfer of

information, it is important to realize that the qual- ity of information at each node affects the efficacy of decision making at subsequent nodes.

Velocity Velocity refers to the speed of incoming information at a node. In recent times, such terms as flow and velocity are used more extensively to indicate the speed of change in the economy. Bill Gates has argued that the primary driver of organizational change will be the flow of information. The Federal Reserve Board routinely tracks the velocity of money to guide its Hscal and monetary policy. Michael Dell uses the phrase "inventory velocity" to refer to rapid inventory flow in business transactions [5].

Velocity's effect was particularly evident during holiday seasons when several e-commcrce retailers

Figure 2. Information flow in the procurement process.

were unable to handle the deluge of seasonal orders. Therefore, systems that handle millions of e-com- merce transactions (such as Web servers, database servers, and payment servers) require a design that is robust enough to sustain wide variations in the velocity of information flow without an adverse effect on their performance. Further, the existing business infrastructure (such as warehouses and delivery trucks) supporting order fulfillment processes should also be sufficiently robust to accommodate different speeds.

Typically, inventory and fulfliimeiu systems can- not manage high velocity better in a situation where the subsystems are partially automated and poorly integrated. The CIO of a major tier-one supplier to the three largest automotive manufacturers com- mented, "Ir takes two or more weeks for information from the automaker regarding the increase in the sales of a specific type of model, that translates into materials requirements for our company, to get to us. This leaves us with about a week to manage our sup- ply chain, leaving our inventory management ad hoc

at best." Covisint (www.covisint.com), an exchange system based on standards agreed upon by the three major automakers, will facilitate exchange of the type of information that compa- nies need. This exchange could lead ro an increase in the velocity of infor- mation flow.

Clearly, systems designed to facilitate the automa- tion of information exchange help to streamline the organizational processes. However, it is not always true that automated processes are less prone to influ- ence velocity. Some processes could potentially suf- fer from automation when information flows too quickly. This might occur in ERP environments where users are unaware of the consequences of their actions. For example, in the pre-ERP days, if a sales clerk entered an incorrect order (wrong specifica- tion, price, or shipping address), there usually was time to correct the error. In an ERP environment, sales order information is directly routed to the man- ufacturing module where it is scheduled into pro- duction eventually waiting to be shipped. Since the sales order module is also integrated with the accounting module, it is likely an undelivered prod- uct will result in unpaid invoices because of which the customer's credit status could possibly be down- graded. It is also possible to imagine the conse-

COMMUNICATIONSOFTHE ACM 200J/Vol. 46, No 2 79

quences of manufacturing a product without the correct specifications. In the past, when most processes were manual and paper-based, busi- nesses managed to cope with these problems because there was more time available to them to react and correct some of the inaccuracies.

Viscosity Viscosity reflects the degree of conflict at the node. The con- flict arises due to the presence of contradictory information components known as infor- mation particles-—the smallest component of" information rhat can exist independently Figure 3. Impact of flow and still retain the characteris- parameters in business

,- - (• • I I transactions,

tics or lnrormation. In such cases, viscosity appears in the form of multiple val- ues of information (multiple information flows feed similar information content to a node) that must be resolved before the node can begin processing. If there is lesser conflict between the multiple values, then a quicker resolution can occur—a situation characterized by low viscosity. However, a high degree of conflict will likely delay the resolution time—a situation characterized by high viscosity.

Consider the following example of Toys-R-Us, which illustrates the effect of viscosity on business processes. During a past holiday season, Toys-R-Us was one of the poorer performers in order fulfill- ment. The company handled both offline and online sales orders during that period. The number of online orders outweighed many times the avail- ahle processing capacity and the inventory the com- pany had in its warehouses. The company's less than satisfactory performance stemmed from the perplex- ing nature of inventory management. Prudent man- agement practice dictates maintaining inventory at lower levels to avoid storage-related costs. Contrar- ily, there are longer-term costs associated with stock- outs that include lost sates, impaired goodwill, poor customer resource management, damage-control expenditures, and changes in customer loyalty. These alternatives represent information particles of inventory cost management.

Inventory management presents an interesting administrative dilemma: maintaming excess inven- tory versus stock-out possibilities. Planning deci- sions in such cases involve seeking an optimal

inventory level-—a tradeoff between demand projec- tions by the marketing department, and inventory cost control by the production department. The constraints imposed by the two opposing elements render the decision making relatively inflexible. An understanding of the interaction and effect of such viscous information flows would have helped the company better manage the costly and lasting effect Toys-R-Us had to cope with.

The Toys-R-Us experience demonstrates the potential for adverse consequences when organiza- tions are unable to manage viscosity. The cause of such consequences is usually a lack of accurate and streamlined information across the supply chain. Viscosit)'-related uncertainty eventually results in what is known as a buUwhip effect 16]. In a bullwhip effect, entities along the supply chain resort to stock- piling (for just-in-case scenarios), thereby eventually leading to excess inventories.

Volatility Information volatility denotes the as.sociated uncer- tainty in its content, format, and/or timing. The degree of volatility may depend on the impact of exter- nal forces based on either industr\'wide or economy- wide factors. Thus, changes in economic policies or interest rate by the Federal Reserve Board (perturba- tion) are likely to affect the operating performance of an organization. Depending on the effect such changes have on the organization, they would gener- ate either laminar or turbulent information flows.

For example, an average daily volume of a few thousand transactions over a month with a variance of more than 5.000 or 6,000 could be characterized as a turbulent flow (high volatility), whereas an aver-

8 0 Fcbfu.iry 20D3/Vol 46. No 2 COMMUNICATIONS Of THE ACM

age daily volume of a thousand transacrions over a month with a variance of 500 or 600 could be con- sidered as a laminar flow (low volatility). More specifically, when the distribution of transactions is comprised of several peaks, there is a higher likeli- hood that the flow is turbulent. Similarly, spikes in transaction volume for an online trading system when there are widespread sell-offs are representative of a turbulent information flow.

It is difficult for an organization to control the timing, content and, extent of turbulence. However, knowledge of relationships between external forces and internal processes can help manage the effect on the system. Consider the example of online retailers who frequently face the problem of preparing for the surge in demand during the holiday season. Prepara- tion could entail making necessary investments In both the technical infrastructure (such as increasing the number of servers) as well as the business infra- structure (such as more efficient arrangements with delivery companies). Planning for such capacity alternatives requires an assessment of anticipated demand and subsequent translation of demand into resource requirements. Thus, an online toy retailer could project an increase in page views (resource requirements) based on fourth-quarter estimates and historical data on the average number of page views required per order.

Organizational Implications Organizations invest in e-business drivers to improve operational and financial performance [2]. Examples of such drivers include system integration, internal orientation of information technology, and cus- tomer/supplier-related processes. For successful implementation of these drivers, careful attention should be given to the parameters influencing the flow. Figure 3 outlines the role of flow parameters in influencing the nature of interaction between an orga- nization and its various stakeholders. For example, a customer order triggers various interdependent busi- ness processes and the associated information flows. Performing a credit and inventory check through related subsystems will validate the order. A satisfac- tory evaluation should initiate information flows related to the generation of production schedules, contacts with suppliers, arrangements with logistics providers, and realization of cash from customers. Hence, the relationship between the organization and the external stakeholders (such as customers, suppli- ers, and service providers) can be affected by varia- tions in flow parameters. Some illustrative implications of flow parameter variations in the con- text oi Figure 3 are described in more detail here.

Customer relationships. The number of orders per unit time would constitute the velocity of flow and may be affected by the number of nodes through which the order reaches the implementation stage at the back end ot data processing operations. Further- more, variations in the number of orders processed per day can increase the volatility of incoming traffic at electronic trading sites. Hence, back-end applica- tions must actively control the infltience of velocity and volatility of the incoming traffic.

Supplier relationships. As a general rule, the Rir- ther a decision point is along the value chain, the higher the likelihood it will be affected by node den- sity. Nevertheless, the nature of controls and interac- tions may cause node density to become critical at any decision point. For example, controls for the purchase function in Figure 3 occur at a relatively early stage in the value chain, and may affect the node density. Node density can be a factor along two dimensions: internal and external. Internal sources stem from the organization of the purchase department. A large pur- chase department witb a centralized authority struc- ture has the potential to congest the decision-making nodes. External source is a function of suppliers' pop- ulation in this example: selection amongst a large number of suppliers injects increased node density (and externally induced volatility) into the system every time a purchasing decision is involved. Further, the ability to respond to fluctuations is limited wben information systems of the company and its suppliers are not effectively integrated.

Outsourcer relationships. If credit assessment is outsourced, it might affect the nature of information flow in the value chain. The level of integration of the outsourcer s information system with the company's system will determine the velocity of information. If the credit-check function is built into the company's system, then the degree of integration within the ERP system would influence the velocity of flow. In specific instances, the velocity will also be affected by whether required information to perform the credit check is available internally or help is needed trom a credit bureau system. Also, when the organization has some information for performing credit-check activities internally but requests assistance from credit bureau systems for additional corroborating informa- tion it could end up receiving contradictory informa- tion leading to higher viscosity.

Environmental factors. Besides intrabusiness transactions, the nature of flows between corporate intranets and external entities would also be moder- ated by various environmental factors. For example, a 500-point drop in the Dow Jones Index can increase the velocity and volatility of incoming traf-

COMMUNICAT1ONS OF THE ACM February 3003/Vol 46. No 2 8r

Pic at electronic trading sites. The uncertainty in the oil market or a global financial crisis can result in high viscous flow of information hetween business entities. Additionally, legal factors sometimes force an increase in the number oi nodes along the infor- mation flow—for instance, certain states do not allow cardirect.com to sell cars directly to consumers over the Internet; such restrictions would introduce more intermediary nodes before the final decision maker processes the information.

Finally, the overall utility oi the information flow parameters has specific significance b r enhancing system design and improving management control. System design provides the necessary infrastructure for information processing functions that influence the quality of the resulting output. The architecture of the design should, therefore, incorporate the flow parameters and their changes in real time. Knowl- edge of factors affecting process efficiencies via flow parameters assumes significance in providing effec- tive management. In the future, organizations will evolve into intricate networks of dynamic relation- ships with external entities. The complexity of the resulting processes can best be managed by analyz- ing the parameters of information flow. Q

REFERENCES 1. Baron, J., Shaw, M.J.. and Bailey, A.D, Web-based e-catalog systetiis

iti B2R procuremL-nc. Commun. ACM 43, 5 (May 2000). 9.1-100. 2. Barita, A., Konana. P., Whitistoti. A., atid Yiti. T. Measures for e-busi-

ness valtie assessment. IEEE !T Professional {}an.-l-cb. 2001). 35-39. 3. Boulton, R,F,.S,. Libert. B.D., and Samek, S,M. Cracking the Value

Code. Harper Collins. 2000. 4. Davenport, T.H, and Beers, M.C. Managing information about

processes. Journal of Management Information Systems 12, I (Summer 1995). :57-80,

5. Dell. M. Where house. Eor/>rs{No\: 30, 1998), (i. Hau. L.I... Padmanabhati, V,. and Whang. S. I he bullwhip effect in

supply chains. Sloan Mitnagetneni Review (Spring l')")7). 9.^-102. 7, Magretia.J, The power of vinual ititegration: An intcrvit-w with Dell com-

puter's Michael Dell, Harvard Business Review (Mar.-Apr, 199H), 74-84, 8, Patton, D. Supply side: New software can strcamhne the aften-onerotis

task of procurement. Wall Street Jotirnal (Nov. 15, 1''99). 9, Rathniitn, S., Maliajan. V., and Whinston. A.B. Facilitating coordina-

tion in custotner support teams: A framework and its implications for the design of information technology. Management .Sciemr 41, 12 (Dec. 1995), 1900-1921,

10. Roberson. J,A. and Crowe, C.T. Engineering Eiuid Mechanics. WiLy, NY. 19%.

R A V I N D R A K R O V I (krovifi'uakron.cdu) is a professor in the

Deparrmenc of Marî if;etnent at the University of Akron, OH. A K H I L E S H C H A N D R A (aclO^uakron.cdu) is an associate professor in

the School of Accountancy at the University of Akron. OH. B A L A J I R A J A G O P A L A N ([email protected]) is an assistant

professor in [he Department of Decision and Information Scietices ai Oakland University in Rochester. ML

ACM 0002-0782/0S/02

Because usability is no longer a luxury -

it's a necessity.

For Subscriptions Call 800-342-6626

Association for Computing Macfiinery The First Society in Computing www.acm,org

f • r ' ariMfli

8 2 Febru.iry 2003/Vol 46, No J COMMUNICATIONS OF THE ACM