Decision Environment Improvement using Data Warehouse for Efficient Organizational Decisions - Making

by malhotra_nixie8 on 2006-05-02
Decision Environment Improvement using Data Warehouse for Efficient Organizational Decisions - Making
Navneet Malhotra1 & Anjana Gosain2

School of Information Technology, GGSIP University Kashmiri Gate, New Delhi- 110006,
1 Email:,
2 Email:


Data warehousing technology aims at providing support for decision making by integrating data from various heterogeneous systems in the data warehouse.
A data warehouse is used mainly for making effective decisions that can be used for the organizations for improving their business.
In this paper we have focused on the concept of making better decisions using data warehousing. Researchers have shown a lot of interest in the development of data warehouses (DWs) to support decision-making activities in an organization. The issue of DW quality has also been of interest to the research community. There is a true saying “Unless you plan to invest in the care of your information assets for as long as you use it, you will have wasted every penny you spent”.
Relationship between the conceptual objects in a data warehouse has been shown in the GDI (Goal-Decision-Information) approach[9]. The GDI approach addresses the needs of an organization by identifying goals of an organization thereafter possible decisions are determined to achieve the goals and then the decisional information required for taking decisions. This decisional information is kept in the Data Warehouse. However, the quality of the decisions that can be taken to achieve the goals of the organization is not addressed in the GDI approach.
Here in this paper, author has proposed a little modification in the GDI approach to handle decision quality aspect. In the proposed approach, emphasis is on the Decision Conceptual Object of the Warehouse.

In this paper focus is on the decision Environment present in an organization and its different views with the help of the Decision-Decision Model. Different views are explained to make a better understanding of the decision making in the organizations.
Author have categorized in this paper the Current data warehouse decision-making as Experience based or Reasoning Based and decision determining strategies as Goal driven or Process Driven.These strategies are fully explained with the help of a Decision-Decision Model which shows the association between the different components making the main focus on the decision and all the issues related to the decision-making are shown. There are different parameters affecting a decision made in an organization so here we have also focused on all these parameters and their effect on decision-making has been explained.

1. Introduction

Data warehousing is a new paradigm specifically intended to provide vital strategic information .The executives and managers who are responsible for keeping the enterprise competitive need information to make proper decisions. They need information to formulate the business strategies, establish goals, set objectives and monitor results. Executives and managers need to focus their attention on customer’s needs and preferences, emerging technologies, sales and marketing results and quality levels of products and services. The type of information needed to make decisions in the formulation and execution of business strategies and objectives are broad based and encompass the entire organization and this whole information can be included in one group and called strategic information.
In the last decade a lot of interest has been shown in the development of data warehouses (DWs) to support decision-making activities in an organization. There are the data-driven and the requirements-driven approaches seen in the literature. In the former, data is gathered from operational systems into DW. On the other hand, Requirements-driven approaches try to identify the information needs to be met by the DW. In these approa-ches, the real issue has been that of Data Warehouse design: given data needs what is the logical structure of the Data Warehouse.
Whereas the main focus of former information systems has been the automation of operating processes, the strategic importance of information technology for decision-making, and its use as an instrument for achieving competitive advantages has increasingly been recognized later on. Providing relevant, accurate and qualitative information is a main prerequisite for decision-making. For this purpose, various kinds of so-called management support systems (MSS) have been developed in the last 30 years. MSS can be considered as a generic term for information systems, which provide support for management tasks. Examples of MSS are management information systems (MIS) in the 60s, decision support systems (DSS) in the 70s, and executive information systems (EIS) in the 80s.
An ideal decision environment would include all possible information, all of it accurate, and every possible alternative. However, both information and alternatives are constrained because time and effort to gain information or identify alternatives are limited. The time constraint simply means that a decision must be made by a certain time. The effort constraint reflects the limits of manpower, money, and priorities. (You wouldn't want to spend three hours and half a tank of gas trying to find the very best parking place at the mall.) Since decisions must be made within this constrained environment, we can say that the major challenge of decision-making is uncertainty, and a major goal of decision analysis is to reduce uncertainty. We can almost never have all information needed to make a decision with certainty; so most decisions involve an undeniable amount of risk.
Two different approaches for the development of Data Warehouses have been proposed as the data-driven [10] and the requirements driven [11] approaches. In the former, data is gathered from operational systems into Data Warehouses. On the other hand, Requirements-driven approaches try to identify the information needs to be met by the Data Warehouse. In these approaches, the real issue has been that of DW design.
Prakash et al address conceptual objects for a Data Warehouse in the GDI approach [9]. According to the GDI approach, an explicit Requirements Engineering (RE) phase is proposed in DW development referred as Data Warehouse Requirements Engineering [DWRE] phase. In the DW domain, the information contents must be closely related to the decision-making capability to be supported. Indeed, this capability should drive DW development. However, decision making itself does not occur in a vacuum but in the larger context of the goals However, the quality of the decisions that can be taken to achieve the goals of the organization is not addressed in the GDI Approach.
Here in this paper, author propose an enhancement to the GDI model to highlight the decision making process in an organization, all the parameters affecting it, its different types etc. and this modified GDI model is named as Decision- Decision Model. This model will help the decision makers in understanding the process of decision-making in the organization more effectively, the issues related to the decision-making, its different perspectives etc.
An overview of the GDI approach for data warehouse requirements engineering [DWRE] is discussed in section 2 and Decision Decision model is also shown, which is discussed in detail in section 4, decision’s categories is discussed in section 4.1 and parameters affecting decision in 4.2 and different perspectives for determining decision quality are given in section4.3 followed by the conclusion in section 5.

2. Overview of the GDI Approach:

According to the GDI approach, to obtain a Data Warehouse that is a good fit for an organization, the DWRE process determines the goals of an organization, uses these to arrive at its decision-making needs, and identifies the information needed for the decisions to be supported. The aim of the Requirements Engineering (RE) phase is to identify the decisional information to be kept in the Data Warehouse. It deals with the identification of the goals of the organization, decisions that can be taken to achieve these goals and the information needed for decision making. There are two kinds of relationship (a) is satisfied by between goals and decisions and (b) is required for between decisions and information. The RE stage is completed when the goals, decisions, information and their relationships are all determined.
The GDI model is shown in Fig.1. A goal as an aim or objective that is to be met. A goal is a passive concept and unlike an activity/process/event it cannot perform or cause any action to be performed. A goal is set, and once so defined it needs an active component to realize it. The active component is decision. Further to fulfill the decisions appropriate information is required. A goal can be either simple or complex as shown in fig.1. A simple goal cannot be decomposed into simpler ones. A complex goal is built out of other goals which may themselves be simple or complex. The component goals of a complex one may be mandatory or optional. A decision is a specification of an active component that causes goal fulfillment. It is not the active component itself. When a decision is selected for implementation then one or more actions may be performed to give effect to it. In other words, a decision is the intention to perform the actions that cause its implementation. A decision can be either simple or complex as shown in fig.1.The simple decision can not be decomposed into simpler ones whereas a complex decision is built out of other simple or complex decisions.

3. Decision- Decision Model (Modification to GDI Model)

Fig 2

4. Decision-Decision Model

As shown in the model, a goal in an organization is satisfied by the decision. If the decision is satisfactory then the goal will be achieved else not. Hence to improve the quality of a data warehouse it is necessary that all the decisions should be satisfactory. Decision-Decision Model focuses on all the aspects related to the decision as explained in the following subsections. According to [1] decision is only as good as weakest link. And the frame used for making the decision matters. He has mentioned different decision points. According to him Frame, People and Process comprise decision declaration. Alternatives, values and information lead the decision maker to final decision

4.1 Categories Of Decision and Decision-Making

In case of warehousing; designing a data warehouse and its maintenance are also two entirely different phases. Hence in warehousing after the design, equal attention should be paid for maintaining the quality of warehouse. Warehouse is a strategy used for making efficient decisions hence the quality of decisions in warehousing should be taken care.
A decision can be categorized on the basis that whether it will satisfy the goal or not. Hence a decision can be categorized as Good quality Decision or Bad quality Decision.
Good Quality Decision is that decision which satisfies the goal up to a desired level, which is called threshold value. Threshold value of a goal is that value, if that is achieved then goal is considered to be satisfied.
Bad Quality Decisions are those decisions, which are not able to reach the threshold value of the goal hence they are called Bad Quality Decisions.
For a Quality Data Warehouse most of the decisions made should be of good quality otherwise the quality of the warehouse would deteriorate.
A Decision- making can be done by two ways: Experience Based meaning by here the decision Maker makes the decision on his previous experience and the Reasoning Based where the decision is done entirely by following proper procedure, by exploring all the alternatives and then choosing the best one. Botth have their own benefits. Sometimes experience based decision give better results than reasoning based.

3.1 Parameters affecting Quality Of Decisions

Unless you plan to invest in the care of your information assets for as long as you use it, you will have wasted every penny you spent. Following are the factors, which affect the quality of decisions in warehouse.

Process used:
Process followed for making a decision plays a major role in determining the success of goal. A decision process includes a sequence of steps followed for coming to a final decision
The Process for making a decision includes following steps:
1. Define the situation when and where to make a decision.
2. Declare the decision completely and then decide what the decision exactly is, how you’ll work it.
3. Select the criteria and make it clear in your mind that what exactly you need.
4. Work the decision: generate a complete set of alternatives; explore each and every alternative by gathering the information, and ultimately make a choice that best fits your values.
5. Finally choose the alternative that best suits you as a decision and Commit it.
6. At last step evaluate your decision that whether it has satisfied the goal or not.
A decision is categorized as good or bad quality decision and quality of decision is not related to its outcome rather it is related to the process followed to take that decision meaning by good quality process for making a decision may lead to either good outcome or bad outcome [4].
So whatever process we have followed for decision-making does not determine its outcome directly. If a good process is followed for making a good decision, it might lead to good or bad outcome. i.e. goal may not be achieved & vice a versa[4]. Hence focus is on the process followed to make the decision if good process followed means that decision is good quality, never mind whether outcome is coming satisfactory or not.
So [4] considered the process aspect and goal satisfaction as two entirely different aspects.

Goal State as a Parameter:

Sometimes we don’t focus on the process used to achieve the goal rather we focus on the result itself. If the result is satisfied then the decision is said to be of good quality, doesn’t matter what process has been followed to reach that decision. Mostly in our day today life we judge decision depending on the result meaning by if we achieve our aim then we said that our decision was good otherwise we always regret on our decision if it does not provide us result without considering the process we have followed for making that decision.
Quality of decision can be related to its outcome, if outcome is good i.e. if goal is achieved then decision is categorized as good quality decision else bad quality [2]. Criteria for determining the quality of decisions here is outcome perspective, and] a decision is good quality if goal is satisfied[2]. And goal is satisfied if decision achieves the threshold value.

Time effect

Time also is a great parameter in affecting the quality of decision. If proper Time management is done then only decision comes out to be of good quality otherwise even if a lot of time is provided to decision makers and if they do not manage it properly then it leads to a bad decision.
[5] Emphasized on time factor i.e. time management for making efficient decisions. He has given seven techniques for exploiting variable time.
Faster decisions in less time are generally not successful Hence time should be properly coordinated for making effective decisions.

Dirty data is one of the topmost reasons for the failure of warehouse. Improved data quality boosts confidence in decision-making, enables better customer service and reduces risk for disastrous decisions.
Data is the raw material from which information is derived and is the basis for intelligent actions and decisions. Organizations therefore need to understand data quality, and establish procedures to assure the quality of data in data warehouses.

If data is bad then decision coming out from it will also be bad called Bad Data In Bad Decision Out [8]. Data quality problems often cause all of the following issues and probably more [8]:
• Your offer won't reach the customer (because of a wrong/undeliverable address).
• Your offer reaches the wrong customer (because you no longer have the right contact due to job change, lost interest).
• Your offer is irrelevant and annoying, not anything the customer is interested in.
• Your offer actually conflicts with some other message the customer is hearing from your sales team or sales partners.
• Your offer medium conflicts with a specific request from the customer (no e-mail, no phone calls).
Size of Information Samples

If data is the raw material, information is a finished product. Information is
Data in context. Information is usable data. Information is the meaning of data,
So facts become understandable.
The Triad is a conceptual and strategic framework that explicitly recognizes the scientific and technical complexities of site characterization, risk estimation, and treatment design[7]. In particular, the Triad Approach acknowledges that environmental media are fundamentally heterogeneous on a variety of scales, a fact that complicates sampling design, analytical method performance, and toxicity estimations.
Cost for analyzing the samples before actual use is high because it requires sophisticated instrumentation along with experienced and well-educated operators. Hence financial motivation is also to minimize number of samples. Data quality must be cleaned otherwise decision will contain errors.
Small information sample is sufficient to make a good decision[3]. If an animal exposed to a smaller number of learning trials may be more likely to choose the safer of two territories than one exposed to a large sample of trials[3].


Information is an important ingredient for taking a good decision. Utility of information cannot decrease. Information is the very important factor in affecting the decision because if the raw material will be not of good quality then how one can expect the decision to be of good quality.

Information quality requires quality of three components: clear definition or meaning of data, correct value(s), and under-stand able presentation (the format represented to a knowledge worker)[6]. No quality of any of these three components can cause a business process to fail or a wrong decision to be made.
Information is applied data and may be represented as a formula:
Information = f (Data + Definition + Presentation)
From a business perspective, information may be well defined, the values may be accurate, and it may be presented meaningfully, but it still may not be a valuable enterprise resource. Quality information, in and of it, is useless. But quality information understood by people can lead to value.

4.3 Decision Quality Determining Strategies

After making a Decision it become important to determine its quality so that it can be used for the future references also. There can be the two ways for determining the quality of a decision and these are Process Perspective and the Outcome Perspective which are explained below:

Decision Quality Process Perspective:

In process perspective we focus on the process followed to achieve the goal. Hence here the process followed to give a decision categorizes the quality of decision whether it is good or bad. If the process followed is systematic and satisfactory then that procedure is said to be of good quality, it does not matter whether the goal is satisfied or not.
So here it need not to wait for the result to come because the goal result might take a long time. Process here means all the steps followed for making the decision. It is Scalable that means we can extend the process if our goal is more refined. Sometimes a good process may lead to a bad outcome, although probability is low but it does happen and also it takes long time to follow a process for determining a decision because At each step a number of different choices need to be considered, and this is a little more hectic to determine the quality of decision by observing the process followed to reach that decision
But the benefit is it evaluates Quality of decision even if output of goal is delayed.

Decision Quality Goal Perspective:

In outcome perspective we focus on the outcome of decision that is the effect of decision on the goal. For outcome perspective we are considering that if the decision satisfies the state of goal to a desired level called threshold then it is good quality decision else bad quality. So in outcome perspective the process followed is not paid attention rather the outcome of goal matters. If goal is achieved then decision is said to be good else bad quality decision.
A data warehouse is a set of databases created to provide information to managers and decision makers through an integrated software and hardware environment that is optimized for retrieval rather than for update integrity and transaction throughput. In order to gain long-term advantages, it is not sufficient merely to declare the decisions.
Rather these decisions must be evaluated; their effect on the goal must be measured and should be stored in the warehouse for the future use.
Quality information will be used in many new ways in the intelligent learning organization.

5.Conclusion: -

In this paper we have focused on the Decision Environment in an organization and we have considered all the issues related to the Decision by using Decision Decision Model.
Now these issues must be taken care for making better organizational Decisions and these should be stored in the warehouse along with goals for the future reference. Hence through this way a better decision environment can be made.

6. References: -

1 Hoffberg K., Decision Journey/Work,
2. Kumar M., Gosain A., Data Warehouse Development Approach for
Improving organizational decision-making. National Conference on Communication and Computation Techniques: Current and Future Trends
(10-11 Feb 2006).
3. Fiedler K., Kareev Y. Does Decision Quality Increases With Size of
Information Samples?,
4. Harris R., Introduction To Decision Making, version date: July
2, 1998;
5. Mankins C. M., Better Decisions Faster, Techniques for exploiting top management
6. Hungerford M., Defining Information Quality; book- Defining_info_quality.pdf
7.Crumbling M. D., Griffith J., Powell M. D., Improving Decision Quality: Making the
Case for Adopting Next-Generation site characterization Practices, /tio/download/char/spring2003v13n2222p91.pdf
8. Kincaid W. J., Bad Data in Bad Decision Out, Darwin:3/1/2003;
9. . Prakash. N., Gosain.A., Requirements driven data warehouse development, Proceedings CaiSE-03, PP 13-16.
10. Inmon W.H., Building the Data Warehouse, John Wiley, New York.
11. Ballard C., Herreman D., Schau D., Bell R., Kim E., Valencic A., Data Modeling Techniques for DataWarehousing,

© copyright 1999-2023, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.