Limiting Business Intelligence Project Failures: Some Practical Hints, Part 2
For any business intelligence (BI) project to succeed, proper project management guidelines, requirements gathering and expectations have to be created. In addition, because of the complexity involved in evaluating and purchasing a data warehousing and/or BI solution, businesses should develop an understanding of the architectural options available beyond what IT requires for development, testing and maintenance. Decision makers require this understanding to identify how data quality initiatives or data latency requirements will affect service level agreements as well as the integrity and value of the information and analytics being provided.

This article, a continuation of Part 1 of this series, discusses these concepts and looks at the importance of focusing on the customer, the benefits of developing broad-reaching BI initiatives, and how to tie these to project success.

Understanding Architecture and Associated Options
In order to build a strong BI infrastructure, the organization should understand what options exist and how each would affect the overall environment. Selecting the right solution means evaluating what offerings exist and how each would benefit the organization. This means looking at things such as:

•    Current infrastructure – what currently exists in house, potential integration requirements, and whether other departments are using BI to identify whether another data warehouse already exists. Also, depending upon the source data and solutions being used in house or being provided to the company, certain data warehouse options might be more feasible. For instance, if a business relies heavily on their SAP data and most of their information resides in SAP systems, it might make more sense to look at SAP when evaluating a new data warehouse. Similarly, many organizations gravitate toward Microsoft SQL Server because of the developer expertise they already have in house. Although relying on what currently exists to drive future software and hardware decisions may be limiting, it also makes sure that IT considerations are taken into account in advance and are not overlooked during the evaluation process.

•    Data warehouse options – these include the types of databases available, such as appliances, analytical databases, traditional data warehouses and hosted solutions. Each comes with its own value proposition, benefits and challenges. Overall choice depends on internal resources, budget, data volumes, number of data sources, required analytics, general purpose and the like. Even within organizations that are similar in size and structure, their data warehousing requirements will differ if their goals are different.

Some considerations for organizations include understanding the difference between disparate offerings, and looking at how hardware and software fit together (i.e., selecting an appliance that is plug and play versus an analytical database that requires separate hardware and software components). Smaller businesses may find appliances beneficial because less internal resources are required to develop and maintain the data warehouse, whereas in enterprise organizations with large IT departments, looking at broader data warehousing options becomes more feasible.

•    Pricing and licensing – overall, pricing varies greatly within the world of data warehousing. Price considerations include how much data, how many data sources, general usage, etc. Depending upon the model used, there may be licensing and support costs as well. In many cases, hosted solutions offer subscription-based pricing that may depend upon the volume of data being housed. This means that if data volumes increase exponentially, costs may as well. Traditional data warehouses require hardware and software that will have to be upgraded over time as BI expands and the infrastructure gains more complexity. Therefore, what might be feasible price-wise in the short term might not be in the long term.

•    Integration and data management – Data management is becoming integrated with business intelligence. Data integration and getting cleansed information into the data warehouse has always been a key ingredient to a successful BI initiative. Now, areas such as master data management, data quality, data governance, and data profiling are becoming integrated within BI and data warehousing offerings to provide organizations with the tools required to integrate advanced data integration and data management within their data warehouses. Developing a holistic approach to data management means that businesses are better able to manage customer experience, reduce redundant marketing costs and work with suppliers. Although not all organizations require data management aspects within their data warehousing environment, areas such as data integration, data cleansing and continual data quality cannot be overlooked. Without data that is valid, the information being interpreted won’t provide the desired value. Consequently, companies do have to identify how to ensure data quality over time and, in many cases, this falls outside the realm of standard data warehousing practices and requires additional time and money.

Identifying and Focusing on the Needs of the Consumer
Defining the customer is no easy task. Each department within the organization may define a customer differently. For sales and marketing, a customer is the consumer, for IT the customer represents other departments within the organization, for those in supply chain management a customer may be partners or suppliers, and within a BI initiative the customer will be the consumer of the solution. For BI to truly be successful each customer has to be taken into account.

Usability and interactivity is essential to today’s successful BI solutions. Businesses want end users to interact with data, create their own dashboards, and make decisions and collaborate based on the information gathered. The question becomes whether users are in the position to interact with BI on a deeper or more superficial level. The technical know how and the role of the business user should define how BI is developed and the level of interactivity required and provided. Because different people in the organization have varying skill sets, the solution developed should be tailored to the varying needs of BI consumers while taking into account that skill sets will improve over time.

Additionally, depending upon the end goal, external customers should also be taken into consideration. To provide better customer service, to identify sales patterns, or to increase supply chain efficiencies requires an understanding of the needs of the customer. Whether this means collecting specific data or developing customer-focused dashboards for collaboration, BI should be synonymous with an increase in customer visibility and satisfaction.

Enterprise-Class Business Intelligence 
Even though the goal of business intelligence will differ from organization to organization and from department to department, for business intelligence to be successful, businesses should look at how all areas within the organization can benefit from its use. Even though Part 1 of this series focused on starting small and making sure that expectations are set in the beginning, the reality is that many enterprises have standardized on a single BI platform or have BI solutions within their company and want to ensure that future expansions within the organization fall within the current offerings being used. In some ways, this is positive because a BI foundation already exists; but in other ways, this limits what can be achieved because expansions are limited to what is currently available. In these situations, BI projects may not achieve effective end goals because the level of interactivity and feature and functionality sets of current BI solutions are much more robust in comparison with solutions and infrastructures designed several years ago.

An organization should not limit its solution scope to those solutions already being used within the company. Any BI project should be evaluated on its individual merits. Even though it may seem sensible to fall into an easy solution by choosing what is already available, organizations should remember that even if something is lower in cost and easier to expand versus starting new, that does not mean that the solution already in use is the best answer to the business pain.

Conclusion
The topics discussed in this series are not the only ways to limit project failures. But the value they provide interrelates as organizations are forced to validate solutions and IT projects in relation to time and budget. By developing a scope and setting expectations, and by understanding the infrastructure options, the importance of the customer, and how to create an organization-wide approach to BI without limiting BI use to what already exists, businesses can develop solutions based on addressing business pains and gaining overall information availability to make more informed decisions.

Published by B-Eye-Network.com, August 10 2010

Written by Lindsay Wise
Lyndsay is the President and Founder of WiseAnalytics, an independent analyst firm specializing in business intelligence, master data management and unstructured data. For more than seven years, she has assisted clients in business systems analysis, software selection and implementation of enterprise applications. Lyndsay conducts regular research studies, consults, writes articles and speaks about improving the value of business intelligence within organizations.