Summary
In many organizations, business intelligence (BI) and data warehouse systems play a key role in the delivery of data for strategic decision-making and financial analysis. However, as these systems reach into more data sources and provide reporting, data access and analysis to a larger population of operational users, complexity has become a barrier to companies’ realizing full returns on their investments. Most BI and data warehouse systems grew up piece by piece, and this unplanned structure makes it a slow and difficult process to implement changes requested by users or to take steps to improve data quality, consistency and relevance. Extraction, transformation and loading (ETL) routines for integrating and preparing data still require manual work and custom coding; if users discover that they requested the wrong data or need to make changes to reporting parameters, usually entire ETL routines must be started over from the beginning. Inflexible BI and data warehouse architectures make it difficult for IT organizations to allow users greater independence in customizing their use of information, which our research shows is an important factor in the maturity of operational BI deployments.
Kalido, a provider of information management software, recently introduced a new product and revisions to existing products that are designed to address these difficulties. Kalido’s existing Master Data Manager (MDM), Dynamic Information Warehouse (DIW) and Universal Information Director (UID) offer alternatives to standard data warehouse modeling and ETL-based data integration. The new product, Kalido Business Information Modeler, is a graphical tool that business and data analysts use to collect user requirements and build logical or conceptual models for BI and data warehouse systems.
Assessment
In our benchmark research study on Operational BI, we found that inadequate understanding of user requirements is a major reason why deployments fail to satisfy and have to be redone. Miscommunication between IT developers and analysts on one side and business users and analysts on the other often cause BI and data warehouse models to misrepresent what users actually need. The resulting systems require significant customization to meet the real requirements, which increases the overall complexity of the system. When a system is hard for most users to work with, data analysts and technically knowledgeable users have to do more ad-hoc querying and reporting, which can make system performance less predictable and open the door to “queries from hell” that drain computing resources and reduce availability. Confronted with unsatisfactory performance, many business users will resort to loading data into their personal spreadsheets, a move that can exacerbate data quality problems.
Traditional modeling and design tools and general-purpose diagramming software have not effectively facilitated communication between business and IT about requirements. The models produced with them can be unwieldy for ordinary users to work with, hard to visualize in their entirety and difficult to share. Specialized knowledge is often required to understand the models and change them when business structures or objectives change. In practice, such models have limited the expression of business requirements to what is possible with existing physical platforms and data warehouse architectures. In Business Information Modeler, Kalido offers a graphical tool that enables business users and analysts to develop conceptual models that accurately represent their information needs and processes. The modeling tool is an outgrowth of the “generic data modeling” method that the vendor introduced with DIW. Generic data modeling encourages a clear separation of conceptual and logical business and data model layers from the underlying physical schema that maps how data is actually stored. Business Information Modeler employs “gesture-based” modeling, which allows designers and analysts to point to objects and capture business classifications, relationships and other requirements by clicking and dragging objects in different ways across the computer screen.
The tool can be set up to inform users automatically when models will not work with currently validated data structures; or, conversely, the tool can help data analysts and architects see where to make changes to existing data warehouse designs to provide the information business users require. Models generated by Business Information Modeler also can help data warehouse managers and analysts evaluate the relevance of current reports, ETL routines and other procedures, which they can then revise or eliminate. In these ways, improved models can reduce the manual work and number of iterations traditionally required to make changes or develop new BI and data warehouse systems.
Business Information Modeler is integrated with DIW, MDM and UID, which together comprise the Kalido Information Engine suite. In DIW release 8.4 and MDM, Kalido has further automated the management of master data and data warehouses, including staging tables that hold data before it is loaded into the data warehouse. Kalido MDM is used for managing data quality and consistency in multiple data warehouses, dimensional data cubes and reporting hierarchies. The UID product enables data analysts to use the business models to guide and automate the transformation of data warehouse metadata into that used by BI tools for reporting, data access and analysis.
Market Impact
In an uncertain economy, IT organizations must reduce the time required and cost of making changes to current data warehouses and to develop new systems. At the same time, many business users are frustrated by having a limited and indirect role in determining the architecture of systems whose purpose is to deliver data and information to suit their needs. Kalido’s Information Engine suite helps organizations lower costs and address information management problems that generate complexity and slow down response to change requests. These revisions to the suite strengthen Kalido’s focus on business needs for BI and data warehouse systems improvements, which is a more comprehensive approach than its earlier focus on providing technology tools for master data management and data warehouse management.
Business Information Modeler, which Kalido currently offers for free on its Web site, should attract interest from business users and analysts who would prefer to model information requirements themselves and use the models to communicate with data analysts and IT developers. This focus makes Business Information Modeler distinct from competitive products for data modeling from CA, Embarcadero, IBM, Microsoft, Oracle and other vendors.
Recommendation
Ventana Research believes that organizations seeking alternatives to traditional practices, processes and technologies for BI and data warehouse systems should evaluate Kalido Information Engine. Master data management – the focus of the Kalido MDM component of Information Engine – is an important technology and process for managing and integrating master (or “reference”) data about customers, products or other information items that physically reside in multiple systems. Kalido’s integration of MDM with DIW and UID could enable organizations to streamline distributed data and metadata management and make the development of ETL processes more efficient. We recommend that organizations that are planning to give business analysts and users greater capabilities for modeling their own information needs and collaborating on systems with IT data analysts and architects evaluate Kalido Business Information Modeler.
About the Author
David Stodder is a Research Director for Ventana Research.
