To paraphrase the Great Bard: There is a tide of data in the affairs of men which taken at the flood leads on to fortune! The concept of the data warehouse is not new. Whether it starts out as an executive information system (EIS); or a data mining strategy; or an advanced decision support system; it essentially translates to copying data from existing transactional systems into a centralized physically separate data repository, after ensuring data integrity and reliability. Companies maintain huge aggregations of data at great cost for many years. The question is not so much about copying data per se, as much as it is about coping with the plethora of data. Different methods and different designs of data warehouses depend upon circumstances.
|“Today, the new paradigms that drive companies are business value and the ability to create competitive space”|
|Vikram R SriHari|
To resolve this the qualifiers that spring to mind are: What components are needed, and how should they be integrated? What sort of data is relevant, and at what cost should data be warehoused? How does one manage the data warehouse at a reasonable cost, and how does one weigh benefit vis-Ã -vis the cost? The uphill task ahead of businesses today is to manage divergent data, and facilitate retrieval in a cohesive and meaningful form. They have to be in a position to access new opportunities, shorten planning cycles, and protect their competitiveness.
It is critical that information access, which encompasses sales, finance and operations, be melded together to provide instantaneous user access. With data at ones fingertips, the challenge is to provide top management and analysts with the ability to summarize business trends, analyses and forecasts, which can be continuously manipulated.
The problem that needs to be addressed by a mature solution in managing the data warehouse is one of disparate data, which is of variable quality and coding, spread over several databases.
Using an OLAP (online analytical processing) system by monitoring key performance business indicators helps to spark sales opportunities and areas of profit (or lack of it). This is achieved by analyzing and identifying customers; not so much by asking the right questions, but by providing answers to questions that should have been asked.
For instance, corporate data can be crunched to highlight answers to questions such as: Which product is performing? Which product is under-performing? Is it a combination of product and region that is the cause? How did we do last month? Last quarter? Last year? Last three years?
Answers to these questions can only be supported by large volumes of data. The key is to manipulate the data to provide the bonafide user with answers to questions he should be asking. The ability to identify
your cash cows, dogs and rising stars by comparing profit against volume is critical to any organization. It is essential to ensure that the OLAP server supports the ability to drill down your multi-dimensional database into relational database detail. The details must be available in an easily accessed
manner to support the assumptions that are portrayed.
Security is another important issue, and the user must have privileges that can be controlled to enable the database to be well administered. Features such as recovery, backup and robustness are also relevant. Maintenance must be simple and should be automatic in terms of reassigning members to different parents.
The choice of analytic tools and mining tools while essential for providing the “business intelligence”, is not as complex as the ETL–extract, translate, and load; into a central repository or data warehouse. The hardware architecture must be open to enable the client to use different servers, and vice versa.
An organization that seizes control of an industry and creates future markets, is not determined by downsizing, restructuring or re-engineering alone. It is determined by the very basic fundamental of any successful business, which is information–the ability to manage it, and the ability to successfully warehouse it.
Vikram R SriHari
The author is director business systems, Coca Cola India