Subscribe

Data warehousing: It`s the journey


Johannesburg, 13 Aug 2004

Today`s data warehouses have entered that predictable stage of technology evolution called the "love, hate relationships" phase. Despite a global understanding of data warehouses, many projects are still not delivering that critical business return on technology investment.

Alarmingly, the failure rates of data warehouse projects - depending on what analysis you`re reading - is anything between 40% to 70%, which counts for a lot of unfulfilled promise.

The logical conclusion is to blame technology. But, unfortunately for those IT sceptics out there, the problem doesn`t lie with technology but with people`s failure to think through the intended business and organisational value of their data warehousing projects.

Current data warehouse veterans realise that the value delivered from a data warehouse is removed from the value delivered by, for example, a transaction system.

It is imperative that a transaction system must be accurate and deliver sub-second response as well as availability, failure to do so can result in a company money loss.

A data warehouse, on the other hand, is an example of a decision support system as it essentially integrates information from transactional systems and even external sources to enable companies see a larger piece of the enterprise and identify new business opportunities.

In short, a data warehouse exists to support management`s decision-making processes. It is not a system that makes report generation easier or faster, or works to reconcile operational reports.

Companies have to clearly define what they require from their data warehouse, an "any big business should have one" approach is not going to cut it.

Delivering reports to users is an important facet to any data warehouse project, but this should to not be the prevailing reason for spending hundred of thousands or even millions of rands.

There must be a clear business demand for the capabilities and functions that a data warehouse provides. Defined business goals and objectives will ensure that the project team`s task is driven by the business drivers and not the latest data warehouse technology trends.

In saying this, it is still important that business should select technology that fulfils both the present and future needs of a company, otherwise it becomes a colossal waste of time.

This doesn`t mean that a company should buy every server and disk drive it will need in the future - the technology selected for the first release should be scalable and flexible enough to support future upgrades.

Although this has probably been mentioned more than once, the "big bang" approach doesn`t work. Instead, break the project into parts or releases with each going through the complete system development lifecycle of design, build, test, implement and support.

Due to its iterative nature, a data warehouse is a journey, not a destination - as a business continues to grow or change there`s always more data that can be integrated.

Lastly, remember that the success of a project will be determined by the communication and management skills of the project leaders. Marketing the data warehouse projects, keeping the stakeholders informed and managing expectations are some of the factors that will keep the project afloat.

Share

CA

Computer Associates International, Inc (NYSE:CA), the world`s largest management software company, delivers software and services across infrastructure, security, storage and lifecycle management to optimise the performance, reliability and efficiency of enterprise IT environments. Founded in 1976, CA is headquartered in Islandia, New York, and operates in more than 100 countries. For more information, please visit http://ca.com.

Editorial contacts

Willie Bezuidenhout
Computer Associates Africa
(011) 236 9111
Willie.bezuidenhout@ca.com