Subscribe
About

Development standards: the way to consistency

Standards need to be adhered to in a BI environment to ensure data consistency.
Ren'e Muiyser
By Ren'e Muiyser, Principal BI consultant at PBT Group.
Johannesburg, 03 Jun 2008

It does seem somewhat strange that one has to motivate to developers as to why they should adhere to standards, whereas it is rather those that want to deviate who should go through the "please explain" process.

While creativity in developing certainly has its place, the primary reason standards need to be adhered to in a business intelligence (BI) environment, is to ensure consistency in the data.

Through data warehouse architectural guidelines and standards, various benefits can and should be realised. These include greater agility, cost reduction and most importantly - quality. In most cases, the architectural recommendation should be that projects make use of the recommended nomenclature of the data warehouse architecture as part of a BI initiative.

Creating standards

Firstly, an overall observation needs to be made - when reading the standards, methodology and guidelines, which should be put in place by data warehouse architects, the emphasis is on enablement and not enforcement. The opposite, however, does sometimes apply for the relatively closed world of an enterprise solution.

Maintaining high standards, ensuring consistency and cast-iron quality, where the development teams reside within one department, makes the enforcement approach more viable. For a federated approach, however, enablement is more applicable especially if the BI departments each have their own development teams. As in this case, whenever rules are laid down by a data warehouse architectural governing group, it will create resentment; it will not be seen as fostering a "customer relationship" and be perceived as an unnecessary burden.

What is needed is for the most pragmatic and business-value-delivering data warehouse standards and practices to be created and strengthened and yes, it is and always will be a difficult task to create standards in such a way to enable solution designers to follow the preferred path. Sometimes, however, it will be necessary to create architecture standards that are made rigid and unwieldy; they read as if they are aimed at enforcing 'rules and regulations'. It is therefore important that the authors of the architectural documents have a very good understanding of data warehousing and BI to know what is indeed needed and what will work best.

In light of this, it is essential to keep the central purpose of any data warehouse standards top of mind to ensure the following are adhered to:

* Clarity and maintainability of code
* Data load and exploitation performance
* Consistency in the data
* Enablement of effective solution designs

Standards, guidelines and best practices should include the above and evolve based on the experiences and insights of the solution designers and developers.

Implementing standards

Managers and users should not be allowed to dictate what standards to use due to some bizarre notion that time and money will be saved.

Rene Muiyser is principal BI consultant at PBT Group.

Once created, it is important to implement the standards. There are cases where the data warehouse standards, methodology and guidelines are put together by a mixture of managers and users rather than solution designers and architects. The end result is usually always disastrous.

Managers and users should not be allowed to dictate what standards to use and what not to use due to some bizarre notion that time and money will be saved. They should be assisting on what the data warehouse needs to deliver and not how to go about it.

Furthermore, there will always be people resistant to change and those who find it difficult to embrace new technology that can assist their current BI environment in terms of better performance, quicker enhancements and lower maintenance cost.

In today's fast-moving world, data is growing at a very fast rate. BI is moving at a phenomenal pace - from just being seen as a reporting solution to a strategic operational BI solution.

The key driver behind this is the business requirement to speed up and improve the organisation's decision-making ability. This means the operational data has to move as quickly and accurately as possible into the data warehouse. However, this can only be done by having the finest and most recent technology available, including the application of best standards, practices and guidelines.

Organisations should not be afraid to take the time to retrofit new and improved standards into existing programs. In fact, it should be seen as an opportunity and investment. The business community should also see it as an opportunity to increase return on investment and decrease total cost of ownership.

Given the benefits of proper development standards and the need to improve decision-making, following such standards should be seen as an inspiration to successful BI and not a limitation.

* Ren'e Muiyser is principal BI consultant at PBT Group.

Share