INFORMATION TECHNOLOGY

Solving the complexity challenge, sort of (Part 3 of 3)

March 24, 2008

Editor's Note: The following is the third of three excerpts from Business Intelligence and Performance Management, a new Analytics Report produced by InformationWeek Analytics. In this excerpt, author Erik Pieczowski explores the impact of BI deployments on IT infrastructures and companies' use of BI capabilities.

Download the PDF

According to the survey, companies find BI deployments extremely complex.

Considerations such as how best to approach the creation of new data structures to deliver BI are only the tip of the iceberg.

The majority of respondents’ organizations are addressing this issue with data warehouse technologies. (See Figure 1.)

Figure 1: Database Architectures. Click to close.

Figure 1: Database Architectures. Click to enlarge.

But once data structure decisions have been made, the complexities mount.

Data extraction, transformation, and loading activities are in play and analytical tools must be selected.

For many, setting up information for BI processing alone can be a daunting task.

The world of data extraction, transformation, cleansing, and loading is incredibly complex.

Furthermore, at least half of the respondents indicated they need to access multiple types of data sources.

This requires significant time, energy, and a detailed knowledge of the data held in existing systems.

Issues such as the size of the dataset, multiple formats of duplicate data, and data constantly being added or updated are just some of challenges that organizations need to overcome.

The companies already using BI analytical tools indicate they have made inroads in addressing these concerns, to some degree, as shown in Figure 2.

Figure 2: Database Architectures. Click to close.

Figure 2: Analytic Tools in Use. Click to enlarge.

A large percentage of respondents currently using data mining as an analytical tool show maturity in BI deployments.

Data mining is not a generic tool; there are various algorithms available in data mining tools that help businesses think outside the box.

These tools are typically customized for a particular business, and they often identify and extract information that would otherwise be lost.

However, respondents still consider analytical tool sets to be immature, and stability is noted as a concern.

At least half of the respondents indicated they need to access multiple types of data sources.

It seems that organizations that are “once bitten” are “twice shy.”

Decision-makers find product stability a major concern when considering a BI offering.

Somewhat surprising is that more than half of the respondents say that their BI deployments have had minimal impact on their infrastructure.

This indicates that less than 10% of their infrastructure required an upgrade.

Considering all the complexities of deployment, and the size of the adopters organizations, respondents’ may have already made the investment in infrastructure typically required by a BI deployment.

It’s also possible that organizations have not undertaken capacity-planning efforts to capture sizing estimates.

Whatever the reason, a BI deployment is complex and requires a high level of commitment from the IT organization to be successful.

Summary

Successful BI implementations involve far more than technology.

Their ultimate goal is identifying and improving efficiencies to achieve strategic business advantage, but meeting that goal requires high-level buy-in.

Our survey results showed that BI efforts, whether initiated at a departmental or organizational level, soon require a commitment from the enterprise.

The top three barriers to BI adoption are cost, complexity, and the inability to demonstrate benefit or ROI to internal stakeholders.

Respondents’ healthy skepticism about tool sets and management techniques offers an opportunity and incentive for organizations adopting BI to consider plans carefully and appreciate the benefits that BI can bring at a strategic level.