I spend a lot of time at research gatherings, where IT executives join with academics and experts to share information on running organizations better. It seems like I keep hearing the same complaints: "There's no predictability in my IT area." "Each year, more and more money goes into maintenance and operations, but there's still new work to be done." "We never know a project's on the rocks until it's too late to fix it."

One of the biggest IT-related questions facing executive teams today is whether or not their investment in information technology actually delivers business value. At a macro-level, this question has long since been resolved. Properly implemented, IT can and does benefit the implementing firm. The problem is that many managers have trouble believing this fact when it comes to the specific projects with which they have dealt.

Metrics is a people business. Having spent more than 15 years in the metrics field, this concept has reinforced itself with every engagement I've undertaken, first as a project leader within large companies and later as a managing partner in a private consulting and training firm. Measurement may initially seem to be about benchmarks, trends, and data, but what comes first is getting the data. And to do that, you have to be with and talk to people.

Software metrics is one of those great ideas that just never seems to take hold. It certainly has vociferous advocates, and it is mandatory for Level 2 certification on the Software Engineering Institute's Capability Maturity Model. Software metrics is available as dashboards, scorecards, and instrument panels. There are plenty of qualified consultants ready to help would-be users, but it just hasn't caught on.

The area of measurement is a fascinating one because everyone thinks they measure lots of things, yet almost no one does. We think we know what we have, who works for us, their skill sets, the applications they use, how happy our customers are, and the rate at which we're really growing. But, most often, we don't.

In fact, most organizations have barely inventoried their assets, their business processes, or the business and technology outcomes that should matter the most.

It's time we at CBR step back from benchmarking specific topics and industries to consider the process of performance measurement and benchmarking itself. How should you use benchmarking information? What information should you gather about your company's own internal operations? In general, how should you gather and use performance data? This issue of CBR considers these questions.

"System" is a word frequently spoken by anyone working with software. But few of those who speak the word realize that "system" denotes a well-defined concept of great importance to the success of software development. Even fewer have a thorough understanding of the system concept and what it implies about how we should manage our businesses and the software that automates them.

Every day we struggle to get to work via a transportation system, are kept informed by a news system, and put up with the antics of the global weather system. As developers, we build software systems that are part of an enterprise information system that is part of a business (system) that, in turn, is part of a supply chain (system) and the global economic system.

Integration remains one of the most important and difficult issues facing IT organizations today. Whether the integration is internal using enterprise application integration (EAI) or external using business-to-business integration (B2Bi), estimates are that at least 40% of IT budgets are directed to integration infrastructure. Multitiered computing, including Web-based applications, is quickly enhancing enterprise IT systems and provides the best hope for efficient and cost-effective integration.

IT organizations today are tasked with a variety of development projects ranging from new Web-based applications to business-to-business integration (B2Bi). In all of these projects, integration of existing systems plays an important role. Traditional enterprise application integration (EAI) has delivered on connecting applications together, but it has failed in other major requirements for IT development, namely: cost (development and maintenance), time to market, flexibility, extensibility, reusability, and return on investment (ROI).