Mar 112008
 

The majority of organizations that have implemented or are planning to implement business performance management solutions rely on a data warehouse to support the data integration requirements of their performance management initiatives. This finding comes from a Cutter Consortium survey conducted in January 2008 of 101 end-user organizations based worldwide. The survey was designed to measure the extent that organizations are implementing business performance management and the techniques and tools they are using and the issues they are encountering.

Specifically, when asked, “Does your organization have an existing data warehouse that you are using/plan to use to support the data integration requirements of your business performance management initiative?” survey participants responded as follows:

  • “We are using/plan to use our existing data warehouse” (44%)
  • “We do not have an existing data warehouse/data integration architecture but plan to develop one ourselves to support our business performance management efforts” (19%)
  • “We do not have an existing data warehouse/data integration architecture but plan to buy a packaged business performance management application that includes a data integration architecture” (10%)
  • “We do not have an existing data warehouse/data integration architecture and do not foresee having to develop one to support our business performance management efforts” (15%)
  • “Don’t know” (12%)

That the overwhelming majority of organizations (i.e., 73%) are using or plan to use some form of data warehouse to support their performance management applications makes sense when you consider that data integration and ensuring data integrity were found to be the third most-difficult issues confronting organizations implementing business performance management initiatives, according to the survey.

This finding is important for several reasons. For one, data integration requirements for business performance management can be daunting because of the number of sources that need to be accessed/integrated in order to feed the performance analytics and KPIs. In addition, over the last several years, some have pushed the idea that the data warehouse is no longer essential for performance management and other BI-related applications. The argument is that the availability of virtual data integration brokers using real-time data capture technologies means it is no longer necessary for organizations to undertake complex (and risky) data warehouse projects.

I have long argued that the data warehouse is still essential for comprehensive BI applications (those spanning multiple departments, business units, and users, etc.). This is because, although virtual data broker technology excels when it comes to capturing and integrating real-time data, questions exist as to how well the technology supports complex data cleansing — especially when the data is drawn from multiple, heterogeneous, distributed sources. Moreover, although virtual data brokers are excellent for displaying real-time information in dashboards, scorecards, and other “œat-a-glance” style applications, they may not be able to present a historical view. Such a view is essential for adding business context to recent developments (e.g., contrasting regional sales or shipments for the current week’s marketing campaign with those from last week, last quarter, last year, and so on). It is for these reasons, among others, that I have always urged organizations to consider data warehousing and virtual data broker/real-time integration tools as compatible rather than mutually exclusive technologies.

In conclusion, the survey’s finding that the overwhelming majority of organizations implementing business performance management applications are relying on a data warehouse to support their efforts tends to throw cold water on the “no need to utilize a data warehouse” argument — at least as far as business performance management is concerned. As a result, I urge organizations that are considering foregoing the use of a data warehouse to carefully reexamine their requirements — especially when it comes to the need to ensure data integrity for performance management applications.

But this is just my opinion. I’d like to hear your view on this matter as well. Does business performance management, done correctly, require a data warehouse?

Discussion

  4 Responses to “Is a Data Warehouse Essential for Business Performance Management?”

  1. Thanks for posting the survey results.

    I would agree with your conclusions about the role of a data warehouse in the success of a cross-functional BI initiative: while not a requirement, it does make information delivery easier. Of course, this is assuming that the DW is at the right level of data granularity (some BI initiatives need to drill-back into the transactional detail, which may not be housed in the DW).

    However, Business Performance Management (the other ‘BPM’) is more than BI. While many analysts, vendors, and clients include BI as part of BPM, there is much more to BPM than reporting, analysis, data mining, etc. For example, BPM includes Budgeting, Planning & Forecasting, consolidation & statutory reporting, business modeling, Balanced Scorecards, and OLAP.

    I wonder if some of your survey respondents thought of business performance management as a synonym for BI.

    In some of the other areas of BPM, especially Planning & Consolidation, a data warehouse is not essential. Planning systems have their own repository (usually multidimensional) and may be a net contributor (data source) to a DW, but not often — variance reporting, for example, isn’t usually done out of the DW. And consolidation and statutory components usually go directly off of G/L or ERP source systems.

    So, overall, I would say DW is part of the BPM strategy and roll-out, but you can certainly start your BPM journey without one. I would also add that many of our clients do both, and eventually they meet-up, but do not have to be dependent.

  2. I agree but I am finding there is a debate regarding if Performance Management is a subset of BI or the reverse. How are the two similar yet different. I did some unscientific research asking around and published it on April 2, 2006 in my monthly DM Review.com column. I welcome thoughts.

  3. avatar

    BI with or without DW appears to be a hot topic these days. While these survey results are very comforting to the BI and DW professionals, I see or hear about almost as many BI solutions being developed without DW as I see with. If you are interested in this topic, be sure to check out the upcoming Cutter IT Journal (April issue), which is dedicated to this debate. The opinions and experiences of the article authors may surprise you!

  4. avatar

    I think a lot of the confusion surrounding business performance management and BI arises from the fact that both business performance management and “traditional” BI applications can use the same technologies (e.g., analytics, KPIs, report generation, dashboards).

    One way that separates business performance management applications from BI applications is that the former tend to utilize some form of business management methodology (the most popular is the Balanced Scorecard) as well as provide support for specific processes for optimizing performance. Methodology may be “baked into” the application in some kind of process or analytic workflow (as in the case of a focused scorecard application), or it may be referred to when analyzing results or applied during decision making. In other words, I would generalize by saying that the typical business performance management application tends to have more built-in capabilities for automating analysis and/or applying the findings to obtain a desired result than a generic BI application has.

    In reality, however, trying to differentiate business performance management from BI based on support for methodology and process is not failproof, because it’s easy to point out exceptions. For instance, most of the BI vendors offer digital dashboards and other front ends for their BI environments that can be used to display a quick summary view of some analysis (for example, sales for the recent quarter compared to other quarters). Although this is a straightforward application that probably doesn’t employ much, if any, in the way of business methodology or process support, it’s somewhat hard to say that it is not measuring business performance.

    Another way I’ve seen to differentiate business performance management from BI applications is by the kind of end users each is built to support. Business performance management applications tend to be designed primarily for managers and executives (hence the high-level summary view), followed by business and then financial analysts — in that order. (I’m not reciting conventional thinking here, but am basing my opinions on findings from various surveys we’ve conducted to determine the primary end users that organizations are choosing to support with their performance management solutions.) Conversely, BI applications tend to be used more by analysts.

    Still, I am somewhat uncomfortable with this way of classification because the possibility for many exceptions is considerable. In particular, many companies have striven over the past few years to leverage their investments in BI applications by extending their capabilities to different classes of end users.

    To be honest, I feel that the argument regarding the differences between business performance management and BI–or whether one is a subset of the other–is irrelevant. I believe that business performance management represents the ongoing evolution of BI and its blending with business PROCESS management (i.e., the “other” BPM). In reality, BI and business performance management are really overlapping application domains that rely on many of the same technologies and techniques. The key elements of business performance management involve the use of these technologies and techniques to implement a set of processes and/or analytic workflows in the form of a framework — backed up by one or more business methodologies — designed to help organizations optimize their business and operational performance. Ideally, this framework should automate the processes for measuring performance as well as for applying the findings or results to the operational systems and/or activities (including human) required to optimize business performance. However, while much desired, automation is not always essential. In short, business performance management is a process for formulating strategy and for connecting goals, metrics, and people in order to drive improved management, analysis, and decision making across the organization.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

(required)

(required)