Our software platform helps business managers make assumptions about business actions so they can understand the full economic impact of their decisions. Business models, developed using this platform, have the ability to express a holistic picture of a company; however, there is one essential ingredient required – data. Our platform models need data from a variety of company sources, for without data, a thorough and effectual model cannot exist. (Note: Models, regardless of the tool in which they reside, need data.)
Most data sources are not within easy reach of IT. Generally, the most valuable data needed to build a driver-based model of a company originates in Operations. Here one can expect to find spreadsheets, as generations of Operations managers have used spreadsheets as analysis tools. Additionally, as a result of the ubiquity of Microsoft Excel, these spreadsheets have been revised and disbursed as regularly as clockwork. Despite this ‘lack of control,’ this unstructured information is likely the most critical data within the company.
For many types of analyses and decision support, spreadsheets are fine; however, this approach has limitations, because spreadsheets embed knowledge at a very low level, and while they can be shared, they do not support the system’s capture of knowledge. Nevertheless, they are in wide use and this usage is not expected to change. Microsoft has acknowledged that it’s important for organizations to perform analysis more easily (their studies show that as many as 60% of decisions are supported by ad hoc analysis, usually based in Excel); hence, they’ve built PowerPivot, a tool which allows users to query hundreds of millions of rows with an equation language.
Back to the point, how does IT support this most basic approach to business modeling while exercising some control over unstructured data?
I suggest employing a combination of constraint-based modeling and a publishing paradigm. Constraint-based modeling represents business processes, market opportunities, and financials as specified constraints. The specification can combine multiple forms of constraint representations, including graphical, symbolic, quantitative, and relational. The publishing paradigm will lend a level of control to the unstructured data. We have a series of dataset management tools that allow data ‘authors’ to originate and update data, and then modelers can configure the model to look at the ‘published’ version of the data so modeling is done with the most current data.
The dataset management tools are three-pronged. First, there is a user management/dataset segmentation component that allows administrators to manage who is originating and publishing data. Second, users can publish and update content, usually via import from a spreadsheet. Third, solution integration allows the published data to be integrated into models and planning solutions. All this functionality facilitates more efficient decision support processes which Operations managers and financial analysts appreciate.
About the Author
F. Shan McAdoo, Vice President of Technical Development Mr. McAdoo has extensive experience in technology project management and coordinates all technology development projects for River Logic. Prior to joining River Logic, Mr. McAdoo held several senior level positions at leading software companies, including Oracle Corporation. Mr. McAdoo earned his Masters of Regional Planning at the University of Massachusetts in 1991.