If you made a list of common adult fantasies, total control over enterprise-wide business processes probably wouldn’t be on it. Nevertheless, plenty of thought has gone into meeting that goal. The idea of an all-seeing, all-knowing, all-doing system appeals to that little bit of megalomaniac in each of us. Depending on your perspective, it is either the culmination of previous trends to automate and integrate ever-larger portions of business operations, or a desperate last chance to control madly proliferating chaos. Either way, it implies a system with three broad functions:

• gather information. This seems the simplest goal, since it only involves reading information already assembled elsewhere. But extracting and consolidating data from numerous independent source systems is traditionally the largest job in building a business management system. The good news is that, in most major businesses today, some of the work will already have been done and accessible in a data warehouse. Even if the work isn’t complete, how to do it is widely understood.
  
  
• make decisions. Of course, making decisions is easy: it’s making the right decisions that’s hard. In business, the right decisions require a model of how the business works so the full consequences of any decision are clear. This implies both a functional view of how different parts of the business are related, and a temporal view of the results over time. Such models can be built at different levels of detail, and there is considerable art in determining the correct level for a given situation. Picking the right decision also requires some way to identify the plausible options-you can’t just model every potential choice-and a way to compare expected results of one choice versus another.
  
  
• execute. Finally, the decisions must be transmitted back to operational systems for execution. This certainly sounds challenging, but-according to people who’ve done it-is not as hard as it seems. Most modern operational systems have documented interfaces to accept external instructions or can otherwise react to decisions reached elsewhere. Even if a direct, automated interface is not practical, operational systems increasingly let users change business rules without involving a computer programmer. Although older systems are more rigid, workarounds can usually be found.

Since information gathering functions are widely available and execution is fairly straightforward, developers of business management systems tend to focus their efforts on decision making. This includes planning, budgeting, forecasting, analysis and reporting in addition to modeling, although models provide the structure on which these other functions are built. Therefore any look at a business management system must start with a close examination of its modeling features.

SymphonyRPM (SymphonyRPM, Inc, 650-935-9400, www.symphonyrpm.com) offers an impressively powerful modeling engine. It uses analytical database technology originally developed by WhiteLight Systems and acquired in 2002 by SymphonyRPM. The WhiteLight database was noted for its ability to handle very large sets of multidimensional data.

No surprise, then, that SymphonyRPM takes a multidimensional approach to business modeling. After connecting source data to the analytical engine, users map the inputs to dimensions, measures and hierarchies. Models are then built by constructing rules that define calculations using the mapped data.

The power of the models lies in these rules. They can mix multiple versions of the same data measure, such as historical and planned values, and resolve differences such as U.S. vs. U.K. product codes. Rules can work backwards from total values to components: for example, distributing limited inventory across markets in fixed proportions. They can also link values across models, using the output of one model as input to another. The system automatically verifies the integrity of complex rules sets, checks that required data is available, and executes related rules in the proper sequence.

Additional power comes from the analytic engine. It supports difficult time calculations, such as period-to-period comparisons and custom time periods, in addition to advanced math and statistical functions. It can apply multiple hierarchies to the same measure, such as organizing products by line, supplier and customer segment. The database is highly scalable, allowing models with up to 250 dimensions and hundreds of millions of cells.

If you’ve ever done serious business modeling, you know this is Good Stuff.

Other portions of the system are similarly powerful, if less unusual. Planning functions capture budgets and updated forecasts. Multidimensional data analysis tools evaluate model results, compare these to each other and actuals, and drill down for more details. Collaboration features distribute information for review, comment and approval. Users can define personalized home pages, dashboards and aggregates and receive automated alerts when measures exceed defined limits. The system can write results back to other systems, including operational databases, to help execute any decisions its users agree upon.

SymphonyRPM does have its limits. It is for business performance management, not business process management, meaning it does not control operational decisions directly. Rather, managers model different decisions and then decide which to implement. And it is not something to try at home: models are built by SymphonyRPM experts and delivered to clients who can make limited changes to parameters but generally not to the model logic itself. In fact, although SymphonyRPM does sell software and implementation services directly to individual clients, its main strategy is to help other firms develop applications they can then sell to their own customers. SymphonyRPM-based applications are already available from SunGard for Basel II reporting, profitability analysis and performance management at banks, and from market data vendor IRI (also owned by Symphony Technology Group) for more than a dozen marketing management functions. Custom-developed SymphonyRPM applications range in cost from $200,000 to over $1 million.

In short, SymphonyRPM is a yacht: elegant and powerful, but something you need a crew to operate and can’t afford if you have to ask the price. But, as with yacht, you might be able to rent it for limited use and, even if you don’t, just knowing what it has gives you ideas of what to look for when buying something more suited to your budget.

*                       *                        *  

David M. Raab is president of Client X Client, a consulting and software firm specializing in customer value optimization. He can be reached at draab@clientxclient.com.

Copyright 2006 Raab Associates, Inc. Contact: info@raabassociates.com

Back to Search