Dan Power, Editor of DSSResources.com, conducted an email interview with Jim Franklin in March 2006.
Q1: How did you get interested in decision support and especially modeling and simulation?
Franklin's Response: I was naturally drawn to the quantitative methods used in my MBA at The College of William and Mary (i.e. I became a spreadsheet modeler). When I first saw spreadsheet simulation in 1994, my perspective on how people should evaluate data-driven decisions changed from deterministic to stochastic.
Q2: How do you define the term decision support system? What is the overlap between Monte Carlo simulation and DSS?
Franklin's Response: For me a DSS is an integrated Excel model that ties together assumptions based on data, the art of spreadsheet modeling and the power of persuasive charting. Monte Carlo simulation is a key part of the DSS toolkit because so many decisions are influenced by numbers. Very often those numbers mask uncertainty that should be unveiled by using Monte Carlo simulation.
Q3: Your definition of decision support system seems very narrow and specialized. Why have you limited the concept so much?
Franklin's Response: I limited the concept to the area of DSS where I think Crystal Ball fits in best. In terms of a broader view of decision support systems, I would say they help decision makers to use data or other information to make better decisions and solve problems. In those regards, a Microsoft Excel model can be key to a decision maker. But a spreadsheet model is not enough and that is where Monte Carlo simulation becomes a key part of the DSS toolkit.
Q4: In general, what should managers know about simulation, risk and modeling? How can and should that information be used in decision making?
Franklin's Response: Managers should know that it is essential to base decisions and decision policy on the realistic range of outcomes that could occur. For example, a CFO making a capital budget decision should base the decision on the rule "there must be a 90% probability that this project will exceed our hurdle rate" rather than just a 'business case' that shows one particular case where the project does meet the hurdle rate. It is also critical to understand the stochasic sensitivity: this tells the manager where the risks are and where to focus his/her efforts as well as how wrong he/she may be on an assumption and still achieve their target. Managers should also know that almost all MBA's today (in the US) are learning simulation and risk modeling; if the managers aren't literate in it, they will be passed by.
Q5: In general, what computerized decision support do you think managers need and want?
Franklin's Response: Managers want the answer. In our case that means simulation and optimization. Simulation gives the probabilities of outcomes occurring. With stochastic optimization, you first get clarity by segregating those elements of the decision that you control from those that you do not. You gain additional clarity by articulating your goal. For example, I may want to maximize my mean expected return, by choosing the best subset of projects that meet my resource constraints subject to the uncertainties of those projects.
Q6: Do simulation and optimization really give "the answer"? Are managers expecting too much when they have this expectation?
Franklin's Response: Simulation does not provide answers, rather it gives the user a much better understanding of the possible outcomes. Using stochastic optimization allows a user to determine the optimal solution based off the inputs and factors they have included in their model. Simulation and optimization like many other things are only as good as the data you put into it. Bad data will result in inaccurate outcomes.
Q7: Should managers be "hands-on" users of computerized decision support applications?
Franklin's Response: Yes. Because the tools have become so accessible, the managers should be able to 'play around' with the tools to get a feel for what their analysts have created.
Q8: What are the "good" and "bad" approaches to using Excel for decision support.
Franklin's Response: The good: Fast, Flexible, Transparent and Cheap.
The bad: hidden errors, lack of auditabiliy, lack of control, not scalable.
But wait...Excel 12 is promising to address some of the enterprise concerns by creating a server based version of Excel. This should great help the 'one version of the truth' issue. The limitations of Excel are why it is so important to use a tool like Crystal Ball for Monte Carlo simulation and optimization.
Q9: What new and improved decision support technologies is Decisioneering working on?
Franklin's Response: Mostly proprietary, but I can tell you that we believe it is all about making the existing technology more accessible so that it becomes a part of the analyst's and manager's daily work.
Q10: What do you see as major trends in computerized decision support? Where are we headed?
Franklin's Response: The BI players and the OLAP players before them have largely mastered slicing and dicing the past. Now they can do it in real-time and even send the reports/dashboard updates to your Blackberry, etc. The focus now is on the future: it's nice to know what the past has been, but what managers really need to know is next quarters sales, units, profits, etc. Some companies are trying to use the computer to do predictive analytics; we believe that it is more of an art than a science. In this case the artist is the analyst. His canvas is Excel, his paint is the output of the BI/OLAP/datawharehoused, mined data and his creation is his best assessment of what the future may bring.
About Jim Franklin
Prior to his appointment as CEO in 2003, Jim was Decisioneering's CFO from 1994 to 1998 and Vice President of Sales from 1998 to 1999. Jim left Decisioneering in 1999 to become cofounder of Web Family Ventures/ImagePlay, a Wireless technology that created a new channel for expectant families. From 2001 to 2003, Jim raised $20M as the CFO of Vericept Corporation, a leading provider of network security software. Jim holds a JD/MBA and BBA (Accounting) from the College of William and Mary.
Jim currently sits on several for-profit and non-profit company boards and is frequently called upon to speak at universities around the country. Jim was recently appointed President of Rockies Venture Club, the leading entrepreneurial club in the Rocky Mountain region focused on helping build better entrepreneurs. Jim has passed the CPA exam and is licensed to practice law in Pennsylvania and Colorado. Jim has competed in Ironman triathlons, but since becoming a father he now trains for marathons.
Power, D., "Jim Franklin Interview: Decision Support for Simulation, Risk and Modeling, DSSResources.com, 05/21/2006.