Understanding what ‘What-if?’ is and how critical it has become

I just finished a customer demonstration of our software and heard the typical question mostly asked in demos: “Do you handle “What-if” scenarios? I answered with a “Yes,” and then described the method we use and went to that area of the application to show how it works (we are pretty proud of the methods we use). At that point, the customer interrupted and said they were good, and I didn’t need to show them any more.

I always find this question particularly frustrating, because the question should not be “Do we handle it?,” but, “How do you handle it?,” or “Is it easy enough to use that anyone in my finance group can use it?,” or “I have a ‘What if’ example I’d like to see how you would do in your application.” “Do you handle it?” is opened to too much interpretation. Put another way, I can successfully use a screwdriver as a chisel, but it is not the correct tool for the job.

There are many methods of performing “What if” analysis.

Many companies take their current Excel models, copy them, make some high-level formula changes, hope they didn’t break any links, hope they only have to change the driver values in one place, and hope when they copy a template it doesn’t break everything. This method typically means you have a “budget guru” who understands and owns the Excel model. The good news is it works almost okay, as long as the budget guru isn’t sick. It also means you can’t ask for to many “What ifs” at a time because replicating all the Excel models gets very confusing.

Other companies may have purchased sophisticated and expensive planning tools and their budget guru has become the “calc script guru,” or the “PIM guru,” or the “stored procedure guru.” The net-net is the company still has limited ability to do “What if” scenarios with only one or two people. Worse yet, the sophisticated software could give you all kinds of flexibility (and breakability) by allowing you to handle it in Excel. All a person needs to do is “lock and send” or upload the information from the Excel template back to the database (hmmm… is database a finance term?). In this situation, if a single driver changes across a hundred entities, your guru gets to open each Excel file, F9 (calc) the spreadsheets, and save them back to the database, while not falling asleep, remembering where they left off, and pushing people away when they say, “When you are done running it with the cost of gas at $4.00 per gallon, I’d now like you to run it at $1.86.”

The real problem is, “What if” analysis happens at both a macro level within the planning model and at the micro level, and a tool needs to be able to address both of those levels of planning. Examples of a macro level “What if” include:
-· What if we buy one of our competitors?
-· What happens if two of our largest customers merge?
What if Lehman Brothers goes belly up while holding our commercial paper? (Oops, sorry, that may be too close to home for some of us.)
What if we take our travel expense budget down by 20%? What problems will that create?

Examples of micro level “What if” scenarios include:
-· What if the price of a barrel of oil goes to $40?
-· What if we negotiate the labor contracts to a base $18 per hour rather than $18.50?
-· What if the shipping rates for region four go up?

The key is to provide a user friendly method (or methods) that does not require scripting or tedious Excel management, and that can handle both the macro and the micro “What if” scenarios.

I’d propose you need two methods, actually. One to handle the macro “what if” analysis, and another to handle the micro “what if” analysis. (And, of course, there is a gray area in between). To nail it down even further:

-· Macro level “What if” requires procedural modeling. That is the ability to model using business rules that can be applied to the entire model or just portions of the model (Host Business Rules).
-· Micro level “What if” requires discrete modeling. Discrete modeling emulates the type of calculations you do in a spreadsheet (Host Simulation Engine).

The reason for the two methods is many calculations can be cumbersome to construct in a procedural/rules-based modeling language, but can be expressed rather easily in a spreadsheet model.

As the speed of business increases with the economic turmoil, it is clear that “What if” analysis and alternate scenarios will become more frequent, and the required precision will go up. Companies can’t plan to survive and thrive in this economy with old planning and modeling tools.