Decisions

manager can look at prediction results, make hypothetical decisions and see how they affect bottom-line ROA. new decision document

Product Manager input

• must get PM interested in project
• give ideas for decisions to model
• where intial efforts should go
• provide source for existing model

Phases

One of the greatest advantages of the scenario planning application will be letting managers test decisions and see how they unfold in each scenario. This part of the project has many levels of possibilities, each improving the usefulness of the scenario planning application.

1. Default data without manager control- Initially, projected variable costs can be computed as a function of projected sales. Period costs can be directly fed from the current costs stored in the financial model.
2. Percent increases/decreases - Managers will be able to directly control the default values fed into the financial model by indicating percent increases and decreases.
3. Constraints - Constraints must be maintained to keep production and inventory levels realistic.
4. Real-world decisions - Ultimately managers should be able to make decisions in areas such as job shifts, factory shutdowns, source suppliers, and captial improvements. The "real-world" decisions would be better supported by completed source & supply, factory, and marketing models. Development of these real-world models will require the cooperation of domain experts.

Terminology

Decision is a heavily overloaded word. With this in mind, I'll use the following conventions:

Atomic Decision

A change is made to a single component of the business model. This change may propagate to other components, but all outside alterations are made at one component. A decision may also be time triggered and have a fixed duration. For example, capacity could be increased by building a new factory, but the increased capacity won't be available for several years. Another example is production could decrease during a temporary shutdown. The production rate will only go down for the shutdown period.

Decision Set

A group of atomic decisions. For example, to decrease production and to increase dealer incentives. Some atomic decisions may need to be prioritized to handle constraints and overlapping correctly.

Complete Decision Set

A decision set with all information necessary to run a simulation.

Real-world Decisions

A higher level decision. In the previous example, the decision set might be a part of a real-world decision to decrease factory inventory. Templates and editors for fine-tuning high-level decisions would be helpful.

Issues

• Default and sometimes stub data will be used until the details of the various models are filled in. The user should be able to see where this data is coming from, to see what assumptions are behind it, and to edit the data as appropriate.
• Qualitative decisions must be quantified - NEED good models for doing this
• The tool must be user-friendly. Selection "wizards" and decision templates would be invaluable as usability tools.
• Real-world decisions must be accurately quantified
• Constraints must be maintained - a factory can't produce more than its capacity
• Comparison options should be available to compare decision under several scenarios or a scenario against several decisions. In a way, selecting a scenario is making a decision.
• How to control who makes the decisions, how decision makers interact
• Real world decision models could vary greatly, but must be user friendly

Examples

• sourcing decision
• job shift changes
• factory shutdowns
• factory improvements
• number of employees

Decision scope

Most decisions (ignoring when scenarios and virtual groups are viewed as decisions)
will influence three things:

1. PINS (Cat sales)
2. Factory Production / Inventory Levels
3. Period Costs

Note that the first two will impact factory and dealer inventories and the sales side of the financial model, while the last two will impact the cost side.

• There is an immediate need to talk to potential customers of the scenario planning tool