DPL Portfolio extends the capabilities of DPL to the level of decision-analytic portfolio analysis. With other versions of DPL (Professional and Enterprise), the focus is usually on analyzing one or more decisions for a single project, asset or business situation. With DPL Portfolio, the level of analysis is elevated. Multiple projects or assets within a portfolio are analyzed using one or more DPL decision analysis models. Then, the overall portfolio is analyzed by aggregating the individual project results and comparing projects to each other. Typical portfolio level results include risk/reward comparisons, project value ranges, capital efficiency rankings and portfolio-wide risk profiles.
Typically, portfolio analysis is appropriate in situations where funds to invest in R&D projects or other uncertain ventures are limited, and not all of the ventures or opportunities can be funded. The central question of decision-analytic portfolio analysis is: "Which subset of the many uncertain ventures in my portfolio should I pursue, and what might my overall risks and rewards be as a result of my selection?"
At the heart of every DPL Portfolio analysis is a set of DPL Influence Diagram/Decision Tree models. That is, DPL Portfolio uses the same modeling environment, concepts and methods as DPL Professional/Enterprise. DPL Portfolio contains a superset of the capabilities of DPL Professional/Enterprise. For example, DPL Portfolio includes the capability to link models to both spreadsheets and databases. DPL Portfolio also incorporates unique and powerful linking and "override" features so that the same model structures and links can be easily used for multiple projects within the portfolio.
Examples of a DPL Portfolio
A DPL Portfolio is a hierarchical representation of a set of projects or assets to be considered using portfolio analytic techniques. Examples of sets of projects might be R&D or new product development (NPD) projects, various business development activities under consideration (such as acquisitions of smaller firms), or financial instruments. Simply put, a DPL Portfolio can represent any set of investments that compete for the same pool of limited funds.
DPL Portfolio Hierarchy
The hierarchical structure of a DPL Portfolio includes three levels:
Each of these levels is described below.
Portfolio Groups
A Portfolio Group is defined by three items: a DPL Model, an Initialization Link type, and a Calculation Link type. All projects represented by a Portfolio Group have the same decision model, i.e., the same DPL Influence Diagram/Decision Tree (DPL Model, for short). These projects share the same set of decisions, sequence of decisions, key uncertainties and relationships among these decisions and uncertainties.
In addition to sharing a common DPL Model, the projects represented by a Portfolio Group all have the same initialization link type. That is, the data used to feed the DPL Model for each project is stored in the same way for all the projects.
Lastly, all the projects represented by a Portfolio Group share a calculation link type. That is, the calculations for each project's value model are executed in a similar way; for example, they may all be done in an Excel spreadsheet or a DPL program.
A DPL Portfolio can have one or more Groups.
Portfolio ElementsEach Portfolio Group can have zero or more Elements. Depending upon the Initialization Link type and Calculation Link type of the Element's parent Group, different information is specified to indicate where the initialization link data is and where the calculation link logic is found. E.g., if the Initialization and Calculation Link types are both DPL program, then the specific program that contains the data for the project represented by the Element and the specific program that contains the value model logic must be specified. An Element may represent a single project or a set of projects. For example, if the Initialization Link type is database, then a set of IDs identifying the projects in the database that are represented by the Element is specified (e.g., 1-5). Often the same value model is used to evaluate many projects in the portfolio. For example, you may have an Element of a Group which a database Initialization Link and a program Calculation Link. In this case, you might specify multiple project IDs in the Element all of which will be analyzed using the same DPL Program for the value model logic.
Portfolio ProjectsThe lowest level in the hierarchy is a Project. A Project is the unit of analysis in the Portfolio. I.e., a Project represent the R&D project, investment asset or business opportunity among which you are allocated development funds.
DPL Portfolio Attributes
To analyze, compare, and aggregate (or "roll up") a whole portfolio of projects, you need a common set of metrics or attributes across the projects in order to do so. DPL Portfolio provides you with a way to define a common set of Portfolio Attributes which are used by each DPL Model in your portfolio. DPL Portfolio requires at a minimum three Portfolio attributes as follows:
These three attributes are required and are always defined in that order. You may name the attributes whatever you like. The Objective attribute must be something that you wish to maximize (not minimize). The Investment attribute must be something that all things being equal you wish to consume less of. The underlying definition of each attribute is entirely under your control.
There is no requirement that the Objective and Investment attributes be financial attributes. For example, suppose you are a low-income housing non-profit organization whose most precious resource is volunteers: Objective could be # Persons Provided Housing, Investment could be Volunteer Days, and Success could represent the successful completion of a new house. You would be trying to maximize the number of persons provided with housing subject to constraints on volunteer time and recognizing that some housing projects may encounter obstacles that lead to failure.
In addition to the three required attributes, DPL provides two derived attributes. These derived attributes are calculated from the three required attributes. The two derived attributes are Objective given success, and Productivity Ratio. Objective given success is the expected value of the Objective attribute in the success scenarios. Productivity Ratio is typically a measure of capital efficiency, and is the ratio of the Objective attribute over the Investment attribute.
In addition to the three required attributes and the two derived, you may define as many attributes as you like (up to a maximum of 1024) to track whatever portfolio metrics are appropriate for your situation.
Versions: DPL Portfolio
See Also