Lee Merkhofer Consulting Priority Systems

Technical Terms Used in Project Portfolio Management (Continued)






























Also referred to as decision objective, an explicit statement of a desired goal of implementing a decision, such as a decision whether to conduct a project. The objectives of an organization represent the desires and interests of the organization's senior decision makers; they indicate the direction that those decision makers want the organization to move. Decision analysts recommend that objectives be SMART—Specific, Measurable, Attainable, Relevant, and Time Based. Such objectives provide a basis for defining attributes and performance measures as required by some multi-criteria analysis methods, including multi-attribute utility analysis. To be technically complete, the specification of a decision objective requires specifying the object of value, its context, and the direction of preference. For example, "Increase (direction of preference) customer satisfaction (object of value) regarding our company's products (context)."

objective function

A mathematical statement in equation form that defines a dependent variable (output) which is to be maximized or minimized through the assignment of optimal values to independent variables (equation inputs).

objectives hierarchy

A graphic display of the objectives for a decision structured as a hierarchy. Objectives in the uppermost levels of the diagram reflect broad or overarching values. Progress towards those objectives is achieved by meeting lower-level, subordinate objectives. The process of identifying and structuring objectives into an objectives hierarchy typically leads to an improved understanding of what one should care about in a decision context. Creating an objectives hierarchy is also a useful initial step for creating a decision model as recommended by formal decision framing processes.

Objectives hierarchies are constructed "from the top down," beginning with the specification of the overall fundamental objective. Typically, the overall fundamental objective is relatively easy to identify—it defines the reason for interest in the decision situation. For example, my tutorial on prioritization presents an objectives hierarchy for a decision of what to pack for a vacation. The overall fundamental objective is specified to be "maximize enjoyment."

Once the overall fundamental objective has been identified, it can be related to more specific objectives that explain what is required to achieve the overall objective. The more specific objectives can often be obtained by asking "What do you mean by the higher-level objective? Or, "What, more precisely, needs to be achieved to achieve the higher-level objective?" The resulting lower-level objectives can be further related to more detailed objectives, thereby building the hierarchical structure. Oftentimes, as in the example below, the objectives in the hierarchy can be structured into different types or categories of objectives.

Example objectives hierarchy (for a utility)

An example objectives hierarchy (for a public utility)

Sometimes, the process selected for structuring objectives does not impose any formal requirements for how objectives are defined. When this is the case, the resulting structure is likely to contain a mix of objectives, goals, constraints, and metrics. Also, both fundamental or ends objectives and means objectives (objectives that specify a particular means by which the desired end might be achieved) may appear in the same structure. Objectives hierarchies with such characteristics are often referred to as value trees or value maps. Value maps are not very useful for creating decision models because they can't easily be converted into quantitative equations for measuring project value.

However, if objectives are defined and structured so as to conform to certain principles, including being fundamental, non redundant and non-overlapping, well-defined, measurable, and preferentially independent, the resulting objectives hierarchy provides a basis for constructing a decision model in accordance with the requirements of multi-attribute utility analysis. In other words, objectives hierarchies (especially when used with influence diagrams) provide a method for defining performance measures and other metrics for assessing alternatives and for deriving appropriate equations for combining the metrics into a measure of the value to be derived from each available option (i.e., obtaining appropriate aggregation equations).

Some project management software tools and a few project portfolio management tools include aids for creating objectives hierarchies.

onsite software

Software that is installed and runs on computers on the premises (in the building) of the person or organization using the software. Onsite software is also referred to as on-premises software or “shrink wrap” software. The onsite approach to deploying and using business software was most common until about ten years ago, when software running at a remote location became widely available. The alternative deployment approach uses the Internet to remove the need for the user to install or maintain any software on premises.

open source

Relates to computer software for which the source code is freely available. Project portfolio management tools that are open source are generally licensed and made available so as to allow modifications and redistribution of its source code.

operating and maintenance costs (o&m)

The ongoing cash operation, maintenance and administrative costs needed to secure the benefits of a project or asset

opportunity cost

A cost attributed to the use of a scarce resource. The opportunity cost is the value that could be obtained if the resource were employed in the best alternative usage. Thus, it is the value of the opportunity foregone if the resource is used in a proposed application. Opportunity cost is often important to determining the true costs of projects. For example, just because the organization's cost accounting system might not charge a project for using certain resources, there is typically a cost in the sense that, were it not for the project, those resources would be used for some other purpose viewed as useful to the organization.


As used here in the context of project portfolio management (PPM), optimal means the project portfolio providing the greatest possible value to the organization subject to applicable constraints (e.g., limits on available capital and other resources). Obviously, identifying optimal portfolios requires formal analysis, and this is (or ought to be) one of the primary goals of tools intended to support PPM.


A mathematical process of identifying from an allowed set of possibilities the choice or choices that are in some defined sense best. In the case of project portfolio management, the goal is to select from the set of feasible projects the best subset without exceeding the constraints that limit the projects that can be conducted. For example, portfolio optimization might seek the set of projects that collectively provides the greatest total value to the organization and that can be conducted given the constraint on available budget. As described in the paper on mathematical methods, various mathematical techniques may be used to solve portfolio optimization problems.

Compared to prioritization, optimization is a more general and powerful tool for selecting projects. Prioritization seeks to rank or list projects based on some measure of attractiveness assigned or calculated for each project (e.g., the project's benefit-to-cost ratio). However, if there are multiple constraints that affect project selection or dependencies among projects, priorities are not unique (e.g., the incremental value gained from adding a project to the portfolio compared to the incremental cost depends on the other projects that are included within the portfolio and also on the impact on ability to achieve the various constraints that may be binding). Optimization takes a more holistic approach, it merely tells you, based on the constraints you have set, whether or not individual projects are in or out of the optimal project set. There is no individual measure of project attractiveness; the contribution of individual investments is measured within the bundle. While prioritization can, in some cases, approximate an optimal solution, a truly optimized result will always identify the most valuable portfolio. See the comparison of efficient frontiers versus ranking curves for more discussion.

optimization engine

Also called optimizer or solver, an algorithm for performing optimization. The term most frequently refers to software or the component of a software program that automates the solution to an optimization problem. Users "run" the optimization engine to find optimal solutions, and they may re-run the engine multiple times to see how the solution changes based on specifying different objective functions or constraints.

Optimization engines are distinguished by the nature of the mathematical problems that they are designed to solve, for example, linear programming, quadratic programming, etc. Some project portfolio management (ppm) tools are advertised as containing an optimization engine. The component so referred to might be code written specifically for the ppm tool, or an optimizer "plug in" provided by a third-party suppler such as FICO's "Xpress" or OptQuest's "OptFolio".

option value

Also called real option value, value attributed to the flexibility that an organization or project has to adjust to uncertain or even unforseen opportunities or threats that may present themselves in the future. Traditional valuation methods, such as net present value, are deterministic and static; deterministic in that there is no accounting for the probabilities and consequences of the full range of possible future outcomes, and static in that (in general) the opportunities for changing operational paths cannot be or are not modelled. The greater the uncertainty and the more flexible an organization or project has, in general, the larger option value is.

The observed gap between the value of a business as calculated using traditional valuation methods and its observed market value is attributed to option value. Decision trees, multi-objective decision analysis, and real options analysis are valuation methods used to quantify option value.

ordinal utility

A measure of an individual's subjective preference or satisfaction expressed on an ordinal scale. With ordinal utility, the utility number communicates order of preference, but the difference between two utility numbers cannot be assumed to measure a difference in the level of satisfaction achieved. Contrast with cardinal utility.

organ pipe chart

A chart, provided by some project portfolio management (PPM) tools, showing the priority order for adding projects to a portfolio. The columns ("pipes") in the chart show the funding levels (the y-axis) at which the various projects (listed across the top) are recommended for funding. The organ chart is most useful in situations where the are interdependencies among projects, because it shows increments in funding at which groups of interdependent projects should be added simultaneously. Also, if the tool solves for the efficient frontier, the chart will show funding levels at which particular projects funded at higher portfolio funding levels are omitted from the portfolio to make room for other projects that utilize more of the budget so as to create the greatest possible value.

Example organ pipe chart

An example organ pipe chart

outranking (OR) method

A category of multi-criteria analysis(MCA) methods that seek alternatives that, based on paired comparisons, dominate ("outrank") other alternatives. The two most well-known outranking methods are ELECTRE and PROMETHEE.

The basic idea of outranking is as follows. Alternative Ai outranks Aj if on some significant subset of criteria, Ai performs at least as well as Aj (referred to as a concordance condition), while its worse performance with respect to the remaining criteria is "acceptable." (non-discordance condition). Every pair of alternatives is compared to determine which alternatives outrank others. Specifically, a subset of alternatives is found such that any alternative outside the subset is outranked by at least one alternative within the subset. This subset can then be considered a "shortlist" within which a good compromise alternative may be found by further consideration or by applying some other method.

In contrast to multi-criteria methods such as analytic hierarchy process (AHP) and multi-attribute utility analysis (MUA), which proceed by constructing aggregation equations intended to measure the desirability or utility of alternatives, outranking methods produce a partial ranking that may not render a single, "best" choice. An advantage, however, is that outranking does not require measuring the performance of alternatives using cardinal scales, ordinal measures are sufficient. Also, the decision maker is not required to make tradeoffs regarding performance relative to different criteria.