Lee Merkhofer Consulting Priority Systems

Technical Terms Used in Project Portfolio Management (Continued)





























Kepner-Tregoe method

A structured decision aiding technique for collecting information and making decisions developed by Charles H. Kepner and Benjamin B. Tregoe in the 1960s. The approach consists of four basic steps, situation appraisal, problem analysis, option analysis, and potential problem and opportunity analysis.

key performance indicator (KPI)

A metric selected to indicate how effectively an organization is achieving its key business objectives. The hundreds of KPI's that have been suggested may be organized based on the type of organizational objective that they serve, for example, KPI's for financial objectives, customer objectives, sales objectives, process objectives, staff objectives, and so forth. To illustrate, examples of KPI's for staff objectives include employee turnover rate, percentage of those who respond to open positions that are judged to be qualified for the position, and employee satisfaction level. Organizations typically select and then use KPI's to help determine their progress toward achieving strategic and operational goals, and also to compare performance relative to other organizations in the sanme industry.

knapsack problem

A mathematical statement of the problem of selecting projects subject to a budget constraint. The name derives from the analogy to the problem of choosing items to carry in a knapsack. To illustrate, suppose there are m items, where item i has a size ci and, if selected, provides a benefit bi, for i = 1, 2,,..., m. The capacity of the knapsack is C. The goal is to select items that will collectively fit in the knapsack and provide the greatest possible benefit. The problem may be expressed mathematically as:

Eq. 1

The xi are decision variables (1 for item acceptance and 0 for rejection). These are the same equations used to describe the project portfolio capital allocation problem.

The knapsack problem is relatively difficult to solve. It has, in fact, been used as the basis for encryption. Methods for solving the knapsack problem are computationally intensive, however, various approximate methods are available that are more efficient and that can be shown to come very close to mathematically optimal solution (see Methods for Solving the Capital Allocation Problem).

What makes the knapsack problem difficult is the "0/1 assumption"—Items must be put entirely in the knapsack or not included at all. You cannot, for example, put part of a soda pop can in the knapsack. Were it not for this requirement, you could solve the problem easily using a "greedy algorithm"—Rank the items based on benefit per unit size. Take as much of the top-ranked item as you can (or enough to fill the knapsack). Then, repeat with the next ranked item until the knapsack is full.

The knapsack problem arises in any activity, including project portfolio management (PPM), that requires allocating finite resources to items that are not infinitely divisible. Thus, a relevant consideration for choosing a PPM tool is whether it provides an algorithm for solving the knapsack problem and the quality of that algorithm.



As in lean project management, lean enterprise, lean production, etc., a business practice that considers the expenditure of resources for any goal other than the creation of value for end customers (or in some uses, for the enterprise itself) wasteful, and, therefore, a target for elimination. Toyota developed and applied the concept to manufacturing in the 1990's, and the auto manufacturer's success during this era helped popularize the practice. Project portfolio management (ppm) applied with the goal of selecting value-maximizing project portfolios can be regarded as consistent with the principles of lean operation, and some ppm tool providers have used the phrase lean project portfolio management to describe their offerings.

lexicographic method

A general approach to decision making or to prioritization when multiple criteria are relevant. Specifically, the alternatives are evaluated relative to the criterion judged to be most important, and the alternative with the best performance relative to this criterion is chosen (or ranked number one), unless there are other alternatives that with regard to the most important criterion tie for first place. In that case, evaluations with repsect to the second most important criterion are considered to break the tie. If such comparisons don't resolve the tie, then the third most important criterion is consulted, and so on until one alternative emerges as a winner. There are several variations on this approach which require more than the minimal information of a strict lexicographic method. Note that contrary to most multi-criteria analysis methods, lexicographic methods are not compensatory.

life cycle portfolio matrix

Also called the product life cycle portfolio matrix and the ADL matrix, a simple tool developed in the 1980's by the professional services firm Arthur D. Little intended to help a company manage its collection of product businesses as a portfolio. The key concept is consideration of where each product is within its business life cycle.

Like other portfolio planning matrices, the ADL matrix represents a company's various businesses in a 2-dimensional matrix. In this case, the columns of the matrix represents the growth stage of the business product (embryonic, growth, mature, or aging) and the rows represents the product's competitive position in the marketplace (dominant, strong, favorable, tenable, or weak or nonviable). This results in a 4 by 5 matrix with 20 cells. The company's various product businesses are placed within the matrix, and the positions are associated with logical business strategies as shown below:

Life cycle portfolio planning Matrix

Life cycle portfolio planning matrix

The distribution and trajectory of the businesses across the matrix helps indicate whether the firm's product mix is well balanced now and in the future. For example, the company will need to maintain a continuing set of mature businesses in order to generate cash to support new embryonic and growth operations.

linear programming

A mathematical method for finding the maximum or minimum solution to a problem where the objective function is a linear combination of the decision variables, for example,

ax1 + bx2 + cx3 ...

and where some or all of those variables are subject to linear constraints, for example,

Ax1 + Bx2 + Cx3N   or   Ax1 + Bx2 + Cx3N

Linear programming has been long applied to resource allocation problems. Typically, the decision variables represent the amount of various resources allocated to various purposes and the constraints specify how much of each type of resource is available. So long as the objective function is a linear function of the amounts allocated, the solution can be found using linear programming.

The importance of linear programming derives in part from the efficiency of the algorithm, known as the Simplex Method, by which a linear program may be solved. Linear programming can handle very large numbers of variables and constraints. Some applications, for example, have involved millions of variables and hundreds of thousands of constraints.

The main disadvantage of linear programming, of course, is that it requires that the optimization be conducted based on a single, linear objective function. Project selection decisions, as well as most other decision problems, require multiple, generally non-linear, objectives to be simultaneously optimized. (Goal programming and multiobjective linear programming are variations of linear programming that attempt to account for multiple objectives.) With linear programming you cannot, for example, account for project start-up costs, efficiencies of scale, and other considerations that commonly cause the relationship between project costs and project benefits to be non-linear.

Also, linear programming assumes that any solutions within a continuum of possible values that satisfy the constraints are possible (for contrast, see integer programming). In the real world, most often you must choose whether to fund or not fund a project. A linear programming solution that told you to fund 1/5 of the project might not be very useful.

linear regression

The relation between variables in a regression analysis when the regression equation is linear, e.g., y = ax + b.


A term used on this website and by decision analysts and others to refer to probability distribution for some uncertain outcome, typically a monetary payoff. For example, if the consequences of conducting a project are uncertain, the value of those consequences will likewise be uncertain. A project selection decision model might be constructed to derive a probability distribution over the values of possible project outcomes. It might then be said that the decision to conduct the project is like paying to play a lottery with outcome probabilities as specified by the model.