## Maximizing the Value of the Portfolio

Vertex42 The Excel Nexus

Get Instant Access

The methods used to achieve this goal range from financial tools to balanced scorecard models. Each has its strengths and weaknesses. The end result of each is a rank-ordered or prioritized list of "go" and "hold" projects, with the projects at the top of the list scoring highest in terms of achieving the desired objectives: the portfolio's value in terms of that objective is thus maximized.

Rank Projects Using Their Economic Value or Net Present Value.

The simplest approach is merely to calculate the NPV (net present value) of each project on a spreadsheet. Most businesses already require the NPV and a financial spreadsheet as part of the project's business case, so the NPV number is already available for each project.

The NPV, a proxy for the economic value of the project to the business, can be used in two ways. First, go/kill decisions at gates are based on NPV. Project teams should use the minimum acceptable financial return or hurdle rate (as a percentage) for projects of this risk level as the discount rate when calculating their projects' NPVs. If the NPV is positive, the project clears the hurdle rate. So NPV is a key input to go/kill decision at gates. A best practice here is for the business's finance department to develop a standardized spreadsheet for this calculation so all project teams produce a consistently calculated NPV. Also, the finance people should develop a table of risk-adjusted discount rates for project teams to use for different risk levels of projects: low risk (such as a cost reduction project) to high risk (a genuine new product, first of its kind).

Second, at portfolio reviews, all projects are ranked according to their NPVs. The go projects are those that are at the top of the list. One continues to add projects down the list until resources run out. The result is a prioritized list of projects, which logically should maximize the NPV of the portfolio. In the example in Table 7.2-2, the top four projects are Foxtrot, Beta, Echo, and Alpha, but there is a resource limit of \$15 million in the development budget. Thus, only two projects are go: Foxtrot and Beta (they consume almost all of the \$15 million budget). The value of the portfolio is \$115 million from these two projects.

This method is fine in theory, but there are some problems. The NPV method assumes that financial projections are accurate for development projects (they usually are not); it assumes that only financial goals are important, for example, that strategic considerations are irrelevant; it ignores probabilities of success and risk (except by using risk-adjusted discount rates); and it fails to deal with constrained resources, that is, the desire to maximize the value for a limited resource commitment, or getting the most bang for the limited buck. A final objection is more subtle: the fact that NPV assumes an all-or-nothing investment decision, whereas in new product projects, the decision process is an incremental one, more like buying a series of options on a project.13

This NPV method has a number of attractive features, however. First, it requires the project team to submit a financial assessment of the project. That means they must do some research, make some fact-based projections, and think through the commercial implications of the project before development begins. Second, a discounted cash flow method is used, which is the correct way to value investments, as opposed to EBIT (earnings before interest and taxes), ROI (return on investment), or payback period. Finally, all monetary amounts are discounted to today (not just to launch date), thereby appropriately penalizing projects that are years away from launch.

Rank Projects Using the Productivity Index Based on the NPV.

Here's an important modification to the NPV ranking approach that recognizes that resources are limited. The problem is that

 PV NPV (present value of Development Commercialization (net present Ranking Project future earnings) Cost Cost value) Based on NPV Decision Alpha 30 3 5 22 4 Hold Beta 64 5 2 57 2 Go Gamma 9 2 1 6 5 Hold Delta 3 1 0.5 1.5 6 Hold Echo 50 5 3 42 3 Hold Foxtrot 66 10 2 58 1 Go

Note: All figures are in millions of dollars

Note: All figures are in millions of dollars some projects (for example, Foxtrot and Beta in Table 7.2-2) are great projects and have huge NPVs, but they consume many resources, thus making it impossible to do other less attractive but far less resource-intensive projects. Other projects, although having lower NPVs, are quite efficient: they can be done using relatively few resources. How does one decide?

Simple. The goal is to maximize the bang for buck, and the way to do this is to take the ratio of what one is trying to maximize (in this case, the NPV) divided by the constraining resource (the R&D dollars required). (This decision rule of rank order according to the ratio of what one is trying to maximize divided by the constraining resource seems to be an effective one. Simulations with a number of sets of projects show that this decision rule works very well, truly giving "maximum bang for buck.") One may choose to use R&D people or work-months or the total dollar cost remaining in the project (or even capital funds) as the constraining resource. This bang-for-buck ratio or "productivity index" is shown in column 4 in Table 7.2-3: Productivity index = NPV of the project/Total resources remaining to be spent on the project.

Now it's time to re-sort the list of projects. But first consider the constraint: the R&D spending constraint is \$15 million for new products in this business (the resources required to do all the projects in Table 7.2-3 adds up to \$26 million). To select the go projects, one simply reorders the project list, ranking projects according to the productivity index (this reordering is shown in Table 7.2-3). Then one goes down the list until out of resources. Note that column 6 shows the cumulative resource expenditure. One runs out of resources (that is, hits the \$15 million limit) after project Alpha.

The point to note here is that introducing the notion of constrained resources, which every business has, dramatically changes the ranking of projects. Compare the ranked list in Table 7.2-2 with that in Table 7.2-3. Note that Foxtrot, the number one project in Exhibit 2.7, drops off the list entirely using the productivity index in Exhibit 2.8; the resulting portfolio contains more projects; and its overall economic value is higher.

 Project NPV Development Cost Productivity Index = NPV/ Development Cost Sum of Development Costs Beta 57 5 11.4 5 Echo 42 5 8.4 10 Alpha 22 3 7.3 Limit reached Foxtrot 58 10 5.8 23 Gamma 6 2 3.0 25 Delta 1.5 1 1.5 26

Note: The Productivity Index is used to rank projects until out of resources. The horizontal line shows the limit: the \$15 million in development costs is reached. Go projects are now Beta, Echo, and Alpha. Foxtrot drops off the list. The value of the portfolio is NPV = \$ 121M from these three projects.

This NPV productivity index method yields benefits in addition to those inherent in the straight NPV approach above. By introducing the productivity index ratio, the method favors those projects that are almost completed and have little cost remaining in them (the denominator is small, hence the productivity index is high). And the method deals with resource constraints, yielding the best set of projects for a given budget or resource limit.

Introduce Risk by Using Expected Commercial Value. This method seeks to maximize the commercial value of the portfolio, subject to certain budget constraints, but introduces the notion of risks and probabilities. The expected commercial value (ECV) method determines the probability-adjusted value of each project to the corporation, namely, its expected commercial value. The calculation of the ECV, based on a decision tree analysis, considers the future stream of earnings from the project, the probabilities of both commercial success and technical success, along with both commercialization costs and development costs (see Figure 7.2-6 for the calculation and definition of terms). Because the method treats new product development investment decisions in a series of stages, the solution a close proxy for options pricing theory or real options.

The ECV can be used at gate meetings as an input to the go/kill decision, much like the NPV, except risk and probabilities are built in. For portfolio reviews, in order to arrive at a prioritized list of projects, what resources are scarce or limiting are identified, much like the NPV productivity index example above. Then the productivity index ratio is computed: what one is trying to maximize (the ECV) divided by the constraining resource. Projects are rank-ordered according to this new productivity index until the resource limit is reached. Projects at the top of the list are go, and those at the bottom (beyond the resource limit) are placed on hold. The method thus ensures the greatest bang for the buck—that the ECV is maximized for a given resources limit.

This ECV model has a number of attractive additional features. It includes probabilities and risk, which are inherent in any new

FIGURE 7*2'6 Determining the Expected Commercial Value of a Project

A model of a two-stase Investment decision process. First, Invest \$D In development, which may yield a technical success with probability PT, Then Invest \$C In commercialization, which may result In a commercial success with proablllty Pcs, If successful, the project yields an Income stream whose present value Is \$PV, More sophisticated versions of this model entail more stases than the two shown here and an array of possible outcomes from each stase.

Commercial Success

\$ECV = Expected commercial value of the project PTS = Probability of technical success

Pcs = Probability of commercial success (given technical success) D = Development costs remaining in the project C = Commercialization (launch) costs

PV = Net present value of project's future earnings (discounted to today)

Source: Cooper, Edgett, and Kleinschmidt, Portfolio Management for New Products.

product project; it recognizes that the go/kill decision process is an incremental one (the notion of purchasing options, a stage-wise decision process); and it deals with the issue of constrained resources and attempts to maximize the value of the portfolio in the light of this constraint.

Use a Simulation Financial Model for Major Projects. Another way to introduce risk and probabilities is the use of a computer-based Monte Carlo simulation model, such as @Risk. Here's how these models are used. Instead of merely imputing a point estimate for each financial variable in the spreadsheet, such as year 1 sales, year 2 sales, and so on, one inputs three estimates for each variable: a best case, a worst case, and a likely case. A probability curve (much like a bell-shaped curve) is drawn through each set of estimates. So each financial estimate (sales, costs, investment, and others) has a probability distribution.

The model begins by calculating multiple scenarios of possible financial outcomes, all based on the probability distributions. Tens of thousands of scenarios are quickly generated by the computer, each yielding a financial outcome such as the NPV. The distribution of the NPVs generated in these thousands of scenarios becomes the profit distribution—an expected NPV as well as a probability distribution of NPVs.

One can use the NPV and its distribution to help make the go/kill decision at gates, much as in the method for ranking projects using their economic value or NPV; take the expected NPV and divide by the costs remaining in the project; and rank the projects according to this probability-adjusted NPV, much as in ranking projects using the productivity index based on the NPV.

These simulation models, such as @Risk, are commercially available and relatively easy to use. But there are a few quirks or assumptions in the model that cause problems. For example, the model fails to deal with the options notion of a new product project, and it permits the generation of all-but-impossible scenarios.

Nonetheless, it's a solid method and particularly appropriate for projects that involve large capital expenditures and where probability distributions of input variables can be estimated.

Score Projects Using a Balanced Scorecard Approach. Scoring models or balanced scorecards are based on the premise that a more balanced approach to project selection is desirable—that not everything can be reduced to a single NPV or ECV metric. Thus, a variety of criteria are used to rate the project. These criteria are based on research into what makes new product projects successful, and hence are proven proxies for success and profitability.

In a scorecard system, each senior manager rates the project on a number of criteria on 1-5 or 0-10 scales. Typical criteria include:

• Strategic alignment

• Market attractiveness

• Ability to leverage core competencies

• Technical feasibility

• Reward versus risk

The scores from the various senior managers at the gate review are tallied and combined, and the project attractiveness score is computed: the weighted or unweighted addition of the item ratings. This attractiveness score is the basis for making the go/kill decision at gates, and can also be used to develop a rank-ordered list of projects for portfolio reviews. A sample scoring model for well-defined new product projects is shown in Exhibit 7.2-2. (Different scorecards with different criteria should be used for different types of projects: one scorecard for simple projects such as line extensions and modifications; another scorecard for true new products, as in Exhibit 7.2-2; and yet another scorecard for major platform projects.)

EXHIBIT 7.2-2 A Typical Balanced Scorecard for New Product Project Selection

Factor 1: Strategic Fit and Importance

• Alignment of project with our business's strategy

• Importance of project to the strategy

Factor 2: Product and Competitive Advantage

• Product delivers unique customer or user benefits

• Product offers customer/user excellent value for money

• Competitive rationale for project

• Positive customer/user feedback on product concept (concept test results)

Factor 3: Market Attractiveness

• Market growth and future potential

• Margins earned by players in this market

• Competitiveness—how tough and intense competition is

Factor 4: Core Competencies Leverage

• Project leverages our core competencies and strengths in:

Technology

Production/operations

Marketing

Distribution/sales force

Factor 5: Technical Feasibility

• Size of technical gap

• Familiarity of technology to our business

• Newness of technology (base to embryonic)

• Technical complexity

• Technical results to date (proof of concept?)

Factor 6: Financial Reward versus Risk

• Size of financial opportunity

• Productivity index

• Certainty of financial estimates

• Level of risk and ability to address risks

Projects are scored by the gatekeepers (senior management) at the gate meeting using these six factors on a scorecard (0-10 scales).

The scores are tallied, averaged across the evaluators, and displayed for discussion.

The project attractiveness score (PAS) is the weighted or unweighted addition of the scores, taken out of 100.

A PAS score of 60/100 is usually required for a go decision.

Sources: R. G. Cooper, Product Leadership: Pathways to Profitable Innovation, 2nd ed. (Reading, Mass.: Perseus Books, 2005); R. G. Cooper, S. J. Edgett, and E. J. Kleinschmidt, Portfolio Management for New Products, 2nd ed. (Reading, Mass.: Perseus Books, 2002).

Scoring models generally are praised in spite of their limited popularity. Research into project selection methods reveals that scoring models produce a strategically aligned portfolio and one that reflects the business's spending priorities; they yield effective and efficient decisions better than the financial tools outlined above; and they result in a portfolio of high-value projects.14

0 0