Project Selection

Once you have the information on the candidate projects, you can move on to portfolio selection. Three primary principles differentiate TOC portfolio selection from many conventional approaches:

• Recognition that there is (usually) an ongoing portfolio

• Ranking based on impact on the goal through throughput, inventory (investment), and operating expense

• Consideration of uncertainty and dependent events

The first and last points can be shared with some other approaches to portfolio selection, but often they are not addressed.

The TOC approach to ranking of projects can parallel conventional techniques, ranking according to the risk-adjusted return on investment (ROI): Risk ROI = ROI x (1 - R), where ROI = return on investment and R = risk factor, ranging from 0 (no risk) to 1 (maximum risk).

Companies often assume that ongoing projects have higher priority than new project proposals. Sometimes they make this assumption implicitly by considering only new projects for addition to the portfolio and not considering elimination of ongoing projects at the same time. This is a version of the sunk-cost fallacy, wherein people judge additional investments based on how much they have invested so far. Sunk costs are in fact sunk and should not influence decisions going forward. It is very hard for most people to decouple their thinking from what has been spent to date, but unless there could be some way to recover the sunk investment, it truly has no direct relevance for decisions regarding future investment. This means that evaluations of ongoing projects need consider only the future investment, not the investment that has already been committed.

It logically follows that you can consider project portfolio decisions one new project at a time.11 The standard of comparison is the portfolio of ongoing projects, not the other projects proposed, co-incidently, at the same time. It's okay to have a periodic process to dream up and propose new projects in a bunch. But when you do so, you should compare them to ongoing projects on the basis of remaining investment or latest estimated impact on operating expense. It is equally okay to propose new projects one at a time, at any time, and compare them to the current portfolio.

When evaluating a new project proposal against the existing portfolio, it can be important to update the return on investment calculations for the existing portfolio. One reason is that the investment calculation need only consider the estimated remaining investment to complete the project; that is, it should not include the sunk cost. In addition, often there is better information available on the range of investment and throughput impact as a project progresses.

The TOC approach to portfolio selection focuses on the goal of the organization. Project selection (including cancellation) should maximize the achievement of the goal over time. This means increasing throughput while minimizing increasing (or decreasing) inventory and operating expense. Since operating expenses are ongoing while increases in inventory (keeping in mind the TOC definition, which includes investment) are frequently transient, the impact on throughput per impact on operating expense is usually more important than the impact on investment. Thus, if one wished a simple ratio for project comparison for project ranking, TOC suggests:

Where

AT = Probable impact on throughput

AOE = Probable impact on operating expense

AI = Probable investment increase required by the project. This includes the estimated project cost and can be negative for projects directed at reducing work in progress (WIP) inventory.

The deltas signify that each project has an incremental effect on the company throughput, operating expense, and investment.

TOC developed from understanding variation. This understanding carries into project selection. No one can predict the future exactly. All predictions involve some (usually considerable) uncertainty. The amount of this uncertainty is important to portfolio selection. Many approaches to portfolio selection consider only the mean (or some other central tendency) of the return and investment. Such approaches miss a key element of understanding variation and uncertainty. Consider the two projects illustrated in Table 8.1-2. The projects have identical estimates of the mean return and investment.

Table 8.1-2 Investment and Return Comparison of Two Projects with Different Risk

Table 8.1-2 Investment and Return Comparison of Two Projects with Different Risk

Project

Investment

Return

Minimum

Mean

Maximum

Minimum

Mean

Maximum

A

20

30

40

40

60

80

B

10

30

50

20

60

100

Would you rank one more attractive than the other? Which one? Why?

Note that based on these predictions, project B could have a best-case ROI of 10 (maximum return/minimum investment), while the maximum ROI of project A is only 4. On the other hand, project B could have a worst-case ROI of only 0.4 (minimum return/ maximum investment); that is, it would lose money, whereas project A is predicted to at least break even in the worst case. Which you prefer can depend on many factors, including the relative size of the project compared to the financial position of your company (Can you afford the loss?) and the other projects currently in your portfolio. Substituting increase in throughput for return and increase in operating expense for investment does not materially alter the decision you have to make.

The project ranking included the impact of uncertainty or risk into a single number (ROI) for project portfolio ranking. This requires using a risk-adjusted metric. One way is to rank risk on a scale of 0 to 1, where 0 represents no risk (no uncertainty about either the investment or the return) and 1 represents a finite probability of losing your entire investment with no return.

You can estimate the risk factor a variety of ways. One way is to use a table like Table 8.1-3, estimating likely maximum and minimum impacts on throughput, operating expense, and investment. One way of estimating the risk is to compare the relative range of the variation. Table 8.1-3 illustrates an example of doing this. You can evaluate each project over a standard period of time, say five to ten years (or eight, as in the example), or you can use differing times, since ROI brings it all down to one number.

Table 8.1-3 provides a best-case and worst-case estimate for each of the elements of ROI. It then calculates the totals for the best and worst case for each element and estimates the variation (risk) for the investment and net profit. The illustrated approach uses the statistical term s, or standard deviation, as the measure of variation. I have assumed three standard deviations between the best- and worst-case estimates. Psychological studies evaluating people's ability to estimate reveal an overconfidence bias. That is,

Project Risk-Adjusted Return on

Investment

Total

Year

1

2

3

4

5

6

7

8

Best

Worst

Average

S

Revenue

Best

3,000

3,300

3,630

3,993

4,392

4,832

23,147

Worst

1,500

1,575

1,654

1,736

1,823

1,914

10,203

Raw material

Best

50

52

50

50

50

50

302

expense

Worst

100

103

106

109

113

116

647

Throughput

Best Worst

2,950 1,400

3,249 1,472

3,580 1,548

3,943 1,627

4,342 1,711

4,782 1,798

22,845

9,556

16,201

Operating

Best

50

52

53

55

56

58

323

502

expense

Worst

100

105

110

116

122

126

680

Net profit

Best Worst

2,900 1,300

3,197 1,367

3,527 1,437

3,888 1,511

4,286 1,589

4,724 1,671

22,522

8,876

15,699

4,549

Investment

Best Worst

800 2,000

600 1,000

1,400

3,000

2,200

533

Net return

21,122

5,876

13,499

4,580

ROI

16.1

2.0

7.1

Risk (s/Xbar)

0.3

Risk ROI

10.6

1.3

4.7

people tend to underestimate the range between the best and worst cases. It really does not matter much what you assume, as long as the spreadsheets you develop give the right behavior as a comparison tool between candidate projects (for example, the risk increases as the difference between best and worst case increases). You are interested in a relative ranking of projects.

Table 8.1-3 calculates risk R as the ratio of the standard deviation for the net return to the value of the net return. It calculates the standard deviation for the net return as the square root of the sum of the squares of the standard deviation of the net profit and investment. The reason is that in statistics, variances add when you add quantities. Variance is the square of the standard deviation. The main thing to understand about this is that if one of your estimates (net profit or investment) is significantly more uncertain than the other, it will dominate the risk calculated this way.

You don't have to be a statistician to use this information, but it is important to recognize that if one component of the risk is significantly larger than the other components, variation in the smaller component is even less significant than it looks. Usually the uncertainty in the project benefit is much larger than the uncertainty in the project cost. But people tend to focus on the cost because it seems more tangible and controllable.

You should use the average-risk ROI from a Table 8.1-3 calculation for each project to rank the projects for inclusion in your portfolio.

This ratio approach to risk tends to rank low-risk projects higher than high-risk projects regardless of the absolute value of potential return or loss. If your organization is risk seeking or risk averse, you may want to use a nonlinear multiplier or adjust the method of determining the risk factor.

Was this article helpful?

0 0
Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook


Post a comment