quantitative portfolio management · ©2014 murray r. cantor applying this approach ! getting the...

©2014 Murray R. Cantor

Quantitative Portfolio Management Murray Cantor Cutter Senior Consultant


Some things to consider n  Are you managing your projects to

get the most value from your organization? •  Do you know how much you should

invest in a given effort? •  Are you addressing uncertainty in your

business cases to balance your portfolio between innovative efforts and low hanging fruit?

n  How do you know when you are ready to deploy? •  Do you know how much financial risk

you are assuming when you deploy the offering?

•  Do you know when it makes economic sense to deploy the offering (when the expected benefits outweigh the expected expenses?

2


Strategic alignment may not answer the questions n  The staff with the strongest

opinions will prevail in getting their pet projects funded •  “My project is clearly more strategic

than yours.”

n  Difficult conversations with stakeholders such as the CFO who wants to know if the money is being spent wisely. •  CFO: “Costs are real money, benefits

are soft.”

3


Taking an economic perspective helps the stakeholders reason together

n  What is the value of our project portfolio?

n  How can I get the most value from our finite resources? •  Money •  Staff time •  Calendar time

Key question: “What is the value of an (incomplete) software or IT project?”

4


A conventional answer: $0 till shipped

n  The conventional wisdom: •  Fails to acknowledge value of work already done •  Provides no opportunity for ongoing value management

n  Can only quantify cost, not value

5

0 6 Ship date

Valu

e

If all unshipped efforts are worthless, the only discussion available is some form of “strategic alignment”

5


Another approach: Things are worth what someone (maybe you) might pay for them

n  Imagine (if you will) you could buy or sell incomplete development programs

n  The buyer would spend money now to obtain the option to invest in completing the program to receive its benefits

n  How would one reason about the fair price? •  The buyer, reasoning like an investor, to compute fair price needs

–  The costs to complete, C –  The benefits to be received, B

•  With these probabilities, one can reason about value of V=B-C

The economists call this “incomplete market reasoning”

6


For example: Consider the Airbus A350 program n  They have delivered no planes, but

they have 750 orders @ ~ $350M each (comes to ~ $260B).

n  If they were to sell the program to Boeing, what would be a fair price? •  Certainly not $0!

n  The buyer would get the right to complete the effort to get the future benefits (a kind of option).

7

n  Need to estimate: •  C, the costs of completing development and bringing to market •  B, the benefits: revenue, after delivery maintenance, etc. for the

current and future orders


The two challenges:

1.  The future costs and benefits are uncertain

2.  The benefits may be intangible

8


Dealing with Uncertainty

9


Capturing uncertain quantities

n  Sometimes you do not (or cannot) know a quantity you need to make a decision or carry out a calculation.

n  Even so, you are not completely ignorant. •  For example, I know there is

zero probability that GM will sell a trillion cars next year.

n  One can use probability distributions is to describe what we know or believe the quantity to be.

0 2 4 6 8 10

0.0

50.1

00.1

50.2

0uncertain quantity

Pro

babili

ty

The height shows the probability of the quantity being near a value. In this case, it is most probable value is near 5.

10


Capturing uncertain values probabilistically n  For an uncertain quantity, Q, one can elicit from subject matter experts:

•  High value, H (there is little chance that Q > H) •  Expected value E, (of all possible values Q = E is the most likely) •  Low value, L (there is little chance that Q < L)

n  The inputs are captured in a triangular probability distribution

L

E

H

This is common “best case, worse case, likely case” elicitation

11


One can do arithmetic on uncertain quantities

12

The result is also given probabilistically

Statistics of Sum Mean: 14.667 Median: 14.495 SD: 2.2239 Variance: 4.9459 Lower Percentile: 25.0 (13.061) Upper Percentile: 75.0 (16.191)


Example Software Project Template

13


Example Software Project

14

B - C

C

B


Example of a benefits model: Using models from flight safety engineers and aviation economists

15


This example shows the investment value is close to break even

n  What is this telling us? •  The most likely value is about $285k •  There is a 25% chance this effort would lose more than $390k •  There is a 25% chance that that this effort would yield more than $960k

n  What to explore: •  How does this investment compare to others? •  Could I make this a better investment by changing project plan?

–  Earlier market window? –  Invest more in quality, less after market costs?

16


The investment value distribution gives a view of the value and risk (uncertainty) of the investment

n  The mean of the distribution is its fair value

n  The standard deviation is a measure of its risk.

17


This provides the basis for a portfolio view

The mean

Normalized SD

A program

18


Applying this approach n  Getting the most value:

•  Ongoing portfolio management: Use emerging information to get early indication of failing efforts to “fail fast.”

•  Multiple scenarios: e.g., balance deploy date against feature set •  Instrument Lean Startup: Decide when ready to pivot

n  Manage ready to deploy: •  For example, to balance further investment in quality against less after

market expense

n  Provide economics of efforts to finance, business management (expected NPV, ROI, with amount of uncertainty)

19


Intangibles

20


Dealing with Intangibles n  A government agency, mission-based

organization, or not-for-profit is faced with how best to invest to further in its mission

n  To deal with this challenge, one starts with the questions:

•  What exactly are our goals? •  How would we know they are being

achieved?

n  With the answers, one can determine a useful measure of benefit from the investments.

•  Examples may include lives saved, failures avoided, forested acres preserved

21

Sometimes one can work with economists to ascribe an economic value to the benefit, sometimes not.


Modeling with non-monetary benefits

n  Can’t compute B-C, different ratios

n  But can compute B/C (essentially a ROI)

n  Examples: •  (lives saved)/$ •  (acres preserved)/$

22


To adopt

n  Initial Workshop •  Agree on level of detail needed (simple decision to full financial model) •  Identify first one or two investments to be modeled •  For each investment

–  Agree on cost and benefit streams –  Identify subject matter experts –  Elicit

>  Domain business models >  Initial estimates

•  Build integrated value •  Sanity check output and revise if necessary

n  Expand to more investments •  Build portfolio view

n  Over time, revise models with updated information •  Track trend of program values (well managed programs accrue value over

time)

23


Strengths of this approach

n  Integrates the knowledge, beliefs, and concerns of all the subject matter experts: •  Surfaces the assumptions •  Provides a framework for identifying and addressing risks

n  Enables delivering best case, worse case scenarios for finance

n  Provides a dispassionate framework for decision making

n  Provides an opportunity for incorporating new information to improve models as data rolls in •  Track value creation (or not).

n  Can be elaborated to include financial measures such as probabilities of NPV, RoI(s), etc.

24


Questions? [email protected]

25

quantitative portfolio management · ©2014 murray r. cantor applying this approach ! getting the...

Documents