Product innovation has been described as the way out of today’s difficult business environment. The rate of success of development projects, in particular disruptive innovation projects remains too low.
We believe that a reason for the low success rate is the erroneous application of analysis methods designed for incremental innovation like NPV and DCF to projects with high levels of uncertainty
In this presentation we will discuss the use of @RISK and Probabilistic Decision Analysis in the management of innovation projects with high levels of uncertainty. The launch of the iPad is used as a case study
Quantifying and Forecasting Uncertainty in Innovation Project Management - Dr. Jose A. Briones
1. Jose A. Briones, Ph.D.
SpyroTek Performance Solutions
Twitter: @Brioneja
2. Agenda
Problem Definition
New Framework for Innovation
Management
Discovery Driven Planning
Reverse Income Statement
The “Flaw of Averages”
Probabilistic Decision Analysis
Case Study: Forecast of iPad Sales
www.Brioneja.com
Twitter: @Brioneja
3. Introduction
Product innovation has been described as the
way out of today‟s difficult business
environment.
The rate of success of development projects,
in particular disruptive innovation projects
remains too low.
We believe that a reason for the low success
rate is the erroneous application of analysis
methods designed for incremental innovation
like NPV and DCF to projects with high
levels of uncertainty
www.Brioneja.com
Twitter: @Brioneja
4. Background
In this Chapter of the “Beyond Stage Gate” series
we will discuss the use of Discovery Driven
Planning and Probabilistic Decision Analysis in the
management of innovation projects with high
levels of uncertainty.
Probabilistic Decision Analysis, when combined
with the right management processes like
Discovery Driven Planning is a very effective
approach to evaluate and manage the risk and
potential of innovation projects
www.Brioneja.com
Twitter: @Brioneja
5. Clayton M. Christensen
We keep rediscovering that the root reason for established
companies’ failure to innovate is that managers don’t have
good tools to help them understand markets, build brands,
find customers, select employees, organize teams, and
develop strategy”
• “Some of the tools typically used for financial analysis, and
decision making about investments, distort the value,
importance, and likelihood of success of investments in
innovation”
• “There’s a better way for management teams to grow their
companies. But they will need the courage to challenge
some of the paradigms of financial analysis and the
willingness to develop alternative methodologies”
www.Brioneja.com
Twitter: @Brioneja
6. Classical Stage-Gate
Process
Source: Product Development Institute, Inc.
Project is managed in a linear fashion
“The Stage-Gate system assumes that the proposed strategy is the right
strategy, the problem is that except in the case of incremental
innovations, the right strategy cannot be completely known in advance”
– Clayton M. Christensen
www.Brioneja.com
Twitter: @Brioneja
7. Challenge
Create a unified framework for innovation
projects/product development that
1. Provides the flexibility needed for
innovation to work
2. Still has the metrics needed for proper
measurement of progress and resource
allocation.
www.Brioneja.com
Twitter: @Brioneja
8. A New Innovation Project Categorization
Technical Uncertainty
High
Medium
Low
Low Medium High
Market Uncertainty
Source: Product Development Management Association
This categorization is based on the most critical variable for new
innovation projects: Degree of uncertainty
www.Brioneja.com
Twitter: @Brioneja
9. The Spiro-Level™ 3-D Approach to Innovation
Resources
Launch
Quadrant IV Quadrant I
Roadmap/Timeline
Level Idea
Generation
Risk Analysis 3
2 VOC
Customer
Testing 1
Time Time
Technology
Supply Chain Assessment
Analysis
Business
Value in Use Case
Analysis Regulatory
Quadrant III Prototype IP Strategy Quadrant II
Development Resources
www.Brioneja.com
Twitter: @Brioneja
12. Discovery Driven Planning
(DDP)
Process for management of innovation and
development projects with high levels of
uncertainty such as disruptive innovations, new
product/new market projects or game changers
1995: First Publication by Rita Gunther McGrath and Ian C.
MacMillan (1)
2000: Popularity increases with publication of book “The
Entrepreneurial Mindset” (2)
2008: Endorsed by Clayton M. Christensen as an alternative to
Stage Gate for high uncertainty projects (3)
2009: Publication of book Discovery Driven Growth (4)
www.Brioneja.com
Twitter: @Brioneja
13. Description
Discovery Driven Planning turns the basic
project map on its head
Discovery Driven Planning is a plan to learn, not to
show that you had all the answers when you wrote
the plan
“Learn as you go”
DDP uses four basic documents
„Reverse Income Statement‟ models the business
economics.
„Pro-Forma Operations Specs‟ defines operations
needed to run the business.
„Key Assumptions Checklist‟ identifies the business
hurdles and assumptions for the initiative.
„Milestone Planning Chart‟ specifies when the
assumption needs to be tested.
www.Brioneja.com
Twitter: @Brioneja
14. Reverse Income Statement
Instead of estimating the market demand and
calculating the potential revenue/profit, DDP
starts by defining the desired financial target
that will meet the company‟s targets and
strategy
The method then defines the set of
assumptions, technical and commercial that
MUST be proven true in order for the project
to succeed
www.Brioneja.com
Twitter: @Brioneja
15. Generic Reverse Income
Statement
Reverse Income Statement
Required Profits (MM$/yr) $20
Return On Sales 20%
Target Revenue (MM$/yr) $100
Market Share Penetration 30%
Market Size (MM$/yr) $333
Unit Price, $ $500
Unit Sales M/yr 200
Production Capacity Per Line, Units/Day 100
Production Lines Needed 6
Capital Cost of a Production Line, MM$ $1
Total Capital Investment Needed, MM$ $6
www.Brioneja.com
Twitter: @Brioneja
16. The Flaw of Averages
The Flaw of Averages
states that plans
based on the
assumption that
average conditions
will occur are usually
wrong
Can you drown in a
river that is on
average only three
feet deep?
www.Brioneja.com
Twitter: @Brioneja
17. Demand Forecast Problem
Forecast for annual demand: “Between 150,000
and 250,000 units”
Average 200,000 units.
Plant is sized for average demand estimate
Profit is also calculated using average demand
Flaw of averages: Profit WILL be lower than
forecasted
No upside to exceeding demand estimate due to plant
capacity limitations
Average becomes upper limit
www.Brioneja.com
Twitter: @Brioneja
18. Solution: Ranges and Probabilities
Input data and results are expressed as
ranges and probabilities
80% Chance of number being between 570 and
689K
Source: “The Flaw of Averages” Marketingnpv.com
www.Brioneja.com
Twitter: @Brioneja
19. Probabilistic Decision Analysis
Model uses @Risk probabilistic decision analysis
software
Monte Carlo simulation
Risk and opportunity analysis
Designed for complex projects with high levels of
uncertainty
Inputs contain high number of variables, either technical or
financial with a high degree of uncertainty, assumptions and
dependencies
○ New product development assessment
○ Capital spending decisions
○ Value chain analysis
○ Production and sales forecasting analysis
Eliminates use of “one at a time” cases
Analyzes thousands of cases simultaneously
Generates a range of outcomes
www.Brioneja.com
Outcome charts are analyzed to make decisions on direction
Twitter: @Brioneja
20. Input Ranges
Input values are entered in range format –
Width and shape of range are critical inputs
Definition of the input ranges is the most critical
step
Do not start with the typical value, start with the
range, define the shape of the function (10%, 50%,
90% probability).
There are multiple choices for the shape of
the input range:
Triangular: Most common for initial
assumptions
Normal distribution: Used when more
accurate input data is available 0.4342 0.6448
5.0% 90.0% 5.0%
PERT: When data is in form of probabilities
Gamma distribution: Good to model
pricing distributions in www.Brioneja.com
B-B cases
00
05
00
05
00
05
00
05
00
05
.3
.3
.4
.4
.5
.5
.6
.6
.7
.7
Twitter: @Brioneja
21. Reverse Income Statement
Combined with Probabilistic
Decision Analysis
Reverse Income Statement Min Target Max
Required Profits (MM$/yr) $15 $20 $25
Return On Sales 15% 20% 25%
Market Share Penetration 25% 30% 35%
Unit Price, $ $400 $500 $600
Production Capacity Per Line, Units/Day 75 100 125
Capital Cost of a Production Line, MM$ $0.8 $1.0 $1.4
Define ranges for each input
variable
www.Brioneja.com
Twitter: @Brioneja
22. Revenue vs. Market Size
Target Revenue (MM$/yr)
8.0%
$81.00
81.1%
$120.00
10.8%
Revenue range is
0.030
projected w level of
0.025
0.020 Target Revenue (MM$/yr)
probability
Minimum $64.17
0.015 Maximum $151.41
Mean $101.06
Std Dev $14.69
0.010 Values 5000
0.005
0.000
Market Size (MM$/yr)
$120
$130
$140
$150
$160
$60
$70
$80
$90
$100
$110
$254.36 $436.48
5.0% 90.0% 5.0%
0.008
0.007
0.006
Market Size (MM$/yr)
0.005
Minimum $191.09
0.004 Maximum $553.25
Mean $338.49
0.003 Std Dev $54.93
Values 5000
0.002
0.001
0.000
www.Brioneja.com
$400
$450
$500
$550
$600
$150
$200
$250
$300
$350
Twitter: @Brioneja
23. Unit Sales vs. Capital
Unit Sales M/yr
162.5 250.0
10.5% 79.6% 9.9%
Capital investment is
0.014
0.012
expressed a range vs.
0.010
Unit Sales M/yr
unit sales
0.008 Minimum 114.11
Maximum 338.25
0.006 Mean 203.46
Std Dev 34.00
Values 5000
0.004
0.002
0.000
Total Capital Investment Needed, MM$
250
300
350
100
150
200
$4.20 $8.00
6.2% 83.8% 9.9%
0.35
0.30
0.25 Tidal Capital Investment
Needed, MM$
0.20
Minimum $2.55
Maximum $12.64
0.15 Mean $6.09
Std Dev $1.40
0.10 Values 5000
0.05
0.00
www.Brioneja.com
$10
$12
$14
$2
$4
$6
$8
Twitter: @Brioneja
24. Capital Investment Sensitivity
Analysis
Total Capital Investment Needed, MM$
Regression Coefficients
Capital Cost of a Production Line, MM$ 0.50
Production Capacity Per Line, Units/Day -0.45
Return On Sales -0.45
Required Profits (MM$/yr) 0.44
Unit Price, $ -0.35
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Coefficient Value
Sensitivity analysis is a way to focus on the
variables that are most important to reduce
uncertainty
www.Brioneja.com
Twitter: @Brioneja
25. Case Study: iPad™ Income
Projection
A) Static Analysis
iPad 2010 Unit Sales (MM) 7.1
iPad 2011 Unit Sales (MM) 14.4
iPad 2012 Unit Sales (MM) 20.1
Netbook 2010 Unit Sales (MM) 43
Unit Price, $ $650
iPad Cost of Materials, $/unit $264.3
Static analysis
iPad Cost of Manufacturing, $/unit $10.5 uses projections
iPad Warranty Service, $/unit $20 prior to launch
iPad Indirect Expenses 2010 (MM$/yr) $2,100
iPad Indirect Expenses 2011 (MM$/yr) $2,500
iPad Indirect Expenses 2012 (MM$/yr) $2,700
www.Brioneja.com
Twitter: @Brioneja
31. iPad Income Projections
iPad Income 2010 (MM$/yr) iPad Income 2011 (MM$/yr)
$50 $2,000 $2,250 $12,500
38.7% 54.1% 7.2% 12.0% 77.4% 10.6%
4.5 1.2
4.0
1.0
3.5
3.0 iPad Income 2010 (MM$/yr)
0.8 iPad Income 2011 (MM$/yr)
Values x 10^-4
Values x 10^-4
2.5 Minimum -$1,869.82 Minimum -$830.65
Maximum0.6 $4,703.02 Maximum $26,029.39
2.0 Mean $431.87 Mean $6,784.18
Std Dev $1,010.77 Std Dev $4,206.06
1.5 0.4
Values 5000 Values 5000
1.0
0.2
0.5
0.0 0.0
-$5,000
$0
$5,000
$10,000
$15,000
$20,000
$25,000
$30,000
-$2,000
-$1,000
$0
$1,000
$2,000
$3,000
$4,000
$5,000
iPad Income 2012 (MM$/yr)
$5,000
20.5% 64.4%
$13,750
15.1%
Income analysis shows
1.2
1.0
high potential for losses
0.8 iPad Income 2012 (MM$/yr) 1st year, but huge profit
Values x 10^-4
Minimum $467.14
0.6 Maximum
Mean
$27,088.90
$8,893.70
upside for 3rd year
Std Dev $4,560.12
0.4
Values 5000
0.2
0.0
$0
$5,000
$10,000
$15,000
$20,000
$25,000
$30,000
www.Brioneja.com
Twitter: @Brioneja
32. iPad 2012 Income Sensitivity
Analysis
iPad Income 2012 (MM$/yr)
$5,000 $15,000
20.5% 68.3% 11.2%
1.0
0.8
iPad Income 2012 (MM$/yr)
0.6
Minimum $467.14
Maximum $27,088.90
Mean $8,893.70
0.4
Std Dev $4,560.12
Values 5000
0.2
iPad Income 2012 (MM$/yr)
Regression Coefficients
0.0
$20,000
$25,000
$30,000
$0
$5,000
$10,000
$15,000
iPad 2012 Unit Sales (MM) 0.86
Unit Price, $ 0.48
Sensitivity analysis iPad Cost of Materials, $/unit -0.17
shows where more iPad Indirect Expenses 2012 (MM$/yr) -0.05
market research is iPad Warranty Service, $/unit -0.01
needed to reduce
0.2
0.4
0.6
0.8
1.0
-0.2
0.0
uncertainty Coefficient Value
www.Brioneja.com
Twitter: @Brioneja
33. Actual iPad Sales Figures
2010 – 14.8 MM Units
Original industry forecast 7.1 MM Units
2011
Jan-Mar - 4.7 MM Units
Apr-June - 9.3 MM Units
Jul-Sep – 11.1 MM Units
○ Apple sold more iPads than Dell sold in all its PCs
together (10.6 million)
Oct-Dec 15.4 MM Units
2011 Sales - 40.5 MM Units
Original industry forecast 14.1 MM Units
www.Brioneja.com
Twitter: @Brioneja
34. Use of DDP
DDP has been successfully used by oil
and pharmaceutical companies to tackle
high uncertainty projects
More companies are adapting the
method, but there is still strong
resistance to leave Stage Gate behind.
Stage Gate can be used after the results of
DDP have defined the project and reduced
the level of uncertainty
www.Brioneja.com
Twitter: @Brioneja
35. Summary
DDP combined with Probabilistic
Decision Analysis provides discipline
and framework to the discovery process
but at the same time provides the
flexibility necessary for the project to
change direction or iterate as needed
www.Brioneja.com
Twitter: @Brioneja