MIT OpenCourseWare
http://ocw.mit.edu
15.351 Managing Innovation and Entrepreneurship
Spring 2008
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
15. 351 Managing Innovation
& Entrepreneurship
Fiona Murray
MIT Sloan School of Management
Spring 2008
Class One – Technology Dynamics
AGENDA
z
z
Motivation for today’s material: Why it is important
to assess technology-driven dynamics & why it is
hard.
S-Curves
z
z
z
Defining technology dynamics
Mapping technology dynamics
Managing technology dynamics
Typical definition of an
opportunity/threat
z
z
In a business plan
In a project proposal
OPPORTUNITY
What is the problem you propose to solve?
How do you propose to solve it?
What is the potential market impact?
What is the customer "pain" that you are
attempting to address?
What is the market doing now to address
the problem?
When the opportunity is driven by a new
technological innovation, analysis should
integrate technology & market factors
Technology assessment
& choices
Markets
Technologies
Together technical
choices & market
assumptions lead to the
opportunity – the
concept design that
drives the business
“model”
Market assessment &
marketing choices
What is wrong with this…?
Typical analyses fail to examine the
dynamics of technology & market
factors
Technology assessment,
dynamics & choices
Markets
Technologies
Market assessment ,
dynamics & choices
A more robust
opportunity assessment
is clear about the
dynamics of the
proposed technology &
that of competitors & the
proposed market & that
of competitors
THIS IS HARD – WHY?
Can we forecast the dynamics of
technological change?
z
Hard because:
z
z
z
z
Predicting the future
Hard to get data
Requires expert knowledge (across domains)
Blind spots when considering others’ technology
But….
z
z
z
Wealth of historical data
Trend extrapolation
Robust heuristics – S curve
Early developments in Artificial
Hearts
35
30
*
*
Survival time weeks
25
Jarvik heart
20
*
15
10
5
0
0
5
10
15
20
25
30
35
Cumulative laboratory-years of work
Image by MIT OpenCourseWare.
Moore’s Law at Work
In 1965, Intel co-founder Gordon Moore saw the future. His prediction,
now popularly known as Moore's Law, states that the number of
transistors on a chip doubles about every two years.
Transistors
10,000,000,000
*R
*R
Dual-Core Intel Itanium 2 Processor
*R
MOORE'S LAW
*R
Intel Itanium 2 Processor
*
*
Intel Itanium Processor
R
*R
R
100,000,000
*R
Intel Pentium 4 Processor
*
*
Intel Pentium III Processor
R
R
*R
*R
*R
Intel Pentium II Processor
*
R
Intel Pentium Processor
10,000,000
1,000,000
TM
Intel486 Processor
TM
1,000,000,000
Intel386 Processor
100,000
286
8086
8008
10,000
8080
4004
1965
1970
1975
1980
1985
1990
1995
2000
2005
1,000
2010
Image by MIT OpenCourseWare.
Declining Yield Improvements for
Phthalic Anhydride Production
Selectivity yield (percent)
105
Orthoxylene
100
O
95
Naphthalene
O
90
85
O
80
50
100
150
200
250
300
Cumulative R&D effort (man-years)
Image by MIT OpenCourseWare.
Declining Yield Improvements for
Phthalic Anhydride Production
z
z
z
PA is an organic chemical molecule, a building block in processes that result in
paint thickeners and softer plastic luggage or auto upholstery. It is an important
industrial chemical now and may become more so in the future.
Raw materials: Naphthalene has more carbon in it than required, but lacks
oxygen. Orthoxylene looks like Naphthalene except that it has less carbon. A
pound of Orthoxylene gives you 1.4 pound of PA, vs. 1.2 pound for
Naphthalene. 20% improvement. Worth a lot, margins in the chemical industry
are often 10-15%. This does not mean that Orthoxylene-based PA will be
cheaper. What if Orthoxylene’s price was 20% more expensive than
Naphthalene? That was the case until early 1960s when more ortho became
available as a result of advances in oil refining. That’s when ortho caught up
with naphthalene.
Allied Chemical pursued Naphthalene in its mature stage. Scientists working for
Allied, Monsanto, Chevron, and others spent some 100 man-years of effort
between 1940 and 1958 seeking more efficient ways of making PA from
naphthalene. During that period, performance improved steadily. But from 1958
to 1972, scientists exerted an additional 70 man-years of effort and achieved
only limited progress. Eventually, progress on PA from Naphthalene stopped
completely.
S Curve – Proposed model for
dynamics of technological change
Physical limit?
Performance
Pattern:
Initially increasing then declining R&D
productivity within a given physical
“architecture”
Performance is ultimately constrained
by physical limits
E.g.:
Sailing ships & the power of the wind
Copper wire & transmission capability
Semiconductors & the speed of the electron
Foster’s
S Curve
Effort
The S-Curve
All it says is: things are going very, very slow in the
beginning, the pace quickens in the middle, and
then decelerates in the end. That’s all it says. It’s a
tool for thinking where you are strategically, it’s a
tool for asking questions, like ”what performance
measure should I plot?” It is not a magic forecasting
tool.
Breaking Down the Technology S-Curve
Position
Early-stage, Low R&D
productivity
“Riding Up” the
S-Curve
Hitting Natural
Limits
Why?
Need to Experiment; Lots of Early Failures; Building up
Knowledge about the Area; Bringing Together the “right”
capabilities and knowledge
Focusing on an overall “architecture”; focusing on
narrower and more well-defined technical challenges;
organizational commitment and incentives; leveraging prior
experience
Key physical limits determined by broad technical choices
(e.g., speed of sound; analog versus digital). The
constraints result from key architectural choices.
The Evolution of Palomar’s Products:
Laser Based Skin Treatment
Images removed due to
copyright restrictions.
120 Pounds
Product
EpiLaser™
E2000™
LightSheer™
SLP1000™
EsteLux™
MediLux™
NeoLux™
StarLux™
Lux Handpieces
Home Devices
Material
Price
Cost
$150K
$80K
$130K
$60K
$100K
$40K
$65K
$25K
$40K
$ 4K
$50K
$ 4K
$30K
$ 4K
$80K
$ 5K
$10K
$ 1K
?
?
Year
1996
1997
1998
2000
2001
2003
2003
2004
2002-4
?
Evolution of Measurement-While-Drilling
Tools S-Curve
Co
nti
nu
ou
sM
.P.
Performance = Data Transmission Rate (bit per second)
15
MWD tools are extremely
complex pieces ofCo
nti
nu
electronics.
ou
sM
Used to make directional.P. - B
PS
K
3G
surveys in real time.
Combine accelerometers &
Co to measure
magnetometers
nti
nu
ou
sM
the inclination and
azimuth
.P.
-2
of the wellbore at location,
G
and transmit
hat information
Co
nti
nu
to surface.
ou
14
13
12
11
10
9
8
7
Po 6
sit
ive
5
Ne
ga
ti
M
ud
P4
os
itiv
e
ve
M
3
ud
P2
uls
e
sM
.P.
Pu
lse
M
ud
-F
SK
3G
Physical limit: signal attenuation
-1
G
2n
dG
en
e ra
tio
n
Shallow wells only
All well conditions
Pu
lse
Dominant Design = Continuous
Mud Pulse Telemetry
1
0
1
2
3
4
5
6
7
R&D Effort (measured in Generations = +/- 3 years )
8
9
Example – Smaller Diapers
EXERCISE
z
Sketch the relevant S curves.
z
z
What are the appropriate
(technical) measures of
performance? Are there more
than one?
Where is this industry now? Are
there major growth areas or
discontinuities on the horizon?
Evolution of the Disposable Diaper highlights
Dynamics of Technology S-Curve
The 1960s – Vic Mills’ Pampers
The 1940s – Marion Donovan’s “The Boater”
Polypropylene
Composite fiber
Polyethylene
Polyacrylate
The 1980s – Harper & Harmon
Super-Absorpent
The 2000s – Another 60% smaller
Image by MIT OpenCourseWare.
The Millsian (Absorbent core) diaper faced significant
technological hurdles due to the tradeoff between size and
absorbency...
Benefits of greater
absorbency were at odds with
massive increases in size –
cost of size on the “shelf” and
on the “body” were both
important limitations to this
approach
Physical limits
Absorption
Capacity
(How many times
diaper can hold its
initial weight?
R&D Effort
Donovan
Pampers - Mills
Fluff etc.
SAP technology facilitated a long period of sustaining
innovations along the S curve – changed the absorbency
versus size tradeoff...
SAP
Physical limits
Absorption
Capacity
(How many times
diaper can hold its
initial weight?
INTRODUCTION OF A NEW
CURVE DESCRIBED AS:
• Discontinuity
• “Break point”
OR (incorrectly)
•Disruption
SAP
introduction
1974
Millsian
SAP
patents
1966
R&D Effort
Issues in using S Curves to
analyze technological dynamics
z
z
z
z
z
Progress as a result of the passage of time vs. progress as the
result of returns to effort
Do all good things come to an end?
Which parameter(s) shall I predict?
What level of aggregation – firm or industry
What level of analysis – component vs. system v. process
The S curve is best viewed as a tool
for triggering discussion, not as a
“scientific reality”
Time or Effort?
Image removed due to copyright restrictions.
Figure 2 on p. 338 in Christensen, C. M. "Exploring the Limits of the Technology S-Curve. Part I: Component Technologies." Production and Operations
Management 1, no. 4 (Fall 1992): 334-357.
Figure 2a charts the average area1 density of all
disk drive models introduced for sale by all
manufacturers in the world between 1970 and 1989.
The pace of improvement has been remarkably
steady over this period, averaging 34% per year;
with time as the horizontal metric, no S-curve
pattern of progress is yet apparent.
Figure 2b shows that what appeared in Figure 2a as a
relatively constant rate of improvement over time in area
density appears instead to be an increasing rate of
improvement per unit of engineering e@rt applied.
Source: Christensen 1992
What parameter?
Metrics of interest may change over
the technology S curve
Speed
Scanning Projection
Aligners
Step & Repeat
Aligners
Yield
S curves are probably complex landscapes of performance over a multidimensional surface – easier to plot several S curves with different parameters
Figure 5 on p. 345 in Christensen, C. M. "Exploring the Limits of the Technology S-Curve.
Part I: Component Technologies." Production and Operations Management 1, no. 4 (Fall
1992): 334-357.
Source: Christensen 1992
Firm vs. industry
S-Curve revival
Image removed due to copyright restrictions.
What level of analysis?
Component vs. Architecture
Physical limits
Technical
Metric
Physical limits
Technical
Metric
R&D Effort
Component-level:
No change in overall system
architecture – in disk drives think of
ferrite read/write heads shifting to
thin-film heads
R&D Effort
Architecture-level:
Change in the linkage of components –
14’ -> 8” -> 5.25’ disks. Or generations
of optical photolithographic alignment
equipment – contact-> proximity etc.
What level of analysis
All three “purple
boxes” would
constitute a new S
curve but only one
is labeled “radical”
High impact on architectural knowledge
Architectural
innovation
Radical
innovation
Low impact on
component knowledge
High impact on
component knowledge
Incremental
innovation
Modular
innovation
I WOULD CALL
ALL THREE
RADICAL & THEN
SPECIFY
Low impact on architectural knowledge
Image by MIT OpenCourseWare.
Adapted from: Henderson
& Clark, 1990
Technology S-Curve in practice….the
pharmaceutical industry
Using the S-Curve perspective to analyze the so-called “productivity
crisis” in the pharmaceutical industry
Fewer new drugs approved...
50
40
30
20
10
0
45,000
... despite higher costs
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
0
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
60
19
9
19 6
9
19 7
9
19 8
9
20 9
0
20 0
0
20 1
02
20
0
20 3
0
20 4
05
z
Origins of the Productivity Paradox
Widely debated by CEOs, analysts….surprising given
z
z
z
Massive investments in innovation, yet the level of new FDA
drug and biotherapeutic approvals is comparable with the
1980s
Thirty years of dramatic scientific progress (from genetics to
systems biology)
Emergence of thousands of biotech companies (> 500 public)
What might an S-curve analysis tell us?
z
z
Drawing the curve
Patience is a virtue…
Getting the S-curve #s Right
Number of “New Molecular Entities” approved per year in the U.S.
NMEs
60
Biotherapeutics
50
40
30
20
Source: FDA, Tufts CSDD
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
0
1964
10
Image by MIT OpenCourseWare.
Anomalous period for the X axis…
In part due to an “overhang” cleared during the early years after
PDUFA, the reduction in approvals since the late 1990s may simply be
a “return to trend” (Berndt, et al, 2004)
Getting the S-curve #s Right
$bn
40
35
Worldwide R&D spending
by PhRMA members
30
25
20
15
10
5
1
96
4
1
96
6
1
96
8
1
97
0
1
97
2
1
97
4
1
97
6
1
97
8
1
98
0
1
98
2
1
98
4
1
98
6
1
98
8
1
99
0
1
99
2
1
99
4
1
99
6
1
99
8
2
00
0
2
00
2
2
00
4
0
Nominal R&D spending
Constant 1964 dollars
Data source: PhRMA, NIH Biomedical R&D Price Deflator
Inflating the Y axis…
While most discussions of increasing cost compare nominal expenditures, the
cumulative impact of biomedical price inflation significantly reduces the
measured growth rate in R&D expenditures (Cockburn, 2007)
Getting the S-curve #s Right
80%
70%
60%
50%
40%
30%
20%
10%
0%
ACE Inhibitors
H2/PPIs
SSRI/SNRIs
Data source: Berndt, Cockburn, Grépin (2005) “The Impact Of Incremental Innovation In Biopharmaceuticals: Drug
Utilization In Original And Supplemental Indications.”
Not accounting for all progress on X axis….
High share of revenues for many drugs come from applications and indications
that are only discovered after market introduction, and these uses are not
always approved through formal FDA approval
Patience is a virtue…
Performance
Maturity/diminishing returns
“Biology-based” drug
development
Take-off
“Chemistry-based” drug
development
1970s
1990s
Time/effort
Biopharmaceuticals is going through a familiar process of disruptive (i.e.
costly) technological change from “chemistry” to “biology.” Do not be
surprised if this takes quite a long time to materialize.
Patience is a virtue…
4,500
850
800
4,000
750
700
3,500
650
600
3,000
550
2,500
500
450
19
95
19
96
19
97
19
9
19 8
99
20
00
20
01
20
02
20
03
20
04
20
05
Compounds in preclinical development
19
95
19
96
19
97
19
9
19 8
99
20
00
20
0
20 1
02
20
03
20
0
20 4
05
400
2,000
Compounds in phase I trials
Image by MIT OpenCourseWare.
Data source: Pharmaprojects/Goldman Sachs, PAREXCEL Pharmaceutical R&D Sourcebook 2005/2006
While approvals slowed 2000-2005, there has been a dramatic increase in the
number of promising compounds at earlier stages of the drug approval process
Managerial issues with using S
Curves as part of opportunity
analysis
Idealized approach is fine but
in reality several issues…
•When to switch/join?
•Which S Curve to switch
to/join?
•How to combine
incremental vs. switching?
•How to organize to switch/
join?
When to switch curves
800
600
EFI
400
1992
1990
1988
1984
1982
1980
0
1986
Carburetors
200
1978
Number of models of new cars sold
Car Models with EFI vs. Carburetors
6
5
4
3
Mean for cars
with carburetors
2
1
1992
1990
1988
1986
1984
1980
-1
1982
Mean for all cars
0
1978
MPG above or below the average
Carburetor Fuel Economy
Image by MIT OpenCourseWare.
Adapted from: Snow 2008
“Beware of Old
Technologies’ Last Gasps
When to switch curves?
Which curve to switch to?
Performance
Effort
Balance of incremental vs.
discontinuous
Physical limits
Technical
Metric
Physical limits
Technical
Metric
R&D Effort
IBM – strategic leapers focused on
new component technologies as a
source of improvement with little
movement up a give S curve or system
optimization.
3:4 ratio of incremental vs. radical
sources of improvement
R&D Effort
HP - system masters focus on
squeezing more incremental
improvement out of given components
4:1 ratio of incremental vs. radical
sources of improvement
Wrap-Up
z
CORE DEFINITIONS
z
z
z
S-curve is a useful heuristic to describe robust pattern of
technical change vs. effort
Innovations that move ALONG the curve are “incremental”
Innovations that shift to a NEW curve are discontinuities
(NOT disruptions)
z Discontinuities in “system” sometimes called “architectural”
z Discontinuities in “components” sometimes called “modular”
z I consider all discontinuities to be “radical”
Implications
Use technology S curve to answer the following questions:
z
z
z
z
What are the dimensions of performance in our industry?
Are there natural limits to performance improvement?
Where are our competitors on the S-Curve? Which dimensions of
performance are they working on?
What does the available data tell you about what stage the industry is at and
how much further it can go?
How reliable are your estimates & what are the key
assumptions that justify your opportunity definition?
The S curve is best viewed as a tool for triggering discussion &
revealing assumptions, not as a “scientific reality”
Next Class
z
Implications of technology S curve dynamics for market S
curve (diffusion) dynamics & for competitive dynamics
To put it another way…
z How should we map the ideas in the technology S-curve
to the market S-curve & the definitions of discontinuity to
Christensen: The Innovator’s Dilemma:
Sustaining vs. Disruptive Innovation
z
May seem like semantics but top management teams in
innovation-driven firms spend a lot of time on (often
erroneous) definitions…