Network Effects
•
•
•
•
Increased market size makes product more
valuable to consumers.
This is just like an economy of scale in that it
benefits large firms relative to small ones. Might
lead to natural monopoly.
Since demand increases for large networks, large
networks should have higher prices.
In Microsoft case, judge decided that they were
a barrier to entry, that incumbents were lockedin, or put another way, challengers were lockedout.
1
Network Effects
• Definition: a product becomes more valuable to a
consumer the more other consumers there are of the
product.
• Virtually identical theoretically to economies of scale
• Markets will tend toward monopoly, winner-take-all
result (in simple world).
• Example: Fax machines, some software, languages,
online networks, computer standards,.
• Literal networks (physical connections, e.g., fax
machines) and Virtual Networks (e.g., software)
2
Potential Problems Due to
Network Effects
• Traditional Problem: Network is the wrong
size. Old fashioned negative externality.
– This is just the inverse of the tragedy of the
commons. Here the network is too small. There
the network (commons) is too big (overused).
• New Problem: Getting stuck with the wrong
network—i.e., lock-in.
3
Effects versus Externalities
• Where we have positive externalities,
activity is too small (golf courses,
research?).
• Where we have negative externalities,
activity is too big (e.g., air pollution,
traffic).
• One key element is whether external effects
are internalized, or whether they are truly
externalities.
4
Tragedy of the Commons
•
•
•
•
Example of Negative Externality.
Common Property Resource – lake, forest,
any productive resource that allows free
use.
The tragedy is that the resource is
overused.
Greater tragedy is that it is overused to the
point where its entire value might be
dissipated.
5
Fishermen on Lake
Fishermen Fish per Fisherman Total Catch Marginal Catch
1
10
10
2
9
18
8
3
8
24
6
4
7
28
4
5
6
30
2
6
5
30
0
7
4
28
-2
6
Illustrating Overuse
Fish per
Fishermen Fisherman
1
10
2
9
3
8
4
7
5
6
6
5
7
4
Total
Catch
10
18
24
28
30
30
28
Marginal Opportunity
Catch
Cost
6
8
12
6
18
4
24
2
30
0
36
-2
42
Net
Value
4
6
6
4
0
-6
-14
7
Business Applications
• Should a firm have internal charges when
one division helps another (e.g. technical
support)?
• The relationship to Network Effects.
8
Lock-in
• Claim: Markets do not adopt best products
even when it is efficient to do so.
– QWERTY, VHS-Beta are most common
purported examples.
• Story is one of coordination failure.
– We all prefer Beta. But VHS dominates, and
most movies are on VHS. Since we think that
everyone else will get VHS, we get VHS too.
• My work with Margolis has shown the key
examples of lock-in to be false.
9
Lock-in Table
T able 1: A D O P T IO N P A Y O F F S
N um ber of P revious A do ptions
0
10 20 30 40 50 60 70 80 90 100
T echnolo g y B (B eta)
T echnolo g y V (V H S )
10
4
11 12 13 14 15 16 17 18 19
7 10 13 16 19 22 25 28 31
20
34
• Technology B Wins although Technology A
is better.
40
Figure 1
V
30
B
20
10
Benefits
0
10
20
30
40
50
60
Number of Adopters
70
80
90
• Assumption about different slopes? What
does it really entail, and is it reasonable?
• Assumption that consumers only look at
their original payoff and not their final
payoff is also unreasonable.
Qwerty Story
• Most famous case;
– simple to identify ‘quality’ (speed of typing)
– Consistent with chaos theories of unpredictable small events
leading to big outcomes.
• Usual Story:
– Qwerty Keyboard was designed to slow down typing to keep
keys from getting stuck.
– In a famous typing contest in Cincinnati, pitting touch typist
against ‘hunt-and-peck’ the touch typist used a Qwerty and won.
This is the small accident leading to big outcomes, since people
then associated Qwerty with speed.
– Prof Dvorak designed a keyboard to speed things up in 1930s.
– During WWII Navy tested new keyboard and found that speed
increased 40% and cost of retraining was recouped within 10
days of completed switch.
– Keyboard never adopted. End of war reduced Navy’s need for
fast typists.
12
Qwerty Story
• Actual Story
– No evidence that Qwerty Keyboard was designed to slow
down typing speed, and it wasn’t necessary to keep keys
from getting stuck.
– There were many other typing contest pitting touch typist
against other touch typists on different keyboards and
Qwerty won many of these. The point is that the
Cincinnati contest was not crucial.
– A highly publicized 1954 test of the Dvorak keyboard by
the General Services Administration found no advantage
in switching to Dvorak.
– The WWII Navy study was biased in the way it calculated
it results. All the biases were in favor of Dvorak. It also
appears that the author of the study was probably Dr.
Dvorak, the designer and patent holder of the keyboard.
– Ergonomics studies on keyboard design confirm that
Qwerty is a good design and not much different than
Dvorak in terms of speed
13
Beta-VHS Story
• Story
– Beta was better than VHS.
– VHS got an early lead and network effects allowed it to become
dominant even though it was inferior.
– Network effect was the fact that prerecorded tapes were available
in VHS and people wishing to rent tapes would get a VHS
machine to get a good selection of tapes.
• Actual History
– Beta started first, but could only tape for 1 hour.
– Sony (Beta) and VHS (Matsushita) had a patent sharing
agreement.
– Only real difference was size of tape and how it was threaded.
– VHS always had longer playing time which turned out to be
decisive.
– Beta had better editing and so it is used in professional
equipment
– VHS as good or better than Beta according to independent tests.
14
Other Stories
•
•
•
•
•
•
•
Quadraphonic sound
Railroad gauges
Macintosh versus dos
Internal combustion engine
Metric versus English measurement
Stereo AM
AC versus DC
15
Weaker forms of Lock-in
•
•
Decisions are often made that turn out to
be incorrect yet not worth changing. This
is a weaker form of lock-in.
There are at least two weaker forms that
we can identify, called first and second
degree in the literature.
16
Measuring Network Effects
• Several attempts:
– Gandal examined spreadsheet prices.
• Examined whether Lotus file compatibility and
ability to link to external databases were
characteristics that led to higher prices.
– Brynjolfsson and Kemerer examined
spreadsheet prices
• Examined whether Lotus menu structure and
installed based lead to higher prices.
– Literal networks: fax machines (Saloner and
Shepard).
17
Characteristics of Information
Markets
•
Bigger is better, large firms (standards) have
advantages over smaller firms (standards).
–
–
Economies of Scale: large fixed costs
Network Effects
•
•
–
Definition: a product becomes more valuable to a consumer
the more other consumers there are of the product.
If these effects are not exhaustible, we have ‘natural
monopoly’, which used to be thought to require government
regulation .
Instant scalability: the ability to increase output
extremely quickly. Not the same as zero marginal cost.
18
Network Effects and Software
• Standard Claims:
– Winner-takes-all.
• Market should gravitate toward a single winner.
• Note that ‘instant scalability’ can also lead to winner-take-all
type of result since market can follow preferences of
consumers more strongly than would be typical.
– Lock-In: Network effects presumably strengthen
monopoly of leader.
• Superior challengers can replace leader with built in network
effects until difference in quality becomes
– Tipping: At some point network effects support one
firm or product so strongly that it becomes the standard
and takes the market.
19
Methodology
• We compare market share changes with product quality
changes.
• Quality measured by magazine reviews.
– Magazines could be biased
– Whether reviews are right or wrong, consumers have only this
data and can interpret them the way they want.
• 2 measures of quality: ‘wins’ and ratings.
– ‘Wins’ is based on the top numerical score, Editors Choice, or
some other indication that some product was the best in its class.
– Many fewer reviews give numerical scores. When they did we
set the top score to 10 and normalized the other scores to match.
20
Figure 8.7: Spreadsheet Wins
8
7
Excel
Quattro
Lotus
6
5
4
3
2
1
0
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
21
R atin g s
50
40
30
7
20
106
0
5
Lotus
E xc el
P C M a ga zin e
2/ 18/ 97
P C M a g a zin e
B yt e
12/1/ 96
2/ 1/ 98
PC W o rld
4/ 1/ 96
H om e O ff C om p
2/28/ 94
11/ 10/ 92
P C M a ga zine
12/ 31/ 91
P C M a ga zine
9/ 1/ 91
P C W o rld
6/1/ 91
Quattro
INF OW OR LD
By t e
7/ 1/ 89
Excel
2/ 8/94
8
10/ 12/ 92
60
IN FO W O RLD
70
P ers on a l C o mp u tin g
9/ 1/88
P e rs o na l Co m pu t ing
Figure
S Set R atings
Figure
8 .8 :8.12:
C o mRevenue
parison S hare
preadofshe
100
10
90
809
L otus
Q ua ttro
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
22
F igure 8.12: Revenue S hare of S S
100
90
80
70
60
E xc el
50
40
Q uattro
Lotus
30
20
10
0
1988
1989
1990 1991
1992 1993
1994
1995 1996
1997
23
9
8
7
Ami Pro
MS Word Windows
WordPerfect for Windows
PC Magazine
2/18/97
PC Magazine
11/21/95
PC Magazine
2/8/94
PC Magazine
11/9/93
PC Computing
8/1/93
PC Magazine
2/25/92
Infoworld
2/10/92
Infoworld
1/7/91
ComputerWorld
5/24/93
PC Magazine
12/11/90
Personal Computing
8/1/90
Figure 9.5: Windows WP Ratings
10
Word Processor Market Shares
100%
WordPerfect
Word Star
90%
Microsoft Word
Ami
80%
70%
60%
50%
40%
30%
20%
10%
0%
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
25
Fig 9.21: M idrange DT P W ins
7
6
5
4
3
2
1
0
pre 1990
1990-1991
MS Publisher
1992-1993
Publish It!
94-95
96-98
Other
F ig 9.22: M id ran g e P C D T P S h ares
90%
80%
70%
60%
50%
Mic ros oft P ublis her
E xpres s P ublis her
S erif P ageP lus
O ther
40%
30%
F irs t P ublis her
C orel P rinthous e
P ublis h It
20%
10%
0%
1990
1991
1992
1993
1994
1995
1996
27
F ig 9 .1 0 : P e rs o n a l F in a n c e W in s
9
Managing Your Money
Q uic ken
Mic ros oft Money
8
7
6
5
4
3
2
1
0
86-90
91-92
93-94
95-96
97-98
28
Fig 9.12: M arke t S hare P C P e rsonal Finance
80%
70%
60%
50%
40%
30%
Q uic k en
M ic ros oft M oney
M anaging Y our M oney
20%
10%
0%
1989 1990 1991 1992 1993 1994 1995 1996
29
8.5
11 /1 /9 3
M a c U se r
6/ 1/ 89
10 /2 5/ 93
InfoW orl d
8/ 3/ 92
InfoW orl d
11 /5 /9 0
InfoW orl d
1/ 1/ 90
9
P e rsona l C om p ut in g
9.5
M a c U se r
6/ 20 /8 8
InfoW orl d
10.5
F ig . 9 .1 4 : HIg h E n d M a c in to s h D T P
10
P ageM akerM
Q uarkXP ressM
8
30
F ig u re 9.15: M arket S h are M ac H ig h -E n d D T P
80%
PageMaker
70%
Quark
Frame
60%
50%
40%
30%
20%
10%
0%
1990
1991
1992
1993
1994
1995
1996
31
5
4
6
1 0/10 /95
P C M aga zine
3
32
5 /6 /97
P C M aga zine
1 1/19 /96
7 /1 /95
C om puse rve
P C M aga zine
6 /1 1/9 6
8
M acu se r
AOL
P C M aga zine
6 /1 /95
M SN
P C W orld
2 /2 1/9 5
P C M aga zine
3 /1 5/9 4
P C M aga zine
2 /2 3/9 3
P C M aga zine
1 1/12 /91
P C M aga zine
11
F ig u re 4 1 : O n lin e S e rv ic e R a tin g s
P rodigy
10
9
7
Figure 9.27: Online Market Shares
AOL
CompuServe
MSN
Prodigy
70%
60%
50%
40%
30%
20%
10%
0%
1991
1992
1993
1994
1995
1996
1997
In 1994 Prodigy dropped its $14.95 unlimited use pricing and went to the
same pricing as its competitors ($10 for 5 hours, $3/hr after).
Business Implications
• First to market: How Important??
• How important is large market share early
on?
• If lock-in unimportant, then quality is more
important than the order of entry.
• So what should a firm do to succeed in
information goods market?
34
Descargar

No Slide Title