1. Networks and Positive Feedback
Carl Shapiro
Hal R. Varian

2. Overview
Chapter 7 (this chapter) illustrates the key concepts that underlie market dynamics and competitive strategy in both real and virtual networks. 4 strategies. History lessons.
Chapter 8 discusses how to work with allies to establish a new technology – to launch a new network. Negotiations and standardization.
Chapter 9 examines what happens if negotiations break down: how to fight a standard war.

3. Important Ideas Positive feedback Network effects Returns to scale
Demand side
Supply side

4. Positive Feedback(善歸還) cf. regression 回歸, cycle 循環
Mkt share
PF: Strong gets stronger, weak gets weaker (from electronics term)
Negative feedback: stabilizing  the strong gets weaker, the weak gets stronger
PF makes a market “tippy” (쏠림)
Examples: VHS v. Beta, Wintel v. Apple
“Winner-take-all markets”
100%
50% 0%
tippy

5. Sources of Positive Feedback
Supply side economies of scale
(ever) Declining average cost
Marginal cost less than average cost
Example: information goods, early US car industry
Demand side economies of scale
Network effects
In general: fax, , Web
In particular: VHS v. Beta, Wintel v. Apple

6. Network Effects due to Positive Feedback
Less value
More value
Network Effects due to Positive Feedback
Real networks: telecommunication
Virtual networks: iPhone Users
Number of users
Metcalfe’s Law: the inventor
of Ethernet(LAN)
Value of network of size n proportional to n2≈ n(n-1)
Importance of expectations n
Value to Users
Virtuous
Cycle
Vicious

7. Lock-In and Switching Costs
Keyboard array(천지인) for human
Lock-In and Switching Costs
Network effects lead to substantial collective switching costs b/c of the value of network s/he has to bear if switch
Even worse than individual lock-in
Due to coordination costs
Example: QWERTY
Keyboard array for machine
August 드보락
1874(Type Writer)
1936, 속기사

8. Don’t Get Carried Away
Network externalities (or tipping) don’t always apply
AOL, Compuserv, Delphi (incompatible network services; 천리안, 코넷, 나우누리 PC통신, 90년대 중반부터 www) was going to tip the mkt but introduction of internet change all that  ISPs if with standard. No tipping.
PC production (no PF within the mkt)
Likelihood of tipping
See next slide

9. Likelihood of Tipping due to Positive Feedback
Scale Economy Effect > Demand for Variety Effect
Low Scale Economies (low Network Effect)
High Scale Economies
(high Network Effect)
Low Demand
for Variety
다양성 필요 없는 상품
Ex) Rice
Unlikely
Ex) Computer OS
High
High Demand
다양성 요구되는 상품
Ex) Tennis Racket
Low
Ex) Mobile Phone
Depends:
w std, do not tip
w/o std, tips. ③ ① ④ ②
Chance of tipping increases: ④③②①

10. Chicken & Eggs Problem: which makes mkt tip? Bottle or Wine?
The problem: wine without bottle or bottle without wine? Ex) Fax machines and fax, VCRs and tapes
Internet browsers and Java: Internet provides the contents (wine) AND the viewer (bottle) at the same time
Two strategies to ignite PF: 1) Performance vs. Compatibility 2) Openness vs. Control

11. Strategies to Ignite Positive Feedback 1: Performance vs. Compatibility
Evolution
Give up some performance to ensure compatibility, thus easing consumer adoption
Revolution
Wipe the slate clean and come up with the best product possible
Compatibility
Evolution
Improved
Design or
adapters
Revolution
Performance

12. Evolution Strategy: Smooth Migration
Offer a migration path
Examples
Microsoft (backward compatibility, Window XP vs. Windows 10) Ex) W8 써도 W10 쉽게 사용
Build new network (W10) by links to old one (W8)
Problems: technical and legal (next) + slow consumer adoption Ex) 아직 W 8 사용?
USB-C to USB

13. Technical Obstacles to Evolution Strategy
Need creative product design: compatible but superior, Ex) B/W TV v. Color TV, NBC engineers in 1950s.  Color TV signal but watchable with B/W TV
Need to think in terms of system: you may be making only one component, but the user care about the whole system. Ex) Broadcasting both conventional and HDTV digital signal, lessening switching cost. Analog vs. Digital broadcasting
Need converters and bridge technologies
One-way compatibility (ex: only analog to digital converter. XP to Win 8, not Win 8 to XP)

14. Legal Obstacles to Evolution Strategy
Need IP licensing
Example: IP of Sony and Philips on CDs  DVD makers needed Sony and Philips’ consent to make DVD read CDs.(as an evolution strategy)

15. Revolution Strategy: Compelling Performance
Intel’s CEO
Risky, need powerful allies, don’t know if the tech will take off or crash and burn even for great tech.
Groves’s law: “10X rule”: need to be 10 times better to revolutionize.
But depends on switching costs
Example: Sega v. Nintendo: Sega 64 was not 10X better than Nintendo 32 but switching cost was low (new crop of ten-year-old boys every year)

16. Strategies to Ignite Positive Feedback 2:
Openness v. Control
Mkt Share
CONTROL
OPEN
In choosing btwn them, your goal is to max Your reward (= Total value added to industry x your share), not control.  T&C control
Value added to industry depends on;
Product (what improvement, value of the tech) and
Size of network (how widely adopted; more open  larger network)
Your mkt share or installed base.(more open less share), Depends on how openly your tech used
Total Value
CONTROL
OPEN
Mkt Share
Too
Much
control
CONTROL
MAX your reward
Too open
OPEN
Total Value

17. Trade-off btwn Openness and Control
Your share
Proprietary
= Control = std war (chapter 9)
Optimum (largest area)
Hardware +
Software
Your Reward
Hardware
Open = cooperation (chapter 8)
Total Value added to industry
≒ network size
Software
Info tech is comprised of systems  as a system, more value  need cooperation of others  need to sacrifice mkt share to get them cooperate

18. Openness: emphasis on total value added to the industry
Openness strategy: more details in chapter 8
When no one firm is strong enough to dictate tech standard, or when multiple products work together
Full openness
Anybody can make the product
Problem: no champion, small mkt share
Alliance
Only members of alliance can use
Problem: holding alliance together
 small profit
Emphasis on this

19. Control: emphasis on your share
Control strategy: more details in chapter 9
For only those in the strongest position Ex) MS now, Sony and Philips before.
Control standard and go it alone
If several try this strategy, may lead to standards wars
Emphasis on this

20. Generic Strategies: Combining the 2 Strategies to ignite PF
One Vendor
Many Vendors
Strategies
Control
Openness
Compatibility
(evolution)
Controlled migration
Ex) Window 8  10
Open migration
Ex) generations of fax machine
Incompatibility
(revolution)
Performance play:
Ex) Nintendo in mid-1980s.
Discontinuity: new tech is incompatible but available from many suppliers
Ex) 3.5 inch floppy
New but
Compatible
Tech
Chapter 8 (cooperation)
단종 전략
New and
Incompatible
Tech
Chapter 9
(Std. War)

21. Controlled Migration Compatible, but proprietary Examples
Windows 10 (from Window 8)
32bit CPU  64 bit CPU
Upgrades
The original Pentium processor was a 32-bit microprocessor produced by Intel. The first superscalar(parallel computing) x86 architecture processor

22. Performance Play Introduce new, incompatible technology Examples
Palm Pilot
Iomega Zip, iPhone?
Attractive if
Great technology
Outsider with no installed base
Quantum
Computer

23. Open Migration Many vendors, compatible technology Examples
Fax machines
Some modems
USB 2.0  3.0  4.0

24. Discontinuity (not continued from old tech)
Many vendors, new(incompatible) technology
Examples
CD audio
3 1/2” disks
HDMI

25. Historical Examples of Positive Feedback: Standard Battles 1
협궤(narrow; 1000~1372mm)
표준궤(std)
광궤(wide)
마차 폭 넓으면
똑바로 끌기
어려움. 너무 좁으면
물건 많이 못 실음
RR gauges: 5 foot (152.4cm; South) v. 4’8½” (148cm; North, Roman cart, most efficient for horse pulled cart)  3 diff. widths met at Erie  Westward expansion + Civil War  Standardized in 1860~1890: Penn. Porters’ Riots, insignificant historical events can lead to lasting technological lock-in Ex) Track gauge conversion

26. Lessons from Railroad Gauge Standard Battle
Incompatibilities can arise almost by accident, yet persist for many years
Network market tend to tip toward the leading player, unless the other players coordinate to act quickly and decisively
Seceding from the standard-setting process can leave you in a weak market position in the future
A large buyer(such as U.S. Gov’t) can have more influence than suppliers(sellers) in tipping the balance

27. Historical Examples of Positive Feedback: Standard Battles 2DC AC
AC v. DC (Thomas Edison Direct Current: 1 mile limit but efficient power generation, Westinghouse Alternating Current: efficient transmission)
AC  transformer: allows power to be transmitted efficiently at high voltages and then stepped down to lower voltages for local distribution
DC densely populated area, AC small, remote towns
Edison electrocuted a dog with AC to show the danger of high voltage, but 1) AC got superior due to advances of polyphase 2) the rotary converter allowed DC stations integrated into AC stations 3) Edison established GE, supplying AC as well
Three-phase system transmits 73% more power but uses only 50% more wire.

28. Lessons from Electronic Industry Standard Battle
Technologies can seek well-suited niches (battery) if the forces toward standardization are not overwhelming
Ongoing innovation (polyphase AC ex: 3-phase, 공업용 전기, 380v) can lead to victory in a standard war
A first-mover advantage (of DC) can be overcome by a superior tech (of AC) if the performance advantage is sufficient and users are not overly entrenched
Adapters (rotary converter) can be the salvation of the losing tech and can help to ultimately defuse a standard war

29. Historical Examples of Positive Feedback: Standard Battles 3
Telephone networks
Color TV
HDTV
3D TV
Samsung: Active Shutter LG: Passive Polarization

30. Lessons
Positive feedback means strong get stronger and weak get weaker
Consumers value size of network
PF works for large networks(large gets larger), against small ones (small gets smaller)
Consumer expectations are critical
Fundamental trade-off 1  performance and compatibility

31. Lessons, continued Fundamental trade-off 2  openness and control
Generic strategies (Trade-off 1 x Trade-off 2)
Performance play
Controlled Migration
Open Migration
Discontinuity
Lessons of history

32. Feedback 피드백
한국: 되먹이기, 귀환(歸還), 중국: 反馈[fǎnkuì], 일본: フィードバック(휘도밧구), 反応(はんのうhannō)
전기용어: 입력신호와 동위상으로 귀환 (feedback) 하는 것을 <정귀환> (正歸還), 반대위상으로 귀환하는 것을 <부귀환> (負歸還)이라고 함.
정귀환 = positive feedback 전기용어에서 유래 : 자동 제어계나 증폭기 따위에서, 출력 회로의 전력 일부가 입력 회로에 대하여 그 입력 전력을 증가하도록 가하여지는 것. 이로 인하여 증폭 작용이 강화되고 결국은 발진(發振)이 생김
악순환惡循環=vicious cycle, 선순환善循環=virtuous cycle:

33. Track Gauge Conversion
Bogies exchange operation in Ussuriisk (near Vladivostok) at the Chinese–Russian border
Bogie change station at Chop, Ukraine station, Ukraine