Slide1 : Learning from Heterogeneous Experience: Internationalization of Entrepreneurial Firms
CHRISTOPHER B. BINGHAM
KATHLEEN M. EISENHARDT
Department of Management Science and Engineering
Stanford University
We appreciate the support of the National Science Foundation and the Stanford Technology Ventures Program
Background : Much research suggests that organizational processes are learned via repeated experience (Epple et al., 1991; Levin, 2000; Pisano et al., 2001)
BUT, a striking feature is that research rarely examines how processes are learned or the content of what is learned
RATHER learning is inferred from outcomes
Background
Background : Background More recent research examines the relationship between aspects of experience (similarity, timing) andamp; outcomes (Hayward, 2002; Haleblian andamp; Finkelstein, 1999)
BUT still no understanding of what is learned and how it is learned
Heterogeneous experience is particularly challenging
Research Question : Research Question
How do executives learn organizational processes?
Definition:
An organizational process is a collection of rules, heuristics or guidelines that executives use for accomplishing a key task (e.g., product development, alliances, acquisitions, or internationalization)
Overview : Overview
Given the lack of empirical evidence and theoretical argument, I use an inductive grounded theory-building approach
Focus on internationalization
12 technology-based entrepreneurial firms
HQ in three culturally distinct countries (Singapore, U.S., and Finland)
Research Design (1) : Research Design (1) Research design is a multiple case, embedded study
Replication logic (Yin, 1994)
Research setting is entrepreneurial firms
Small size enables better observation of phenonemon
Young age avoids left censoring
Sample is 12 technology-based entrepreneurial firms
Four each from Singapore, U.S. and Finland
Research Design (2) : Research Design (2) Focal process is internationalization
Each country can be analyzed as a single unit of analysis and as part of a larger sequence
All country entries took place within five years of data collection
Previously entered at least 4 countries and were currently entering at least one new country (retrospective andamp; real time)
All companies now gain a significant portion of their revenues from sales outside their home country
Data Collection (1) : Data Collection (1) Three data sources:
Semi-structured interviews with company leaders
Extensive archival data (public and corporate sources)
E-mails, phone calls, and follow-up interviews
Main source was semi-structured interviews with multiple informants within each firm (Golden, 1992; Miller et al., 1997)
About 70 interviews on three continents
Informants were two types
Corporate level (CEO, founder, VP Intl)
Country level (country manager)
Data Collection (2) : Data Collection (2) Interviews were 60-90 minutes and had variations for the two types of informants
Relied on courtroom questioning (Galunic, 1994)
Interviews consisted of a series of closed and open-ended questions designed in three parts
Prompted with more detailed follow-up questions
Interview data triangulated with archival data (Jick, 1979)
Data Analysis (1) : Data Analysis (1) Synthesize interview and archival data into cases
Narrative of internationalization chronology supplemented with quotes, timelines, tables of key facts (Miles andamp; Huberman, 1984)
Incorporated perspective of second researcher to provide a stronger and better grounded summary for each firm
30-60 pages
Data Analysis (2) : Cross-case analysis
Look for emergence of similar themes and constructs across cases using a variety of lenses (Eisenhardt, 1989)
Developed initial propositions
Constant iteration between theory and data
Resulted in theoretical framework for how executives develop strategically significant organizational processes
Data Analysis (2)
Data Analysis (3) : This presentation focuses on six firms that were able to develop an internationalization process
A second paper will be a variance-oriented inductive study that analyzes why some did and why some did not
A third paper will be a deductive study of processes at the country level
Data Analysis (3)
Slide13 : Description of Cases
Research question: : Research question:
How do executives learn to internationalize?
Process definition: Collection of rules or heuristics that executives use for accomplishing a key task
Measured an internationalization process by assessing the rules or guidelines that executives used for entering each country
A rule exists if it was a recognizable guide for action in internationalization that was understood and used by multiple informants in more than one country entry
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Prelude: Before any country entry : Prelude: Before any country entry Some literature suggests that organizational processes emerge through 'learning by doing'
Too hard to anticipate and so just do it
Need to adjust to serendipitous opportunities
Other literature suggests learning is aided by a 'cognitive template'
If so, what is in the template? What is not?
Where does the template come from (e.g., experience)?
Does the template need to be correct?
Cognitive templates : Cognitive templates Key finding: firms begin with a cognitive template – learning before doing
What did cognitive templates look like?
Cognitive templates took the form:
A few boundary rules – defined as heuristics about which opportunities to target (e.g., countries and customers)
A few how-to rules – defined as heuristics for capturing target opportunities (e.g., sales approach and mode of entry)
Cognitive templates : Where did these cognitive templates come from?
Experience of knowledgeable executives
And, executive foresight, including that of neophyte entrepreneurs
Cognitive templates
Background: Echo :
HQ: Singapore
Date incorporated: 2000
Founders: Two
47, Hong Kong, and had worked with HK telecom for 21 years
45, Hong Kong, and had lived eight years abroad as regional VP for IBM
Idea: Create security products (e.g., firewalls) to help customers secure their IT infrastructure
Background: Echo
Slide25 : Cognitive Template
Background: Crest :
HQ: Finland
Date incorporated: 2000
Founders: Three recent graduates of Helsinki U.
25, Finnish, and had worked one year for McKinsey
24, Finnish, and had done some academic research
24, Finnish, and had a little experience in pharma
Idea: Create software that would help expedite clinical trials data analysis for the pharmaceutical industry
Background: Crest
Slide27 : Cognitive Template
Unexpected Errors in Cognitive Templates : Surprisingly, while cognitive templates served an important role in seeding experience, they were usually flawed in some fundamental and unexpected way
Errors in boundary rules
Errors in how-to rules
Combinations of errors
Unexpected Errors in Cognitive Templates
Slide29 : Unexpected Errors
Slide30 : Unexpected Errors
Key Points : Key Points Processes begin with cognitive templates before any doing
Initial cognitive templates consist of a few boundary and how-to rules
Cognitive templates stem not only from prior experience, but also from foresight
Cognitive templates are usually wrong
Observations : Observations Cognitive templates provide a head start so that experiential learning takes place further in process development (Eysenck and Keane, 1995)
Cognitive templates provide a simple structure within which people can organize change (Okhuysen andamp; Eisenhardt, 2002; Siggelkow, 2001)
Cognitive templates categorize information and so clarifies what is wrong and how to fix it (Friedman, 1979; Sitkin, 1992)
Cognitive templates give executives the confidence to act in uncertain environments (Langer, 1975; Weick, 1993, 1995)
Act I: Early country entries : Act I: Early country entries What happens next?
During initial country entries, organizational processes develop (from a couple of rules) in several ways
One way was expected
Processes continued to develop through the elaboration of existing heuristics (Pisano, 1994; Yip, 1992)
Background: Alta :
HQ: Finland
Date incorporated: 1997
Founding CEO:
50, Finnish, with extensive work experience in Finland’s retail industry
Idea: Create point-of-sale software to help retailers manage inventory
Background: Alta
Elaboration of detail : Elaboration of detail
Also in Act I: Early country entries : Also in Act I: Early country entries Yet, another change was more unexpected
Executives added temporal heuristics to the internationalization process
Temporal heuristics emerged from the data and took the form of rules about:
Pacing
Sequence
Destination
Background: Grand :
HQ: Singapore
Date incorporated: 2000
Founders: Three
30, Taiwan, and had worked at SAP for many years
38, Singapore, and had also worked at SAP (including seven years abroad)
30, Singapore, and had worked for McKinsey in Singapore for several years
Idea: Develop supply chain software for high-tech manufacturing industry
Background: Grand
Slide38 : Adding time
Background: Center :
HQ: U.S.
Date incorporated: 1999
Founding CEO:
55, Chinese-American, who had founded three previous companies
Idea: Semiconductor solutions for GPS enabled mobile devices
Background: Center
Slide40 : Adding time
Key Points : Key Points Expectedly, processes further develop through the elaboration of 'how-to' rules
Building off what was right from the past
Improving understanding of what to do
Unexpectedly, processes also develop through the incorporation of 'temporal' rules
Not just learning what to do, but when to do it
Observations : Observations Consistent with the maturation of sense-making in developmental psychology (Siegler, Deloache, andamp; Eisenberg, 2003)
Temporal heuristics help executives to build experience in appropriate order (Saunders andamp; Jones, 1990) – planned path dependence
Create temporal flow between experiences that enhances awareness (Brown andamp; Eisenhardt, 1997)
Act II: As time goes on : Act II: As time goes on Research suggests that processes continue to develop by becoming increasingly routine with more experience (Nelson andamp; Winter, 1982; Fredrickson, 1984; Argote, 1999)
More taken-for-granted
More mindless and habitual
Finding 3: Creating levels of abstraction : Finding 3: Creating levels of abstraction In contrast, processes can evolve by becoming less, not more, routine through experience
Rather than habitually repeating set rules, executives heedfully shifted the levels of abstraction in relevant rules
Lower level of abstraction – creating a more specific conceptualization of heuristics
Raise level of abstraction - creating a more generalized conceptualization of heuristics
Slide45 : Creating levels of abstraction
Slide46 : Creating levels of abstraction
Slide47 : Creating levels of abstraction
Key Point : Key Point Organizational processes can become less, not more, routine over time
Rather than habitually repeating increasingly detailed rules, managers heedfully shift the level of abstraction and so simplify at least some rules
Observations : Observations Creating levels of abstraction enables a more efficient and flexible, yet coherent strategy (Locke andamp; Latham, 1980; Miner, Bassoff, andamp; Moorman, 2001)
Raising abstraction is consistent with the higher order thinking of experts which enables greater creativity (Dunker, 1926; Scheerer, 1963)
Attentiveness to levels of abstraction reduces failure and improves reliability as executives become more mindful in their actions (Langer, 1989; Weick, Sutcliffe, andamp; Obstfeld, 1999)
Discussion : Discussion I began by observing that research on learning focuses on performance, and so infers learning
In contrast, I ask what is learned?
Specifically, I examine the developing content of an organizational process, internationalization, in six entrepreneurial organizations
Key Findings : Key Findings Organizational processes begin as rough cognitive templates that include simple boundary and how-to heuristics. They seed experience, but are wrong in major ways.
Processes further develop by:
elaborating what to do (i.e., priorities and more detailed how-to heuristics) and
adding when to do it (i.e., temporal heuristics such as pacing, sequence, synchronization).
Processes continue to develop through varied levels of abstraction, and so become less routine.
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Black = Template
Red = Errors
Brown = Elaboration
Green = Temporal
Blue = Abstraction
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Black = Template
Red = Errors
Brown = Elaboration
Green = Temporal
Blue = Abstraction
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Black = Template
Red = Errors
Brown = Elaboration
Green = Temporal
Blue = Abstraction
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Black = Template
Red = Errors
Brown = Elaboration
Green = Temporal
Blue = Abstraction
Development of Echo’s internationalization process : Development of Echo’s internationalization process
Black = Template
Red = Errors
Brown = Elaboration
Green = Temporal
Blue = Abstraction
Contributions : Contributions Organizational learning
Clarify the emergence of strategically significant processes
What is being learned (i.e., types of heuristics, role of opportunities) and when is it being learned (i.e., developmental sequence)
Role of cognitive foresight on accelerating trial-and-error learning in an empirical setting
Link of improvisation to abstraction
Little evidence for codification
Blend of learning approaches
Contributions : Contributions Organizational capabilities, rules and routines
Link to capabilities-capabilities are composed of particular kinds of rules that have idiosyncratic (and indeed strategic) content
Link to rules-rules become simpler,not more complicated
Link to routines-development is lumpy, not incremental and processes are mindful and in flux, not inertial
Overall : Overall This study contributes to:
Organization theory by adding to the growing body of research on how organizations can change
Strategy by indicating that process is not just the formation and implementation of strategy, but rather is the strategy in high velocity environments
More broadly, : More broadly, Learning strategically significant organizational processes involves developing expertise, not creating a routine
Questions : Questions
Contributions to international research : Contributions to international research Uncovers vaguely understood role of sequence
Not just following culturally similarity, but assembling experience in a more appropriate manner
Traditional process theory describes internationalization as an inertial and incremental process (Johanson andamp; Vahlne, 1977, 1990)
In contrast, internationalization as highly dynamic and non-incremental process (e.g., high commitment entry modes and culturally distant countries first rather than last)
Young firms usually follow an ad hoc approach during internationalization (Yip, 2000)
In contrast, some firms are quite systematic
To be successful during internationalization, a company needs a significant competitive advantage that it can leverage abroad (Chang, 1995)
In contrast, firms began in with basic rules and improvised with abstraction in order to adapt to the local environment
Differences at country level : Differences at country level Country of origin differences may emerge in the content of the rules (e.g., Singapore firms tend to be more relational in their patterns for internationalization) or in the pathologies (U.S. may start reactively and do too much at same time)
Real difference is that countries are different when you enter them, but they are different in the same way for everyone (e.g., Japan is very detail oriented and take a long time to make decisions, whereas in Korea they make decisions very fast)
Indeed, a key test of in-country performance is whether managers learn the country specific lessons
Cultural distance among Finland, U.S., and Singapore : Cultural distance among Finland, U.S., and Singapore Rank scores for countries based on 116,000 questionnaires collected from 72 countries (Hofestede 1980; 2001)
Contributions to entrepreneurship research : Contributions to entrepreneurship research Literature assumes up front planning offers little advantage to new ventures – let strategy emerge (Bhide, 2000)
In contrast, this study shows importance of some a priori planning to structure and guide future activity (Delmar andamp; Shane, 2003). Blend of deliberate and emergent (Mintzberg andamp; McHugh, 1985)
Literature assumes new ventures face inherent liabilities of newness (Stinchcombe 1965) and/or foreignness (Zaheer and Mosakowski 1997)
In contrast, data suggestive of learning advantages of newness
Role of abstraction in enabling opportunity capture
Generalizability of findings to other organizational processes : Generalizability of findings to other organizational processes Alliances
Some support for shifting levels of abstraction in rules (Kale et al., 2002)
Product development
Pacing and rhythmic transitions between events (Brown andamp; Eisenhardt, 1997)
Time (Danneels, 2002)
Abstraction to enable improvisation (Miner et al., 2001)
Acquisitions
Cognitive templates and errors - 'Learning is more likely when prior acquisition incur small losses, prompting refreshed search for the right acquisition' (Hayward, 2002)
Support for pacing (Hayward, 2002)
Generalizability of findings to established firms : Generalizability of findings to established firms Cognitive templates and revelatory errors (Pascale, 1984; Aharoni, 1966)
Temporal rules (Gersick, 1994; Brown andamp; Eisenhardt; Vermeulen andamp; Barkema, 2002)
Abstraction (Miner et al. 2001)
Slide68 : Methods: Inductive Grounded Theory-Building
Reactive entry : Reactive entry 'There was no clear picture of the major countries that we should target.'
'It was ad hoc.'
'Country selection took place sort of by fate.'
'The reason for going in…what was it? More like grasping at straws.'
'Many entrepreneurs behave reactively, expanding abroad in response to stimuli rather than according to a thoughtfully crafted strategy.' - Kuemmerle (2005)
Definitions : Definitions Prototype
An original model on which something is patterned (Websters, 1993)
Functional form of a new type (Thomke, 2003)
Schema
outline, diagram, plan or preliminary draft
Script
knowledge structures that encode the stereotypical sequence of actions in everyday happenings
Template
Something that establishes or serves as a pattern (Websters, 1993)
Others: analogue representations, mental models, mental representation
Field guide: a guidebook describing natural objects of some sort that one might encounter in the field:
Guidebook: something that offers basic information or instruction
Fragment: an incomplete piece
Rules vs. Routines : Rules vs. Routines Heuristic: an informal rule to guide behavior
Rule: A prescribed guide for conduct or action (Websters, 1993)
Routine: A regular course of procedure; of commonplace or repetitious character (Websters, 1993)
Capabilities vs. Dynamic Capabilities (Helfat & Peteraf, 2003) : Capabilities vs. Dynamic Capabilities (Helfat andamp; Peteraf, 2003) While some capabilities may deal specifically with adaptation, learning, and change processes, all capabilities have the potential to accommodate change. Learning, change, and adaptation do not necessarily require the intervention of 'dynamic' capabilities as intermediaries
Thus, capability development does not require dynamic capabilities. With a new capability in a new-to-the-world company, dynamic capabilities don’t play a large factor since a new organization has no dynamic capabilities.
In the founding and development stages of a capability, a capability evolves over time without the action of any dynamic capabilities upon it. The development thus applies to dynamic as well as non-dynamic capabilities.
Slide73 : Shifting Levels of Abstraction