The Effect of Market Leadership in Business Process Innovation ...

The Effect of Market
Leadership in Business
Process Innovation:
The Case(s) of E-Business
Adoption


Kristina Steffenson McElheran



Working Paper


10-104

Copyright © 2010 by Kristina Steffenson McElheran
Working papers are in draft form. This working paper is distributed for purposes of comment and
discussion only. It may not be reproduced without permission of the copyright holder. Copies of working
papers are available from the author.

---

The Effect of Market Leadership in Business Process Innovation:
The Case(s) of E-Business Adoption

June 2010

Abstract:

This paper empirically investigates how market leadership influences firm propensity to adopt new
business process innovations. Using a unique data set spanning roughly 35,000 plants in 86 U.S.
manufacturing industries, I study the adoption of frontier e-business practices during the early diffusion of
the commercial internet. Theory predicts that firms with greater market share will be more likely to adopt
innovations that build on their existing strengths, while they will resist more radical technological
advances. While prior work primarily focuses on product innovation, I extend the logic into the business
process setting to find that leaders were far more likely to adopt the incremental innovation of internet-
based e-buying. However, they were commensurately less likely to adopt the more strategically sensitive
and complex practice of e-selling. This pattern is remarkably robust, holding across a wide range of
industries and controlling for factors such as productivity and related technological capabilities. The
results are explicated by a framework I develop for understanding the drivers of this behavior and making
it possible to classify business process innovations as radical or not. While greater market share promotes
adoption of all types of business process innovations, this effect is outweighed by additional co-invention
and coordination costs whenever a technological advance address strategically sensitive and complex
business processes that must also span the firm boundary.
JEL classifications: L21, O33, D24, M15

* Harvard Business School, 431 Morgan Hall, Boston, MA 02153. Thanks are due to Shane Greenstein, Scott
Stern, Rebecca Henderson , Alfonso Gambardella, Steve Kahl, Juan Alcacer, participants of the Colloquium on
Competition and Competitiveness, and numerous others I met while on the job market. I am indebted to T. Lynn
Riggs and other members of the Center for Economic Studies for their support. The research in this paper was
conducted while the author was a Special Sworn Status researcher of the U.S. Census Bureau at the Chicago Census
Research Data Center. Research results and conclusions expressed are those of the author and do not necessarily
reflect the views of the Census Bureau. This paper has been screened to insure that no confidential data are revealed.
Support for this research at the Chicago RDC from NSF (awards no. SES-0004335 and ITR-0427889) is also
gratefully acknowledged. In addition, I am grateful for financial support from the Census Bureau’s Dissertation
Fellowship program and the State Farm Companies Foundation Dissertation Award. All errors are my own.



1

---

1. Introduction
Identifying when market leaders will embrace technological progress has been a long-standing
question in innovation and strategy research. As far back as Schumpeter (1934, 1942), scholars have
debated the impact of market position on innovation – are large incumbents or small entrants more likely
to advance technological change?
The answer has far-ranging implications for technology diffusion, corporate strategy, and
industry evolution. If leading firms can raise the bar for competitors by pursuing innovative products and
processes, then their industries may advance technologically while remaining concentrated among
dominant firms. Innovation in this scenario becomes a tool for leading firms to maintain or improve their
competitive position. On the other hand, resistance to new technological opportunities may invite
incursions by entrepreneurial rivals, causing market leaders to fall behind or eventually be replaced. Thus,
understanding the relationship between market leadership and the adoption of new technologies becomes
central to an appreciation of how firms may maintain or gain competitive advantage over time.
A large theoretical literature in both economics and strategy has explored how firms of different
sizes and market positions will respond to innovative opportunities. Yet consensus remains elusive.
According to prior work, larger, dominant firms have both more to gain and more to lose from disrupting
the status quo. They have organizational advantages that smaller firms may lack, but also greater inertia.
Predicting the ultimate relationship between market share and innovation ends up depending on a wide
range of modeling assumptions – many of which are difficult to distinguish in practice. Relevant
empirical tests remain likewise inconclusive, due in part to econometric problems and a tendency toward
single-industry studies or anecdotes that are difficult to generalize.
The goal of this paper is to empirically explore whether market leaders are most likely to adopt
business process innovations, when this is most likely to happen (or not), and why. In search of a robust
and generalizable answer to the first question, I leverage an empirical context that cuts across a broad
range of industry settings. The data come from a survey conducted by the U.S. Census Bureau in 1999 of
roughly 35,000 plants in 86 manufacturing industries. Designed to measure the use of information
technology (IT) among manufacturers, this survey provides an opportunity to observe IT-enabled process
innovation amongst a representative sample of firms. I focus on the adoption of frontier e-business
practices during the early diffusion of the commercial internet, in particular internet-based e-buying and
e-selling.
Understanding when market leaders might have an advantage in business process innovation
requires a more precise theoretical understanding than is easily sifted from existing work. To gain clarity,
I leverage insights from the economics and strategy literature on incremental versus radical innovation

2

---

(e.g., Henderson 1993). Fundamentally, this framework recognizes that the nature of an innovation
matters. Changes that reinforce the strategic and organizational advantages of incumbent firms are more
likely to be adopted by them. The greater the incumbent’s market share, the greater the propensity to
adopt these incremental advances. In contrast, radical changes that threaten existing market position – for
instance by cannibalizing sales or disrupting strategically sensitive activities – are commensurately less
attractive for leading firms.
Testing whether this distinction holds true in the business process setting presents particular
challenges. To begin, existing research on innovation and technological change has focused
overwhelmingly on the invention and commercialization of new products. Process innovation, by
comparison, has received far less attention (Rosenberg 1982), particularly in the scholarly literature.
Business process innovation, has received even less, despite being an important feature of corporate
strategy over the past 20 years. In addition, very little empirical work to date has been able to distinguish
between incremental and radical innovations in practice, and only in very limited contexts.
My research setting has the unique advantage of allowing me to observe the reactions of the same
firms at the same time to both incremental and radical innovations – and thus compare the market share
effects between the two to establish the importance of this distinction. This test depends critically,
however, on being able to map the theory on incremental vs. radical innovation into the novel business
process setting.
My approach relies on first developing a basic framework for distinguishing radical business
process innovations from incremental ones, and then integrating it with an in-depth qualitative analysis of
the phenomenon. I argue that the higher the strategic importance and complexity of the underlying
business process to be affected, the more radical any proposed change will be Careful qualitative analysis
reveals that, despite a reliance on a common technology platform, e-buying and e-selling in 1999 differed
dramatically along these dimensions. At the time of the Census survey, e-buying concerned electronic
procurement of highly standardized materials such as maintenance and office supplies that, while they
impacted the cost structure of the firm, were not unduly complicated to acquire or strategically sensitive
for the majority of manufacturing firms. E-selling, by contrast, governed not only a far greater complexity
of products but also critical business processes such as sales and customer management that faced
disruption and channel conflict in the switch to internet-based sales. As such, e-buying constituted a
relatively incremental process innovation, while e-selling was far more radical.
I find robust statistical evidence consistent with the hypothesis that market leaders would pursue
the incremental e-buying innovation. In this large, multi-industry sample of firms, firms with the greatest
market shares, sales and profits were significantly more likely than laggards to adopt the internet as their
primary platform for online procurement

3

---

I find the exact opposite in the case of e-selling: market leaders were far less likely than smaller
or less-successful competitors to embrace internet-based sales. While many of the same influences (such
as economies of scale) that promote e-buying adoption by leading firms ought to apply in the e-selling
case, the economic and/or organizational disruption of this innovation is sufficient to induce a
commensurately negative relationship between market leadership and adoption likelihood. Moreover, this
pattern holds even controlling for potentially confounding factors such as other related IT investments and
underlying productivity. This result is consistent with the hypothesis that market leaders will be less
likely to adopt radical business process innovations.
Thus, this paper contributes robust, multi-industry evidence that market leaders are more likely to
adopt certain types of innovations and sheds light on precisely when this is likely to happen (or not).
Moreover, the process of refining and extending existing theory into the novel business process context
yields the insight that many of our product-based intuitions transfer well to the business process setting:
market leaders are far more likely to adopt incremental business process innovations, while radical
changes may encounter resistance amongst the most successful firms.
With evidence on whether and when, the remaining question is why – what specific attributes of
innovations or their potential adopters lie at the root of this behavior? Additional analysis reveals
important insights about the innovations themselves. The qualitative differences between e-buying and e-
selling are consistent with my proposed definition of business process innovations as either incremental
or radical based on whether the process was strategically sensitive (and therefore risky to change) and/or
relatively complex (and therefore costly to change). However, an additional criterion of “radicalness”
proves necessary: whether the underlying business process spans the firm boundary. Leveraging the rich
Census survey, I compare market share effects across a wide range of e-business processes with different
characteristics. My findings support the proposition that a combination of complexity, strategic sensitivity
and boundary-spanning create the particular challenge observed for market leaders in the case of e-selling
– and that all three may be required to make a business process innovation truly radical for leading firms.
This enhanced understanding of why market leaders may enjoy certain advantages or face
additional challenges in business process innovation has important practical implications. Being able to
apply what we know about the innovative behavior of firms requires the means to distinguish, ex ante,
whether a proposed change lies on the incremental or radical end of the spectrum. While far more work –
both empirical and theoretical—is needed to place different types of innovations in a comprehensive
taxonomy, the suggestive evidence on what makes e-selling radical is a step in this direction. My results
indicate that, in cases where firms must coordinate complex and strategically important activities
throughout a larger organization, across more establishments, while spanning the firm boundary, adoption
will tend to take place primarily among smaller firms. This inter-firm coordination challenge is an

4

---

important strategic consideration as businesses grow ever more dependent on the performance of their
extended value chain for success. Lagging firms may discover new opportunities to leapfrog their larger
competitors using certain business-to-business process innovations and IT-enabled supply chain
relationships.
This paper proceeds as follows. Section 2 develops the theory and places this paper in the
relevant literature. Section 3 delves into the details of the phenomenon, forming the link between the
context and the hypotheses. Sections 4 and 5 present the data and econometric model, respectively.
Section 6 reports on the empirical the drivers of e-business adoption. Section 7 refines the definition of
radical business process innovation. Section 8 concludes.

2.
Literature and Theory Development
2.1
Overview

A rich body of work in both economic and organizational theory addresses the question of how
incumbent firms with large market shares will react to –and participate in –technological change. While
the inquiry dates back at least to Schumpeter (1934, 1942), Arrow’s seminal (1962) theory that existing
monopolists will resist innovating to avoid cannibalizing existing sales sparked a surge of research into
the effects of market structure and firm size on innovative activity.1 The results of these models predict
highly contradictory outcomes depending on a range of considerations such as the degree of competition
(Boone 2000, Aghion et al. 2005), the availability and use of property rights (Dasgupta and Stiglitz 1980)
or the nature of the discovery process (Reinganum 1983, Doraszelski 2003) – many of which are
remarkably difficult to measure in practice (Gilbert 2006).
Recent theoretical advances have focused less on market structure itself than on a firm’s position
in the market (e.g., in terms of market share or cost effectiveness) and how it influences the incentives to
innovate. The most recent and generalizable models (e.g., Athey and Schmutzler 2001) predict
“increasing dominance”, whereby leading firms invest to shore up or improve their favorable market
position, in a wide variety of cases. However, exceptions do exist, and cogent arguments can be found
predicting a greater propensity to innovate among both market leaders and lagging firms.
The related empirical literature, likewise, is “notable for its inconclusiveness” (Cohen and Levin
1989). In part, this is due to econometric problems such as poor treatment of industry heterogeneity and

1 Reviews of this sizeable literature are provided by Kamien and Schwartz (1982), Baldwin and Scott
(1987), Reinganum (1989), and Gilbert (2006).

5

---

sample selection bias.2 Additionally, a failure to distinguish product from process innovation or address
the simultaneous determination of market structure and innovation have limited the contribution of this
work (Gilbert 2006). Finally, far less attention has been paid to the adoption of existing innovations than
to R&D expenditure or patenting behavior, despite well-known problems with these latter measures of
innovative activity (Blundell et al 1999).
The dearth of clear predictions and evidence is even greater when it comes to considering the
adoption of business process innovations. Existing innovation research has focused overwhelmingly on
the invention and commercialization of new products.3 Process innovation, by comparison, has received
far less attention (Rosenberg 1982), particularly in the scholarly literature. Business process innovation,
has received even less, despite being an important feature of competitive strategy over the past 20 years.
This class of innovative activity warrants closer attention in light of theory and evidence that
process innovation and its drivers may differ in important ways from innovation geared toward new
product introduction. Factors external to the firm such as product lifecycles (Utterback and Abernathy
1975), degree of competitive pressure (Boone 2000) or customer requirements (Adner and Levinthal
2001) are believed to emphasize process innovation over product enhancement in certain cases. Internal
factors such as scale of output (Cohen and Klepper 1996, Klepper 1996) and organizational routines and
priorities (Henderson et al. 1998) may also create incentives to follow different innovation paths. Firms
that excel in one may struggle with the other, with important implications for competitive outcomes.
Given the state of the literature, a key challenge of my research design is to distill meaningful
testable hypotheses that fit well in the business process setting. To overcome this challenge, I leverage
insights from the economics and strategy literature on radical versus incremental innovation to distinguish
predictions based on the nature of the innovation in question. To supplement the theory development, I
also borrow insights from prior economics research on process innovation.

2.2
Incremental vs. Radical Innovation
According to prior work, leaders will be more likely than lagging firms to innovate when it
enhances or builds on their existing activities. While this ought not to apply to innovations that are
“drastic” (Arrow 1962) or “radical” (Henderson 1993) in the sense that they render existing technologies
or knowledge obsolete, theory predicts leading firms will tend to enjoy both greater incentive and greater
ability to pursue incremental innovations than will smaller rivals (Henderson 1993).

2 A notable exception is Blundell, et al (1999) which finds a strong positive relationship between the
market share and stock market value of publicly traded U.K. firms and their incentives to patent their
discoveries.
3 Exceptions are discussed in section 2.

6

---

On the incentive side, economics and strategy research emphasizes the strategic benefits to a
leader of maintaining its dominant position in the face of technological change. Because a large firm with
some degree of market power will enjoy higher rents per unit of sales than will a smaller firm, it will have
greater marginal incentive to protect that power. Thus, leaders will be more likely to invest in a new
technology (such as a new process innovation) that allows them to lower production costs and thus price
more aggressively in the product market (Fudenberg and Tirole 1984, Sutton 1991) or otherwise pre-empt
potential rivals from entering the market (Gilbert and Newbery 1982).4
Strategic interactions aside, other economics research suggests that straightforward economies of
scale will increase the benefits of process innovation among larger firms relative to smaller ones (Cohen
and Klepper 1996, Klepper 1996). This is because they can spread fixed investment costs over a greater
volume of sales and better appropriate the value of lower production costs in their own operations. The
greater the market share, the greater this benefit. Thus, barring any adjustment costs that increase
disproportionately with firm size, market leaders will tend to enjoy higher net benefits from process
innovation.
In addition to being more willing, leaders will tend to be more able to pursue incremental
innovations. A rich organizational theory literature argues that firms tend to develop routines (Nelson and
Winter 1982) and information filters (Arrow 1974) based on prior experience that embody organizational
knowledge and fundamentally condition how they react to changes in their environment. As a result, the
experience and knowledge gained over a larger scale of activity and/or over time may make it easier for
larger incumbents to identify and pursue innovations that build on existing advantages and extend
existing knowledge (dubbed “competence-enhancing” innovations by Tushman and Anderson, 1986).
Larger firms may also have invested more in resources (Wernerfelt 1984) or capabilities (Barney 1991)
that enable them to more effectively exploit incremental advances.
Combining incentives and abilities together, firms with larger market shares have significant
advantages over smaller firms when it comes to incremental process innovation. While additional details
of the setting will be needed to determine which observable business process changes fall into this
category of incremental innovation, the basic hypothesis to be tested is the following:

Hypothesis 1: Firms with larger market shares will be systematically more likely to adopt
incremental business process innovations, all else equal.

4 Athey and Schmutzler (2001) develop a very general model of how market position influences
investment in cost-reducing or demand-enhancing technology. Their contribution is a framework that
subsumes much prior work as special cases and predicts leading firms will be more likely to invest under
a wide range of common assumptions. 

7

---

In sharp contrast, radical innovations pose greater problems for larger firms. Innovations that
threaten or destroy firms’ existing competencies (Tushman and Anderson 1986), that are organizationally
“radical” (Henderson 1993, Henderson and Cockburn 1994) in the sense that they demand entirely new
knowledge or ways of processing information, or that are otherwise disruptive (Christensen 1997) to their
existing product market offerings, will face greater resistance. Larger firms that have survived longer or
grown faster than rivals will tend to have made greater investments in organizational routines, procedures,
and resources that embody their acquired know-how and lower the costs of day-to-day operations. Thus
any innovation that requires the removal or replacement of those existing organizational capabilities is
disproportionately more difficult and/or costly for larger firms.
Economic theory weighs in here, too, in terms of the opportunity costs larger firms face due to
disruption of its production. To the extent that larger firms may enjoy higher rents per unit of production
due to their greater market power, any disruption or replacement of existing sales due to the
implementation of a new technology or process will be relatively more costly for larger firms than for
smaller ones. Thus, we have the complementary hypothesis:

Hypothesis 2: Firms with larger market shares will be systematically less likely to adopt
radical business process innovations, all else equal.

2.3 Classifying
Innovations
While a useful conceptual tool for grappling with the contradictory theory literature, this
distinction between incremental and radical innovation has found limited empirical support to date.
Meaningful differences can be found only in single-industry empirical studies in food packaging (Ettlie et
al. 1984), footwear manufacturing (Dewar and Dutton 1986), and the photolithographic alignment
equipment industry (Henderson 1993). Other, multi-industry empirical work has failed to find evidence of
this distinction (Blundell et al. 1999).
One reason for the lack of systematic cross-industry evidence may be the challenge involved in
defining a given innovation as either incremental or radical in a way that transcends a particular industry
or firm context. Prior innovation research is of limited definitional help in the business process context
because it either focuses on product innovation5 or tends to equate process innovation with incremental
change (e.g., Cohen and Klepper 1996, Klepper 1996, Bonanno and Haworth 1998, Boone 2000). Notable
exceptions include Tushman and Anderson (1986), Dewar and Dutton (1986), and Sull et al. (1997),
which consider the possibility that process innovation can be radical and thus threatening to market

5 See Pisano (1997) for a notable exception.

8

---

leaders. However, these exceptions are largely based on industry-specific anecdotes and tend to focus on
process innovations that dramatically shift the product attributes that are achievable (e.g., Pilkington float
glass in Tushman and Anderson 1986, or radial tires in Sull et al. 1997) and/or represent radical
departures in the mechanical or scientific knowledge required by a firm. Business process innovations, in
contrast, can quietly transform internal firm operations with little or no visible impact on product-market
offerings and interactions. This makes them extraordinarily difficult to classify, ex ante.
My empirical exercise requires, however, that I do just that. The first basis for classification I
consider is the strategic sensitivity of the business process undergoing transformation. In order for
changes to a process to pose the substantial risks or costs described in the radical innovation literature, the
underlying process must first be strategically important to the firm. The opinion of experts in the practice
of IT-enabled process innovation also supports the notion that these core activities are the hardest ones to
change (Davenport 1993). While a given business process may vary in strategic importance from firm to
firm, we may be able to classify certain processes as typically more sensitive or not across a broad sample
of organizations.
Another criterion – complexity – comes from prior research on the economics of IT adoption. The
adoption of information technology quickly becomes an exercise in business process innovation when
significant investments in time and money are required to match the technology to internal firm activities
– or, more importantly, to match internal firm activities to the technology. Often, these undertakings
entail substantial uncertainty and risk for the adopting firm, particularly if they touch on strategically
sensitive processes that determine a firm’s competitive advantage. Evidence suggests that these “co-
invention” costs and risks are systematically higher wherever there is a greater complexity of processes
(Bresnahan and Greenstein, 1996). Thus changes to more complex firm processes ought to be more
radical than simpler adjustments.
A final dimension of this classification concerns the organizational structure of the underlying
business process. Managing change within an organization is sufficiently challenging when the primary
stakeholders function with the boundaries of the same firm. Hierarchy and shared culture help facilitate
both decision-making and the implementation of process innovations. These advantages do not exist,
however, when core business processes cross firm boundaries. When core activities of the firm require
tight integration with other members of the value chain – especially customers, then the costs and risks
associated with redesigning them increase dramatically. The greater the number of inter-firm linkages,
the greater the difficulty (Davenport, 1993).
These criteria do not provide an exhaustive definition of radical business process innovation.
However, they are both suggestive and prove useful in the next section, where I delve into the e-business
phenomenon to understand the nature of the e-buying and e-selling innovations and how they map into

9

---