entry choices hight-tech.pdf

7/25/2019 entry choices hight-tech.pdf

1/31

The International Market Entry Choices

of Start Up Companies in High

Technology Industries

For a young, resource-co nstrained, technology-based start-up

emba rking on intemational sales, the choice of entry mode is a

strategic decision of major importance . Yet within the emerging

research stream of intemational entrepreneurship, curiously lit-

tle attention has been devoted to the empirical analysis of for-

eign marke t entry forms. The authors address this important

issue by analyzing the determinants of 398 export decisions

taken from a

U K

survey of 246 technology-based start-ups with

intemational activities. The findings show that the entry m ode

decision is necessarily a trade-off hetween the resources avail-

able and the support requirem ents of the customer. Issues of the

innovativeness of the technology and the historic channel expe-

rience of the firm in its domestic market

are

particularly strong

determinants of mode choice. The authors suggest that an

orga-

nizational capability perspective on these firms behavior offers

a better explanation of their entry decisions than either transac-

tion cost

or stage

theory.

During the past decade, the phenomen on of globalization bas

received considerable attention and bas shaped botb tbe dis-

course and tbe actions of managers, policymakers, and acad-

emics. Researcb on the topic is rooted p rimarily in sociology,

cultural studies, political science, and economics. A com-

mon denominator is the definition of globalization as a

process in w hich the constraints of geography on social, cul-

tural, political, and economic arrangements recede (Waters

1995). One manifestation of a globalizing world is the emer-

gence of entrepreneurial start-ups that have an international

outlook from inception. Some of these firms even carry out

different activities along their value chains in different coun-

tries and thereby blur tbe national identity of the firm. The

international activities of these exceptional young firms

labeled infant mu ltinationa ls (Lindqvist 1991), born glob-

als (Knigbt and Cavusgil

1996;

McKinsey Company 1993),

and international new ventu res (Oviatt and McDougall

1994)have received an increasing amount of attention from

researchers over the past couple of years. Yet relatively little

attention has been devoted to tbe empirical analysis of for-

eign entry modes. Most studies provide descriptive informa-

tion on the chosen entry modes, but witb tbe exception of

Lindqvist (1991), Bell (1995), and Shrader, Oviatt, and Mc-

Dougall (1997), few researchers have attempted to explain

A B S T R A C T

Oliver Burgel and

Gordon C. Murray

Submitted December 1998

Revised June 1999

Jouma l of ntemational Marketing

Vol 8, No. 2, 2000, pp. 33-62

ISSN10B9 031X


2/31

PREVIOUS RESEARCH

IN INTERNATIONAL

ENTREPRENEURSHIP

the choice of entry modes. This omission is even more sur-

prising given that it is vy^idely acknow^ledged that the initial

foreign entry behavior of

a

young firm can he of major impor-

tance in its future economic success.

Empirical studies in international entrepreneurship have

heen hoth exploratory and , less frequently, explanatory in na-

ture. Whereas some have had an explicit focus on interna-

tional husiness theories, others have focused more on general

performance issues. Qualitative stud ies have described prod-

uct characteristics (Jolly, Alahuhta, and Jeannet 1992; Miuray

1996; Roberts and Senturia 1996), market entry forms (Jolly,

Alahuhta, and Jeannet 1992; McDougall, Shane, and Oviatt

1994;

Roberts and Senturia 1996), and characteristics of

founders and key employees (Boter and Holmquist 1996; Mc-

Dougall, Shane, and Oviatt 1994; Mvirray 1996; Roherts and

Senturia 1996). Quantitative studies we siuveyed have ana-

lyzed structural characteristics of the firms, including age,

size, and technology intensity (Lindell and Karagozoglou

1997; Lindqvist 1991), market entry forms (Bell 1995;

Lindqvist 1991; Shrader, Oviatt, and McDougall 1997), geo-

graphic spread of foreign sales (Bell 1995; Shrader, Oviatt,

and McDougall 1997), the relation between strategic orienta-

tion and growth/profitahility (Bloodgood, Sapienza, and

Almeida 1996; McDougall and Oviatt 1996; Shrader, Oviatt,

and McDougall 1997), characteristics of founders and key

employees (Bloodgood, Sapienza, and Almeida 1996), and

the role of risk in internationalization decisions (Shrader,

Oviatt, and McDougall 1997).

Three empirical studies that investigate entry modes in more

detail stand out. Lindqvist (1991) reports that the preferred

entry modes of the Swedish firms in her sample were direct

exporting and foreign sales through intermediaries such as

agents and distrihutors. However, she does not investigate the

determinants that influence the choice of entry modes. In-

stead, she explores the propensity to use sales subsidiaries.

Here, she finds no or weak relationships between entry mode

choices and product characteristics, research and develop-

ment(R D)intensity, or market size. Lindqvist's findings sug-

gest that internationalization could be understood as

jumping a threshold ; that is, when the firms have decided

to operate internationally, their choice of entry modes is not

influenced by structural characteristics, such as size andR D

intensity. In Bell's (1995) study of the in ternational operations

of Irish, Finnish, and Norwegian software firms, 70% of all

sales transactions were carried out through either direct ex-

ports or agents and distributors. Few firms engaged in foreign

direct investment, and when this did occur, they were preoc-

cupied with setting up marketing and sales suhsidiaries.

Firms that commercialized highly customized products al-

most exclusively relied on direct exporting, whereas firms


3/31

that sold standard, off-the-shelf products were more likely to

use sales intermediaries. Shrader, Oviatt, and McDougall

(1997) argue that different entry modes represent different de-

grees of resource commitment and therefore risk for the firm.

They argue that firms actively manage their risk in interna-

tional operations by balancing entry m ode risk, country risk,

and risk from overdependence of foreign markets. Their em-

pirical findings on

U.S.

irms

hat high-risk countries were en-

tered using low-risk/low-commitment market entry modes

and vice versa support their theoretical argument.

Two theoretical articles analyze the structural aspects of in-

ternational activities in more detail. Oviatt and McDougall

(1994) examine internationa l new ventures (INVs), a type of

firm that from inception, seeks to derive significant compet-

itive advantage from the use of resources and the sale of out-

puts in mu ltiple coun tries (Oviatt and McDougall 1994, p.

49).They argue that sustainable INVs must have an innova-

tive organizational structurefor example, assuming a con-

trolling position in a network or using hybrid arrangements

involving subcontractors and intermediate sellers. More di-

rectly related to the choice of entry mode, Zacharakis's (1997)

contribution explicitly deals with the choice of export agents

and distributors to enter international markets. Zacharakis

only considers the use of intermediaries, because he claims

that direct exporting is not an appropriate choice for

start-ups. Civen the wealth of studies on exporting (for a

recent review, see Leonidou and Katsikeas 1996) and interna-

tionalization processes (Johanson and Vahlne 1990) that ind i-

cate that direct exporting is actually a widespread entry

mode of small firms, we believe that Zacharakis's perspective

is too constrained to capture the full scope of decisions that

managers of a new venture will face.

These contributions represent important first steps in ex-

plaining the entry decisions of internationally operating

start-ups. These studies have several common elements.

The firms investigated predom inantly operate in high-tech-

nology industries. In addition, the majority of the studies

have an explicit focus on young firms. Findings from quan-

titative surveys agree that the preferred entry modes of tech-

nology-based start-ups are characterized by relatively low

resource commitment and are directed toward commercial-

ization rather than foreign production. However, they also

report that technology-based start-ups have a tendency to

enter several foreign markets within a short time span. Most

articles based on a case study methodology include firms

that made bold commitments to international operations

during their early years. Rapid and sometimes resource-

intensive market entries into different countries were de-

scribed. Some extreme cases were international from incep-

tion and performed different activities along their value


4/31

cbains in different countries McDougall, Sbane, and Oviatt

1994). Together, these findings contrast previous research

that examines the international activities of more tradi-

tional small and medium-sized enterprises Bamberger and

Evers 1994; Leonidou and Katsikeas 1996). However, they

also leave questions about the determinants of the choice of

a particular entry mode largely unaddressed. Acknowledg-

ing the dearth of empirical studies in this area, Oviatt and

McDougall 1997) conc lude that there is a need for further

work to analyze the prevalence and determinants of interna-

tional entrepreneurship.

pjjg purpose of this study was to use a large data set to deter-

RESE RC H OBJECTIVE ND

mine wh at mo des of foreign market entry our resp ond ent

H Y PO T H E SE S firms chose and wh at were the primary reasons for their

elected choices. The modeled entry decision was simplified

to the choice of selling abroad either by direct exporting or

through the use of distributors. Our reasons for this choice

are both pragm atic and the oretically justified. First, we re-

ceive empirical support from previous studies, which report

that exporting and the use of intermediaries are in reality the

two predominant alternatives employed by entrepreneurial

bigb-tecbnology firms Bell 1995; Lindqvist 1991). In sbort,

our model mirrors wbat small firms actually do.

Second, the choice between direct exporting and tbe use of

distributors is of utmost managerial relevance for technology

entrepreneurs. W ith the singular exception of the U.S. econ-

omy, the finite market opportunities in many countries may

not justify the development expenditures for certain highly

specialized niche technologies unless international expan-

sion is considered from inception . Technology-based start-

ups therefore face a dangerous dilemma. On the one hand,

they may be forced to venture abroad to belp amortize their

initial development exp enditures and generate sufficient rev-

enues to finance ongoing development activities. On the

other hand, because many technology-based start-ups experi-

ence negative cash fiows during their early years, they may

lack the necessary human and financial resources required

for the effective commercialization of their products on their

own. Given these resource constraints, identifying end cus-

tomers and providing pre- and after-sales support services

may be handled better by a local partner. Tbe downside of

this arrangement is that revenues must

e

shared between the

start-up and the distributor. Additional costs can be incurred

by the start-up because of tbe need to provide tecbnical train-

ing and the creation of incentives and monitoring mecha-

nism s. Early on, technology entrepreneurs therefore must

make complex and highly strategic trade-offs, because the

choice of the foreign sales mode may have profound implica-

tions for botb costs and revenue generation.


5/31

Third, the main theoretical frameworks in the field of inter-

national business that analyze entry modes come to different

conclusions when they are applied to firms that at the same

time are young (or inexperienced) and operate in high-tech-

nology sectors. Process models (Johanson and Vahlne 1977,

1990) consider commitment to internationalization a func-

tion of experiential knowledge of foreign markets. Accord-

ingly, a start-up company would be expected to gain initial

experience through reactive exporting before proactively

venturing into foreign markets. The choice between direct

exporting and the use of more complex and proactive entry

modes thus depends on firm experience and foreign market

knowledge. Despite having received empirical support,

process models and the subsequently developed stage

mo dels of internatio naliz atio n (e.g., Cavusgil 1980) are

frequently criticized for being too deterministic and for fail-

ing to take firm-specific factors other than experience into

account (Andersen 1993).

A rival approach, the transaction cost economics (TCE) ap-

proach, has been influential, as it provides a decision rule

with regard to individual entry decisions. Firms are ex-

pected to choose the governance or entry mode that mini-

mizes the costs of carrying out particular transactions. In its

application, TCE is concerned with comparing different in-

stitutional arrangements for carrying out economic activity

(Williamson 1985). As the choice between direct exporting

and involving foreign distributors is essentially a choice be-

tween an internal arrangement and an arrangement involv-

ing an external third party, the tools of TCE are applicable

to mod el the decision. Yet the TCE approach assumes a ca-

pacity for discretionary resource deployment. For example,

the com mercialization of a produc t incorporating advanced

technology may require a high degree of asset specificity. If

this asset specificity leads to high costs of involving a dis-

tributor, a firm is expected to switch from direct exporting

to a sales subsidiary when its foreign sales to a particular

country exceed a certain level. However, this option may

not reflect the reality of resource constraint start-ups. They

may rather be inclined, or obliged, to establish collaborative

relationships with intermediaries to get access to assets, re-

sources, and capabilities they do not own. Cooperation may

not be a choice but an imperative for the young firm. The

widespread use of collaborative strategies in technology-

intensive industries has triggered the development of the

organizational capability (OC) perspective. Its proponents

argue that OC, compared with TCE, represents a superior

approach for the analysis of collaborative governance

arrangements (Madhok 1997). To date, we do not know of

any study that has applied the OC perspective to the inter-

nationalization of smaller, entrepreneurial firms.


6/31

The key elements of these three competing internationaliza-

tion theories (stage, TCE, and OC) can be isolated as separate

and independent sets of variables in our decision model. Ac-

cordingly, they are incorporated in our hypothesis construc-

tion. The employment of an integrated multivariate approach

should provide further evidence whether or not these sets of

variables explain the choice of entry modes made by high-

tech start-ups. Following established research practice, we

have chosen the entry decision and not the firm as the unit of

analysis. Similar to many other researchers, we argue that en-

try decisions are a function of firm-specific factors, product-

specific factors, and target country-specific factors (Cavusgil,

Zou, and Naidu 1993; Erramilli and Rao 1993). Essentially,

the choice between direct exporting and sales through

distributors is a choice between an internalized transaction or

an externalized transaction involving intermediaries. We find

ourselves within the established tradition of researchers that

have conceptualized entry modes choices as binary (see, e.g.,

Barkema and Vermeulen 1998; Davidson and McFetridge

98 ;Erramilli and Rao 1993; Kogut and Singh 1988). Ac-

cordingly, we develop a set of hypotheses to test which vari-

ables account for the choice between direct exporting and

exporting through distributors as the two predominant modes

of foreign market entry for technology-based start-ups.

^^ ^^^^^^^^^

This discussion of our research objective leads to the genera-

ypoth s s tion of seven related or closely associated hypotheses per-

taining to those factors that most instrumentally influence

the choice of channel mode used in entering first foreign

markets. As has already been stressed, the choice for the

young firm of either going alone (direct exporting) or allying

with a partner (using a distributor) is of pivotal strategic im-

portance given the risk/reward implications of this decision.

Direct exporting gives the firm autonomy and control but de-

nies it the assets (including networks) and experience of the

distributor. Alternatively, to contract with a distributor raises

the sensitive issues of sharing profit margins. The latter is

likely to be influenced by the distribution of bargaining

power within the manufacturer-distributor relationship

(Zacharakis 1997). From the local distributor s perspective,

the smaller the foreign manufacturer and the lower the po-

tential sales volume, the less profitable the medium-term re-

lationship is. Below a certain projected sales volume, the

decision to provide the service will not be bilateral but will,

in practice, be solely that of the distributor. Aligning the ob-

jectives of the two parties has been studied widely in a do-

mestic context (e.g., Anderson and Narus 1990; Heide and

John 1994). Yet in a foreign country, the task of monitoring a

distributor is likely to be even more difficult for a small en-

trepreneurial firm. Therefore, selling through an established

distributor relationship is likely to require higher up-front in-

vestment than exporting. In the particular case of technology-


7/31

based start-ups, tbe availability of resources is probably an

even more crucial predictor of subsequent firm actions

(Oakey, Rotbwell, and Cooper 1988). Accordingly, we argue

that this route to foreign markets will be employed by start-

ups commanding greater resources.

Hi:

Firms that sell into foreign markets through inter-

mediaries are larger than firms that export directly

to end customers.

related hypothesis pertains to the international experience

of the firm. The internationalization process perspective ar-

gues that firms increase their commitment to international

sales over time as their experiential knowledge of foreign

markets increases (Johanson and Vahlne 1977, 1990). Ac-

cording to this logic, firms that are more experienced in in-

ternational sales are expected to use more resource-intensive

entry modes. This view also posits that, over time , firms that

start international sales with low-commitment entry modes

switch to entry modes that require a higher degree of com-

mitment as a result of better knowledge of foreign markets.

Formalizing a distribution agreement frequently involves le-

gal and other ad min istrative costs. The distributo r s staff

must be trained in selling tbe product, installing it, and pro-

viding subsequent maintenance or upgrades. Tberefore, we

believe that the use of distributors represents a more commit-

ted entry m ode that is used by more experienced firms.

H2:

Firms that sell into foreign markets through inter-

mediaries are more experienced in international

operations than firms that export directly to end

customers.

It is, however, not necessary that experience in international

operations has to be gained by the firm as organizational en-

tity. A young firm initiating its first market en tries cannot be

expected to bave direct experiential knowledge of interna-

tional operations embodied in its processes and routines.^ In

tbe case of start-ups, the founders interna tional experience

can be a substitute for organizational experience. If key man-

agers in tbe young firm have previous experience in foreign

markets, it may initiate cross-border activities using more

complex entry modes. Accordingly, the international experi-

ence of senior managers has in the past been used to predict

the scale and scope of international activities of start-up

firms (Bloodgood, Sapienza, and A lmeida 1996; Reuber and

Fiscber 1997).

H3: Managers of firms tbat sell into foreign markets

tbrougb intermediaries will be more likely to have

international experience than managers of firms

that export directly.


8/31

The experienceof tbefirm isalsoa keyvariableinresearcb

contributions tbat apply

the

OC perspective

to the

choice

of

entry modes (Anlakh

and

Kotabe 1997; Madbok 1997). Theo-

ristsofOC argue that present outcomes arestrongly influ-

enced

by

past experiences

and

routines that bave become

embedded

in the

organization (Madhok 1997).

Ifa

firm uses

a

particular sales channel

in its

domestic market,

it may be ex-

pected

to

replicate tbis familiar practice (i.e., negotiating

contracts, incentivizing,

and

motivating intermediaries)

in

foreign markets. Thus,

tbe

higher costs

or

risks

of

arranging

more complex foreign sales modes

can

arguably

be

reduced

through leveraging experiences gained earlierin adomestic

market.^

H4:Firms will sell into foreign markets tbrougb inter-

mediaries rather than export directly

if

they already

use distributors

for

tbeir domestic sales.

The OC perspective argues that

a

firm's value-creating activi-

tiesare afunction ofits resourceandcapability base (Madhok

1997;Teece, Pisano,andShuen 1997). Madhok distinguishes

between activities

in

which

tbe

focus

is

capability develop-

ment

and

activities

in

wbich

the

focus

is tbe

exploitation

ofanexisting advantage.In tbecaseofstart-upsgivenour

focusonsales modeswe argue thattheobjectiveoftbeirin-

ternational sales

is to

exploit fully

the

commercial value

of

tbeir technological competency to ensure tbeir survival.

Mad-

bok also introduces the notionsof ownership effect and lo-

cational effect.

Tbe

former

is

represented

by tbe

ratio

of

embedded-to-generic firm-specific know-how, whereas

tbe

latter

is

defined

as tbe

ratio

of

embedded-to-generic market-

specific knowledge (Madbok 1997). AccordingtoMadhok,a

firm will carry

out a

transaction itself (internalization)

if

tbere

is

a

high potential

for tbe

erosion

in the

value

of tbe

firm's

know-how stemming from

the

ownership effect.

In

contrast,

a

firm will haveapreference forcollaborationifthereis abigb

potential

for the

erosion

in the

value

of

the firm's know-how

from

the

locational effect.

In

summary, collaborative arrange-

ments will

be

preferred

in

countries where

the

idiosyncratic

waysofdoing business erodethevalueoffirm-specific know-

bow. Conversely,

a

firm

is

less likely to involve tbird parties

if

firm-specific know-bow

is

inimitable

or

immobile, whicb

thereby makes

the

sharing of routines witb intermediaries

dif-

ficult (Hill, Hwang,and Kim1990; Madhok 1997). Tbissug-

gests tbat

a

firm will have

a

higher propensity to avoid the

use

of intermediaries

if

its technology

is

advanced

or if

the firm

is

unfamiliar with potential users

in its

target market. Civen

the

importance oftacit knowledge for such products, market-

based support infrastructures may

not

be effective

or

available

(Meldrum 1995). Therefore, effective commercialization

may

only

be

possible

by

internalization

of

the sales process. Idio-

syncratic ways

of

doing business abroad

may

therefore repre-


9/31

sent a lesser barrier than in tbe case of more established tech-

nologies, especially if tbe product is sold to industrialized

countries. Collaborations are expected to occur more fre-

quently when tbe technology is more mature and established,

a proposition that has received some validity in p rior researcb

on international technology transfers (Davidson and

McFetridge 1985).

H5:

Tbe products of firms that sell into foreign markets

through intermediaries are technologically more

matvu:e tban those of firm s tbat export directly.

Previous research on internationalization bas found that

product characteristics affect the way firms manage their in-

ternational activities (Cavusgil and Zou 1994; Cavusgil, Zou,

and Naidu 1993). Product characteristics have also been

found to influence tbe chosen entry modes in the case of

young technology-based firms (Lindqvist 1991). Tbe impor-

tance of client-specific customization as a barrier to interna-

tionalization has been reported in studies that examine

produc t characteristics of international start-ups (Lindell and

Karagozoglou 1997; Murray 1996; Roberts and Senturia

1996). Tbe transformation from offering customized technol-

ogy solutions to offering standardized or shrink -wra pped

products in whicb the technology is embedded is associated

witb an increasing market orientation of the firm (Roberts

1991). Companies whose pro ducts are tailor m ade for partic-

ular customers bave been found to be more likely to sell

directly witho ut involving interm ediaries (Bell 1995). Fur-

thermore, it is more probable that the technological skills

required to tailor a particular product to the needs of a cus-

tomer reside within tbe company that developed the product

rather than w ith the distributor in a target country.

Hg: The products of firms that sell into foreign markets

through intermediaries require less client-specific

customization tban those of firms that export

directly.

The tools of TCE have also been widely used to analyze tbe

determinan ts of entry mode choices. In essence, TCE is con-

cerned with finding tbe most efficient institutional or con-

tractual arrangement for economic transactions (Hennart

1989). Instead of following tbe m ethods used by researchers

who measure transaction costs indirectly by defining situa-

tions in wbich asset specificity and uncertainty are suppos-

edly high (see, e.g., Anderson and Catignon 1986; Hennart

1990), we follow the approach of Klein, Frazier, and Rotb

(1990) and attempt to measure tbe costs involved in tbe sell-

ing process of high-tecbnology goods directly. During tbe

sales process for high-tecbnology goods, the vendor may be

required to spend considerable time advising and educating


10/31

the potential customer on the key features and relative merits

of the product. After the sale, more complex products may

require installation by trainedstaff, regular after-sales ser-

vice,

and periodic upgrades (Cavusgil and Zou 1994; Hutt

and Speh 1992). To carry out these tasks, the sales and tech-

nical staff of the vendor will require particular skills. In the

case of exporting, these skills are normally resident within

the manufacturer. Selling a product through an intermediary,

however, requires those skills also to be already present

within, or transferable to, the distributor. The producer may

therefore need to provide regular training to tbe staff of tbe

intermediary to transfer the necessary skills and routines and

effectively support the product. High up-front investments

into these specific assets reduce tbe subsequent bargaining

power and margins of the party that incurs these costs.

Zacharakis (1997) argues that a small entrepreneurial firm is

more likely to be obliged to reduce the distributor's set-up

costs through the provision of training or other transfers than

a large established multinational. Furthermore, a distributor

may not be motivated to push a complex product that re-

quires substantial presale consulting and installation efforts.

Therefore, products and services that incur substantial costs

during the sales process should make it more difficult to

align the interests of the start-up and a potential distributor.

However, selling a technologically advanced product is

likely to require an effective support infrastructure (Meldrum

1995).

Using distributors tbat can exploit economies of scale

and scope not available to tbe young firm may be the only

way to provide tbe necessary infrastructure to service foreign

customers. Furthermore, the costs of learning how to perform

the relatively standardized tasks of installation, end-user

training, and maintenance are likely to be relatively low

when the distributor already has a portfolio of related prod-

ucts in place. Therefore, we hypothesize that in international

markets, distributors represent the preferred vehicle for start-

ups to ensure effective customer support for products whose

commercialization is resource intensive.

H7:The pre- and after-sales transaction costs of prod-

ucts sold into foreign markets through intermedi-

aries are higher tban the costs of products of firms

that export directly.

j ~

TTj

^ 6 include R D intensity as a control variable. It has in the

ontrol Variables past been used as a proxy variable in studies that have ap-

plied the framework of transaction costs to international mar-

ket entry choices (Davidson and McFetridge 1985) to

operationalize asset specificity and information asjnnmetries

in exchange relations. We argue that there are two problems

with this measure. First, it constitutes an input variable and

may not necessarily have an impact on the asset specificity

required to commercialize the output of the firm. Higb R D


11/31

expenditures may not give rise to asset specificity or infor-

mation asymmetries per se. Indeed, higherR Dinvestment

may even allow a reduction in transaction costs by designing

out complexity for the customer. Second, it is usually mea-

sured at firm level, not at prod uct level, because it is difficult

to obtain

R D

intensity (R D expen diture divided by sales)

on a product group basis. Althougb highR Dexpenditures

and the need to amortize them quickly probably have an im-

pact on the decision to internationalize, we do not believe

that tbis variable bas an impact on the entry modes chosen.

Therefore, we argue that the dimensions are better repre-

sented by measuring transaction costs directly and by mea-

suring the maturity of the technology. Because of its

widespread use in empirical studies, however, we include

R D

intensity as a control variable to detect any firm-level ef-

fects it might have on the choice of entry mode. We do not

present any hypotheses regarding the target country, but in-

clude it as a control variable as well. We follow the ap-

proaches of Erramilli and Rao (1993), Shrader, Oviatt, and

McDougall (1997), and Barkema and Vermeulen (1998) and

include the additional variables country risk, absolute size of

the target country, and gross domestic product (GDP) per

capita. We also include further dummy variables to control

for any industry-specific effects tbat may affect the chosen

sales mode.

To summarize, similar to otber researchers before us (e.g.,

Cavusgil, Zou, and N aidu 1993), we attempt to explain the

international entry modes using firm-specific variables,

product-specific variables, and variables specific to the

environment in which the firms operate. Firm size and

experience in international activities can be seen as opera-

tionalizations of the internationalization process model.

Experience with tbe domestic sales mode and tbe innovative-

ness of the technology can be seen as operationalizations of

the OC perspectives, whereas product characteristics are

proxies for the transaction cost perspective.

For the pvu:pose of this study, we define a high-tech start-up

as a legally independent company that is not older than ten

METHO

years and that operates in one or more high-technology sec-

tors.

An operationalizable definition of high-technology sec-

tors in tbe United Kingdom bas been established by Butchart

(1987),

who provides a definition of bigh-tech industries

based on the two ratios of

R D

expenses to sales and em-

ployees working in R D to total employees. Employing this

definition, we identified 33 high-technology industries as

having above average expenditu res for R D (for a list of in-

cluded indus tries, see Appendix A). Using a database ob-

tained from Dun and Brad street, we identified firms

operating in tbose industries that had at least three employ-

ees in 1997 and had been founded between 1987 and 1996.


12/31

We acknowledge that this method cannot consider high-tech

firms in industries that are not included by Butchart's defini-

tion. Yet as opposed to targeting low-technology sectors in

the search for high-technology start-ups, we expected the

proposed approach to result in an increased likelihood of ob-

taining responses from firms that fulfil the specified eligibil-

ity criteria. In total, we identified a gross sample of 7788 U.K.

firms. We subsequently screened all identified company

records to exclude those firms whose business activities sug-

gested they were not carrying out any

R&D

activities (e.g., re

tailers, wholesalers, assemblers). As a result, we retained

3590 firms as eligible for inclusion in the research sample.^

We then chose 2000 firms using a random sampling process

stratified by size class and service/manufacturing.

On the basis of a review of the specialized literature, we de-

veloped a foiu:-page questionnaire. We carried out four pilot

case studies to test whether the questions in the survey in-

strument appeared relevant, easy to understand, and unam-

biguous to the target respondents. As a result, we modified

the questionnaire to take into account the expressed con-

cerns. We posted an introductory letter and questionnaire fol-

lowed by three reminders, if appropriate, to the managing

directors of these 2000 firms. Managing directors have been

used in past studies to collect data on the overall performance

of entrepreneurial firms and have been identified as reliable

sources of information (Brush and Vanderwerf 1992). A total

of 134 envelopes came back unopened from companies that

could no longer be located at their addresses in the database.

Nine companies wrote back saying they were in the process of

receivership. Sixty-one firms contacted us, indicating that

they did not wish to participate in the survey. Altogether, 466

firms returned the completed questionnaires, which resulted

in a response rate of24 .After consistency checks to confirm

that each firm fulfilled all the criteria for eligibility (i.e., firms

needed to be less than ten years old and independently

owned firms), 362 (19 ) firms could be retained in the data

set. We performed t-tests using the original Dun and Brad-

street data to check for nonresponse bias. We did not detect

any statistically significant differences using information on

sales, sales change (in relation to the previous financial year),

number of employees, and credit rating. The latter measure

reflects the credit worthiness of the firm as attributed by

credit rating analysts and is therefore of particular interest

when judging the firm's overall performance status. The ab-

sence of significant differences suggests that the firms in oin:

sample do not differ from the nonrespondents.

In specifying the entry modes, we followed the definitions of

perationahzation Root (1994), Klein, Frazier, and Roth (1990), and Aulakh and

of Variables Kotabe (1997). Accordingly, we asked respondents to specify

whether the chosen mode of their international activities was


13/31

direct exporting, the use of an agent or sales representative

selling on a commission basis, the use of a distributor, a sales

joint venture, a wbolly owned sales subsidiary, or licensing.

We asked firms to make similar statements regarding their do-

mestic sales mode. We measured size using two variables:

sales during tbe last financial year and number of employees.

Because of the b igh correla tion of these two variables (r = .74,

p

.0001), we constructed a single measiu-e out of tbe stan-

dardized scores. We decided against using the share of nondo -

mestic revenues as a measure of experience (Aulakh and

Kotabe 1997). In principle, we accept tbe reasoning behind

tbat cboice; that is, firms that score high on that measure have

an understanding of foreign operations. However, we argue

tbat in th e case of a sample com posed of start-ups, it is impos-

sible to distinguish between cause and effect; that is, tbe sbare

of nondomestic revenues co uld at the same time be a function

of the entry modes used. Therefore, we m easure international

experience as tbe number of years tbe firm bas already been

engaged in international operations before entering a particu-

lar market. To operationalize the international experience of

the firms' senior managers, we include dummy variables for

whether the foimders had lived abroad or had worked for in-

ternationally operating companies before starting their present

business (Bloodgood, Sapienza, and Almeida 1996). The inno-

vativeness of the technology employed was measured using a

four-item scale (see Ap pen dix B). Resp onde nts were asked

whether their products are best classified as incorporating

tested combinations of existing technology, new combinations

of existing technology, novel technolog y dev eloped externally,

or novel technology developed specifically for tbis product by

the company. We measured tbe extent to which a product re-

quires client-specific customization and the transaction costs

incurred during the sales process using five-point Likert scales

(see Appendix B). Tbe foiu- items measvu-ing transaction costs

were tben combined into a single scale (alpha .73). Dummy

variables were included to specify wbetber or not the firm be-

longed to a particular industry. We divided the firms into tbe

five indu stry categories: software, inform ation technology (IT)

and conununications hardware, engineering, biomedical tech-

nology, and electronics (see Appendix A). We obtained coun-

try risk ratings from tbe publication Institutional Investor (see

also Shrader, Oviatt, and McDougall 1997). Because of the

bigh correlation between co untry risk data and GDP per capita

data (r =

.93,

p > .0001), we only included risk in our analysis.

Appendix shows a correlation table of our variables.

Of the 362 firms that partic ipated in the survey, we subse- ~

quently only retaine d firms wit h interna tional activities. Al- RESULTS

together, 246 (68%) firms were engaged in international

sales. Table 1 describes these respondent firms. On average,

they were six years old, started with 5 employees, and had

grown to 22 employees by the time of tbe survey. Eight per-


14/31

Table 1.

Descriptive Statistics

Variable

Age

Sales, first year*

Sales, last year*

Employees,

first year

Employees,

last year

R&D intens ity

( of sales )

R&D intens ity

( of employees)

ean

5.8

274

1215

4.4

20.0

16.7

30.7

Standard

eviation

2. 6

508

1722

6. 5

24.2

22.0

23.0

inimum

0

0

50

1

1

0

0

aximum

10

6900

16100

50

180

150

1.00

edian

6

124

6 50

3

12

8

25

Table 2.

International Activities

of Sample Firms

Sales in thousands of English pounds.

cent of tbe firms indicated tbat tbey did not carry out any

R&D activities. The indicators for botb R&D sbare and per-

centage of employees working on product development sug-

gest that tbe rema ining 92 of sampled firms operate in

technology-intensive and/or knowledge-intensive areas and

spen d, on average, the equiva lent of 15 of their ann ual

sales revenue on R&D.

Table gives an overview of the interna tional activities of the

start-ups in our total sample. On average, these firms gener-

ated38 of their total turnover from international sales (me-

dian 30 ) through operations in ten countries (median 6). A

third of these firms (34 ) generated more than 50 of their

revenues from nondomestic sales. A total of28 of all born

international firms generated international revenues witbin

tbeir first year of formation. Tbis latter group now generates,

on average, 46 of its revenues from foreign sales. Rougbly

half the sample firms (46 ) had ente red their first foreign

market by tbe end of tbeir second year. However, foreign di-

rect investment in internationally dispersed assets played a

minor role among the firms in our sample. Only 11 compa-

nies had established either a joint venture or a wholly owned

subsidiary to manufacture their products abroad.

Variable

Intemational sales

{ of tota l firm s)

Share of nondomestic

revenue ( )

Number of

countries entered

Years before first

intemational sales

ean

67.9

38.4

10.0

2.2

Standard

eviation

46.7

31.7

11.9

2.1

inimum

1

1

0

aximum

100

90

10

edian

30

6

2


15/31

Because our chosen u nit of analysis is the market entry deci-

sion, we asked respondents to provide us with information

on their three most important foreign markets, defined hy

their contrihution to the total sales of the firm's hest-selling

produ ct. This resulted in a data set of 547 market entry deci-

sions (see Table 3). Ten cases were excluded from the analy-

sis,because respondents either could not provide country

level sales or indicated that they operated using entry modes

that were difficult to classify.*

Of the 547 market entry decisions in our sample, the pre-

ferred current entry mode used hy the firms was distributors

(42 ) followed by direct exporting (36 ) and use of sales

agents

(11 ).

Market entry m odes that required some form of

direct investment were not extensively used by the firms in

our sam ple. A total of 7entries (5 ) were carried out using

the joint ventu re form, and 15 entries (3 ) were throug h

wholly owned subsidiaries. The generation of international

sales revenues through licensing equally had a marginal role

(9, 2 ). In comparison with the first entry mod es used by

these firms, it appears that aggregate changes of entry modes

over time reflect a tendency to use arrangements that repre-

sent a higher commitment to international sales. Of the 121

observed changes, 95 (79 ) represented a move toward in-

creasing commitment to foreign sales. These descriptive

findings provide further evidence for the validity of our

choice to compare direct exporting and exporting through

distrihutors as the main strategic options for the majority of

high-tech start-ups.

Table 4 gives an overview of the geographic spread of the in-

ternational firms in our sample. The most important markets

for British high-tech start-ups are found in Western European

countries, followed by North American and East Asian coun-

tries. When individual countries are examined, however, the

most frequently entered export market is the United States,

with 96 entries, followed by France (68) and Germany (64).

Entry Modes

Geographic Spread of Entries

Entry Mode

Exporting

Agents

Distributors

Sales joint venture

WboUy owned sales subsidiary

Licensing

Missing/otber

Total

First Entry

24 1

68

198

12

7

1 1

10

547

44

12

36

2

1

2

2

100

urrent Entry

19 9

60

227

27

15

9

10

547

36

11

4 2

5

3

2

2

100

Notes: The table shows the first and current entry mod es used iu the most imp ortant foreign

markets for the company's best-selling product.

Table 3.

First and Current Entry Modes


16/31

Table 4.

Geographic Focus of

International Activities

Unit of Analysis

Entry M ode

European Union/European

Free Trade Association

United States and Canada

East Asia (Japan, Hong Kong,

Singapore, Korea, Taiwan)

Australia and New Zealand

Emerging markets, Europe

South Am erica

Middle East

Emerging markets, Asia

Other

Total

Entries

Current

Entries

307

106

37

25

12

3

22

13

2 1

54 7

56

19

7

5

1

4

2

4

100

Firms

First Country

Entered

138

56

14

6

3

11

4

10

244

57

22

6

1

1

5

2

4

100

For the first market enteredwhich may not be the firms'

largest marketa similar pictiu-e emerges. Although the ma-

jority of firms (138, 57 ) had their first international sales in

Western Europe, the most popular country of first entry was

the United States (52, 21 ). It is noteworthy that43 of first

entries were made to countries that do not belong to the Eu-

ropean Union/European Free Trade Association. When the

second, third, fourth, and fifth market entries are examined, a

similar pattern is evident (not shown here). These results

lend broad support to the findings of Lindqvist (1991) and

Bell (1995). Although a narrow majority of firms chose geo-

graphically close countries for their first international sales,

an important minority of firms entered spatially distant mar-

kets first.

The descriptive results suggest that there is some support

for the theoretical claims of process models when the aggre-

gate sample is considered; that is, firms' entry modes

changed over time to reflect an increase of commitment to

foreign markets, and most firms first sell into relatively close

markets. However, several indiv idual firm s deviate from this

pattern. That the single most important target country, both

for first entry and for absolute numbers of market entries, is

the United States suggests that entry mode choices are also

driven by more compelling strategic reasons than psychic

distance alone. The economic size of the target market or the

recognition of significant, country-specific opportunities

may be more persuasive factors. We next present the results

of the regressions and investigate to what extent the chosen

firm-specific, product-specific, and country-specific factors

influenced the foreign entry decision.


17/31

To test the hypotheses stated previously, we estimated three

probit models with the entry m ode 1= distributor, 0 = direct

Regression nalysis

export) as the dependent variable. These two entry modes

capture 78 of the observed entry mod es in our sample.

Still, we excluded 28 cases because of missing information

on one or more independent variables. We report the mar-

ginal parameter coefficients of the model to compare the

magnitude of the effects of the different variables. We esti-

mated our principal model including all cases available for

analysis (Model

1).

Webase our discussion of the hypotheses

primarily on the results of this first model.

The overall solution is statistically significant at p >

.0001.

Among the 398 entry mode choices for which all variables

were complete, 213 (55 ) firms chose distributors and 174

(45 ) firms chose to export directly. The classification ratio of

more than 70 therefore suggests that the predictive ability of

the estimated model represents a substantial improvement

over the 55 maximum chance criterion (equivalent to as-

suming that all firms are distributors). In all three models, the

measure of size is positively related to the use of distributors,

and the effects are highly significant in a statistical sense.

However, when the marginal effects are examined, the real im-

pact of size on entry mode is very small. Nonetheless, Hj can

be accepted. The international experience of the firm at the

time of the market entry is not significantly related to the

choice of a particular sales mode in any of the three m odels.

H2

therefore cannot

be

accepted.

H3

also must be rejected. Liv-

ing experience abroad and previous work for an internation-

ally operating company are both negatively related to the use

of intermediaries. However, only living experience abroad is

statistically significant.H4was supported. Firms that used dis-

tributors domestically also had a higher propensity to use in-

termediaries for their international sales. A comparison of the

marginal effects also shows that this variable has the strongest

effect of all the variables included in the model.

H5,

following

the organizational capability perspective, states that the prod-

ucts of irmshat export directly incorporated newer technolo-

gies than the products of firms that entered foreign markets

through intermediaries. This hypothesis must be rejected.

Compared with the base case ofaproduct incorporating tested

combinations of existing technology, more advanced technol-

ogy tends to be sold through intermediaries. However, the co-

efficients indicate that this is a curvilinear re lationsh ip, as the

probability to sell through an intermediary is highest for prod-

ucts that had the second highest score of technological nov-

elty.

HQ

is supported by the model. Products that require

extensive customization are more likelyto besold internation-

ally without the use of intermediaries.H7cannot be accepted.

The measures used to operationalize the costs of commercial-

ization turned out to have a statistically insignificant impact

on the decision to involve intermediaries in the sales process.


18/31

Among the control variables, only one industry effect turned

out to be significant. All other things being equal, firms in

biotechnology and medical tecbnology rely much more on

intermediaries to sell their products abroad compared with

the base case of the electronics industry. Effects for software,

engineering, and IT hardware were not significantly different

from the base case. Research and development intensity was

used as a control variable and resulted in a significant, albeit

weak, effect. Accordingly, firms that have higher R&D expen-

ditures have a lower propensity to sell through distributors.

However, as the size of the marginal effect in Table 5 indi-

cates, firm differences of at least one order of magnitude must

be present to bave a substantial impact. Finally, among tbe

country variables, only the absolute market size appears to

bave a significant impact on tbe entry mode decision. Tbe

size of a national market tbus seems to be positively related

to tbe propensity to use distributors. The estimates for coun-

try risk did not have a statistically significant effect within

the model.

To examine the validity of our findings, we tested two addi-

tional models. Over time, some sample firms had changed

the modes of market entry employed. Thus, the current entry

mode could be a result of dynamic learning effects ratber

than of actual differences in firm or product characteristics.

Model 2 therefore only includes those entry modes that have

not been subject to cbanges over time. We make a further

modification in Model 3, in which only those entry modes

that represent at least 10 of the total turnover of the firm

were included. Our results could be biased by the presence of

reactive, unsolicited foreign sales that are not a result of firm-

specific factors or managerial action. Choosing a relatively

high threshold of 10 should lead to the exclusion of the

majority of these cases. The estimation of both Model 2 and

Model 3 confirmed tbe results of our main model. Models 2

and 3 equally represent statistically significant solutions,

and their respective classification ratios of 70.60 and 71.04

compare favorably with chance criteria of50 (Model 2) and

5 (Model 3). Furthermore, as Table 5 shows, the effects of

our explanatory variables in the three models are strikingly

consistent. The only differences were marginal changes

of statistical significance for management's foreign living

experience in Model 3 and the insignificance of one of the

technology dummy variables in Models 2 and 3. The latter,

however, does not contradict the results of Model 1, because

newer tecbnologies remain more likely to be sold through in-

termediaries compared with the base case.

From tbe previous analysis, a couple of results merit furtber

IS USSION consideration. First, we were surprised by the very incidence

of international activities among the firms in our sample. A

majority of start-ups (68 ) engaged in international sales. On


19/31

average, these technology-based young firms sold 38 of

their tvirnover in foreign countries, and33 of the firm s gen-

erated more turnover from international than domestic sales.

Respondents typically sold into ten countries at the time of

the survey and had initiated international activities two

Variable

Firm size

Managementliving experience abroad

Managementprevious work experience

in internationally operating firm

Firm experience in intemational sales

before market entry (in years)

Domestic sales mode

Products based on new combinations

of tested technology

Products based on novel tecbnology

developed outside

Products based on novel technology

developed in-house

Degree of custom ization

Cost of sales

Industry: software

Industry: IT hardware

Industry: engineering

Industry: biotechnology/medical

technology

R&D ex penditu res

Target country: absolute market size

Target country: country risk

Number of observations

Log-likelihood

t (d.f.)

Prob

> x

Pseudo R2

Classification ratio

*10

significance level.

**5 significance level.

1 significance level.

Model 1

.095

(.008)***

-.106

(.081)*

-.020

.734

- .004

.780

.368

(.000)***

.162

(.069)*

.319

(.001)***

.204

(.021)**

-.068

(.001)**

.001

.976

- .084

.424

.070

.435

.122

.179

.429

(.000)***

-.003

(.030)**

-.005

(.006)***

.001

.643

398

-219.26

109.98

.0000

.2 0

70.60

Model 2

.119

(.002)***

-.129

(.058)*

-.017

.801

.006

.680

.457

(.000)***

.128

.188

.346

(.002)***

.179

(.062)*

-.053

(.017)**

-.009

(.800

- .061

.591

.060

.553

.141

.173

.460

(.000)***

-.002

.142

- .033

(.078)*

.001

.538

339

-185.52

98.90

.0000

.2 1

73.16

Model 3

.0 8

(.037)**

-.105

.140

- .091

.199

- .224

.231

.343

(.000)***

.118

.272

.258

(.043)**

.188

(.073)*

-.057

(.016)**

.041

.307

.010

.929

.112

.284

.157

.145

.469

(.000)***

-.003

(.080)*

-.005

(.010)***

.001

.730

297

-167.71

76.22

.0000

.1 9

71.04

Notes: base case = an electronics com pany selling prod ucts that inco rporate tested technology.

Table 5.

Estimation Results of the

Probit Models


20/31

years after formation. Second, given the importance of inter-

national activities, we looked at the target countries and en-

try modes used. The majority of firms chose countries in

Western Europe for their first market entry. However, the

most freque ntly targeted country is the United States (19 of

all entries). A sizeable prop ortion of firms (42 ) entered

more distant non-European m arkets first. Strategic considera-

tions and the exploitation of new opportunities rather than

psychic/economic distance are more likely to account for

these entry decisions.

According to the sample average, these young firms elected

to choose entry modes that were not resource intensive. But

among these low-resource entry m odes, the use of intermedi-

aries was currently more prevalent than direct exporting.

Arguably, selling through distributors represents a more com-

plex and advanced managerial arrangement because of the

requirement to attract, train, incentivize, and monitor

a third-party agent. We therefore expected, all other things

being equal, the use of foreign intermediaries to be more

prevalent among more experienced and larger firms. As hy-

pothesized, we found a positive effect of firm size on the

propensity to sell through intermediaries. However, an

analysis of the marginal effect reveals thatsimilar to the sta-

tistically significant but weak effect of

R D

intensity the

real impact of that variable is quite small. Firm differences of

at least one order of magnitude m ust be present to have a sub-

stantial impact on the choice of entry mode. The influence of

experience on mode choice turned out to be contrary to our

hypotheses. Direct firm experience did not have a significant

impact on the commitment of the chosen entry mode. How-

ever, as intensive experience with foreign operations cannot

realistically be expected among young firms initiating their

first foreign sales, we also tested for management's interna-

tional experience, a substitute for direct company experi-

ence.

This variable had a significant impact on the choice of

entry mode, albeit not in the hypothesized direction. This re-

sult suggests that managers who have lived abroad are more

likely to sell internationally without the assistance of inter-

mediaries. A possible interpretation of this finding is that in-

ternationally experienced managers do not need to rely on

the superior market knowledge or commercial network of a

local distributor to commercialize their products abroad.

Their teniue abroad means that they already have their own

personal networks and can evaluate them in comparison

with the services offered by a distributor. We conclude from

these findings that experiential knowledge, the key variable

in the internationalization process theory (Andersen 1997),

is of limited value to explain the entry mode choices of the

high-tech start-ups in our sample. This echoes the findings of

other researchers (see, e.g.. Bell 1995). In addition, our find-

ings could provide further evidence for those researchers that


21/31

argued that different entry modes do not so much represent

distinct levels of commitment determined by past experience

but rather represent distinct managerial choices determined

by product- and firm-specific considerations (Andersen

1993;Leonidou and Katsikeas 1996).

In the case of start-ups, the choice of foreign entry modes

may therefore represent a compromise between the limited

resources of the start-up and the support requirements de-

manded by the customers of its products. As hypothesized,

our results indicate that a high degree of required customiza-

tion leads to the exclusion of intermediaries during the sales

process. Products that require a high level of client-specific

adaptation are more likely to be sold directly by the manu-

facturer. We argue that this is the case because the expertise

and tacit knowledge required to configure a product accord-

ing to customers detailed specifications are more likely to re-

side with the m anufacturer than w ith the intermediary. We

further hypothesized that start-ups whose products require

extensive pre- and after-sales support will be more likely to

sell through distributors. However, this hypothesis could not

be supported, because the level of required support did not

affect the choice of entry mode significantly.

We suggest the following interpretation of this result. The ef-

fects of the customization (Hg) and after-sales support (H7)

variables are somewhat related, which is also manifest in the

significant positive correlation between the two.^ From the

distrihutor s point of view, they both incorporate the notion

of having to acquire or invest in certain skills to guarantee

the effective comme rcialization of a client man ufacturer s

product. Such a commitment is attractive to the distributor

only if a large volume of

s les

is a reasonably guaranteed con-

sequence of this investment. All other things being equal,

customization and after-sales support should act as a barrier

to the interest or involvement of intermediaries in the sales

process. However, from the point of view of

firm that wants

to serve large or remote foreign m arkets, the use of a local dis-

tributor may be the only practicable way of providing the

necessary customer-focused infrastructure for installation,

maintenance, upgrading, and/or training of end users. We

tried to decompose this dilemma and argue that highly cus-

tomized products are more likely to be exported by m anufac-

turers directly because of their singular familiarity w ith their

core technology. The costs of acquiring these specialized

technological skills may be prohibitive and economically ir-

rational for a d istributor. However, in certain cases in w hich

the d istrihutor already has a portfolio of related products and

technologies in place, the relative costs of both learning and

subsequently operationalizing the standard and routinized

tasks of installation, end-user training, and maintenance are

likely to be much lower because of both scale and scope ef-


22/31

fects. Therefore, in these circumstances, selling a volume

product or product family whose commercialization requires

high levels of support through intermediaries might be less

problematic. That the two variables have different effects in

the regression (significantly negative in the case of cus-

tomization versus insignificantly positive for sales support)

is therefore an interesting result in itself and goes a certain

way to support our reasoning. It suggests that customization

represents a barrier to involving intermediaries, whereas the

attendant cost of sales support can be managed and may even

be a source of attractive profit for the distributor given suffi-

cient trading volume to ensiure scale effects.

The newness of the technology incorporated had a signifi-

cant impact on the choice of the entry mode, albeit in the op-

posite direction to the one we hypothesized. Compared with

the base case of mature and tested technologies, transactions

involving products that incorporated more innovative tech-

nology (and therefore embodied a higher degree of tacit

knowledge according to the theorists of OC) had a higher

chance of being dealt with through collaborative arrange-

ments than being exported. Although this indicates that the

international market entry forms are influenced by product-

specific factors, the effects observed directly contradict the

theoretical prescriptions and findings on technology transfer

modes of larger firms (Davidson and McFetridge 1985). How-

ever, they corroborate the recent findings of Robertson and

Gatignon (1998), who report that firms that experienced

higher technology uncertainties were more likely to engage

in alliances.

Ovu strongest predictor of the chosen foreign entry mode was

the existing, domestic sales mode of the

firm.*^

The effect of

this variable can be explained in two ways. First, it is a proxy

for different strategic and structural influence factors that af-

fect sales channel choice irrespectively of the context. This

variable therefore partly accounts for unobservable effects

whose determination was not among the objectives of the

study. Second, the explanatory effect of this variable is ar-

guably due to the presence of embedded routines and experi-

ences with the domestic sales mode. This finding therefore

supports theoretical propositions that stress the importance

of firm-specific routines and the path dependence of organi-

zational outcomes (Madhok 1997). However, further research

should necessarily use a more refined measure to determine

to what extent foreign entry decisions axe a result of path de-

pendency or company marketing strategy.

The results of this research confirm what the high-tech entre-

M N GERI L IMPLIC TIONS preneurial manager already knows, namely, that the early de-

cision of an appropriate mode of exporting is not a trivial

activity. On the contrary, it is an activity of the most pro-


23/31

found strategic import and has long-term implications. In an

ideal world, it is likely that any firm would prefer to be in

charge ofitsdestiny and would chose a direct export mode of

operation or build up a wholly ow ned subsidiary. Yet the lat-

ter is usually outside the reach of a young start-up company.

Exporting, in contrast, is not always an app ropriate choice. If

there are serious questions regarding the firms ability to pro-

vide user support in distant markets, the firm may have little

choice but to choose a more established and recognized dis-

tributor. In making this choice, the firm is, in effect, subcon-

tracting a part of its growth strategy to an agent. For the

distributor, the young firm is only one of several clients. In

additio n, the firm s bargaining power may be small in com-

parison with bigger, higher-volume, and longer committed

clients of the distributor. As a result, a situation with high

conflict po tential emerges. othe distributor, providing a ser-

vice to this type of immature firm represents a series of nec-

essary but highly speculative, specialist investments or sunk

costs. The distributo r s choice of accepting the innovative

products of an unknown start-up will therefore depend on

the projected sales volume. As a result, young firms that can-

not meet their distributors expectations will be left with no

other choice than to sell directly.

Our research has identified several practical implications

for the managers of technology-based start-ups. We isolate

three findings that we believe professional managers might

find helpful. First, those young firms wishing to use distrib-

utors in foreign m arkets can benefit from using collaborative

relation ships in their domestic market first. Learning effects

in managing relationships with intermediaries may be more

easily gained in the domestic market and at less cost. In

certain circumstances, this learning may also be less risky

given, for example, easier communication and/or negligible

psychic distance. Even for born global firms, market-based

experimentation in the domestic market may still have

these advantages.

Second, firms that sell a highly customized product should

be prepared to commit appropriate resources to their presales

and after-sales service strategy given their reliance on direct

exports. In giving this advice, we are aware that our research

finding s show that firms appear to choose a mode of foreign

market entry irrespective of the resource implications of the

comm ercialization p rocess. Similarly, our findings also show

that the size or resource endow ment of individua l firms has a

weak, albeit statistically significant, effect on the choice of

mode of market entry. In short, it does not appear that re-

source constraints prevent firms from choosing between the

two discussed entry modes. Instead, managers should be

aware that the choice is more likely to be influenced by the

degree of customization of the product.


24/31

Third, it appears that in getting a new, technologically ad-

vanced product into the market, start-ups with a necessarily

limited record of achievement should seriously consider col-

laboration to exploit the track record or reputation of an es-

tahlished intermediary. The much-quoted concept of

liability of newness posits that young firms face disadvan-

tages, because stable relations with clients are not yet estab-

lished (Briiderl and Schussler 1990; Stinchcomhe 1965). In

cross-border business relationships, this may amplify into

what we call the liability of alienness. The reluctance of

customers to rely on small, untested, and foreign suppliers is

particularly important in the sensitive field of medical tech-

nology, as is evidenced in our results. However, this antipa-

thy to uncertainty by customers can be extended to all users

of mission-critical technologies provided by young and un-

proven third-party suppliers. In such circumstances, and

particularly when there are costly implications for product

failure or nonperformance, an exclusive dependence on a

young company provider is not an acceptahle proposition for

any established corporate user. The use of a trusted distribu-

tor, or more accurately a value-added reseller, and its existing

sales force is frequently the only effective way for a young

high-tech company to present products to mainstream

customers early in its development. Distributors may also

be essential for supplying widely dispersed customers in a

continental-scale market such as the United States or in mar-

kets where there are major cultural disparities between

provider and user, such as China.

Our study is not without limitations. First, we focus on only

L M T T ONS two distinct entry modes in our analysis, because licensing

and entry modes requiring direct capital investments were of

only marginal importance in our sample. Fvurther research on

those technology start-ups that engage in a wider spectrum of

entry modes could reveal whether the effects identified by

our research can account for differences between the involve-

ment of intermediaries and the completely internalized entry

decisions when firms set up foreign subsidiaries. Second, for

the moment, our study cannot reveal any performance impli-

cations of the choice of different entry modes. Despite the

availability of data on firm growth, we cannot presently dis-

tinguish between cause and effect for the majority of firms in

our sample; that is, either the growth could be a result of the

chosen entry mode, or vice versa. However, we plan to con-

tact the participants of our study again in the future to inves-

tigate long-term performance implications of different market

entry strategies. Finally, our sample is drawn exclusively

from high-tech start-ups in the United Kingdom. Further

research should therefore address whether similar findings

can be reported from other countries and other industry sec-

tors characterized hy the presence of internationally operat-

ing entrepreneurial start-ups.


25/31

Industry NACE Classifications

Software

ITandcommunications hardware

Engineering

Life sciencesandmedical technology

Other (mainly electronics, components)

7220 7260

3001 3002 3220

3230

3320 3330 3340

2441 2442 331

3110 3120 3210 3530 2416

2417

Notes:TheNACE classification systemis theEuropean Union equivalentofthe standard

industrial classification code system.

APPENDIX

A

INDUSTRY SECTORS

INCLUDEDIN THESTUDY

1

Innovativeness of Technology

How wouldyoubest describetheinnovativenessofyour

productorservice?

Itincorporates triedandtested combinationsof

existing technology.

It

incorporates

new

combinations

of

existing

technology.

Itincorporates novel technology that has been

developed elsewhere.

Itincorporates novel technology thathad to hedevel-

oped specificallyforthis product by

youi

company.

2

Customization

Please describetheextentto which your productor ser

vice requires

Individual client

customization

low

substantial

aaa

doesnot

apply

3 Transaction ost Intensity

Please describekeycharacteristicsofthe product/service,

particularlytheextenttowhichitrequires

low

substantial

Technical consultation

priortosales

Complexortime-

consuming installation

Regular maintenance

and/or upgrades

Specialized training

requiredforfront line

and sales personnel

doesnot

apply

APPENDIX

B

OPERATIONALIZATION

OF VARIABLES


26/31

APPENDIX 3

CORRELATION TABLE

5 S S.

B E. ^

II II II II II II II II

S 1 1

3 I

E a c c o

P ^ n o o

o :r r 3

03 -^ O O O

r i g s

f-i 2 CB CO O

^ P- CO cn y

e. G g

CD

S

L L B .

CD E . CD CD 0 0

m tn g 5 n

Q. O O O H

d f 5:: tJ

I

I i-

CD

2 S -a -a

entry choices hight-tech.pdf

Documents

Transcript of entry choices hight-tech.pdf