An Exploratory Study of Trade Show Formation and Diversity.pdf

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Romanian Ministry Consortium]On: 2 March 2010Access details: Access Details: [subscription number 918910197]Publisher RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Business To Business MarketingPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t792303971

An Exploratory Study of Trade Show Formation and DiversityJianan Wu a; Gary L. Lilien b; Aniruddha Dasgupta c

a E. J. Ourso College of Business, Louisiana State University, Baton Rouge, LA b Smeal College ofBusiness, The Pennsylvania State University, University Park, PA c Massachusetts Maritime Academy,Buzzards Bay, MA

To cite this Article Wu, Jianan, Lilien, Gary L. and Dasgupta, Aniruddha(2008) 'An Exploratory Study of Trade ShowFormation and Diversity', Journal of Business To Business Marketing, 15: 4, 397 — 424To link to this Article: DOI: 10.1080/15470620802325617URL: http://dx.doi.org/10.1080/15470620802325617

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

http://www.informaworld.com/smpp/title~content=t792303971

http://dx.doi.org/10.1080/15470620802325617

http://www.informaworld.com/terms-and-conditions-of-access.pdf

Journal of Business-to-Business Marketing, Vol. 15(4) 2008© 2008 by The Haworth Press. All rights reserved.

doi:10.1080/15470620802325617 397

WBBM1051-712X1547-0628Journal of Business-to-Business Marketing, Vol. 15, No. 4, October 2008: pp. 1–40Journal of Business-to-Business Marketing

An Exploratory Study of Trade Show Formation and Diversity

Wu, Lilien, and DasguptaJOURNAL OF BUSINESS-TO-BUSINESS MARKETING Jianan WuGary L. Lilien

Aniruddha Dasgupta

ABSTRACT. Purpose: To analyze the objectives of trade show partici-pants and to link those objectives to trade show formation and diversity.

Methodology/Approach: We conduct an exploratory analysis of tradeshow formation and diversity and link them to the differences in severalseller and buyer interests across industries. We used data collected from asingle large trade show to examine the nature of buyer and seller participa-tion goals using multivariate statistical analysis. We then used a data setcollected from the entire trade show industry to examine the trade showdiversity via an econometric model.

Findings: We find that higher selling and buying propensity is linked tomore vertical shows whereas higher breadth of product interests is linked

Jianan Wu is associate professor of marketing at the E. J. Ourso College of Busi-ness, Louisiana State University, Baton Rouge, LA (E-mail: [email protected]).

Gary L. Lilien is distinguished research professor of management science atSmeal College of Business, The Pennsylvania State University, University Park,PA (E-mail: [email protected]).

Aniruddha Dasgupta is associate professor of international maritime businessat the Massachusetts Maritime Academy, Buzzards Bay, MA (E-mail:[email protected]).

We gratefully acknowledge financial support from the Institute for the Studies ofBusiness Markets (ISBM) of Penn State University. We thank Hans Baumgartner,Kalyan Chatterjee, Srinath Gopalakrishna, and Arvind Rangaswamy for manyuseful comments on an earlier version of this paper and Exhibit Surveys, Inc. forproviding the data used in the empirical study.

Address correspondence to: Jianan Wu, E. J. Ourso College of Business,Louisiana State University, Baton Rouge, LA 70803 (E-mail: [email protected]).

Downloaded By: [Romanian Ministry Consortium] At: 19:51 2 March 2010

398 JOURNAL OF BUSINESS-TO-BUSINESS MARKETING

to more horizontal shows. We also find that a highly innovative industry isoften linked to more horizontal shows.

Originality/Value/Contribution: Although trade shows are critical ele-ments of the communications mix in business-to-business marketing, verylittle is known about how a trade show is formed and why two differenttypes of shows (i.e., horizontal and vertical) coexist in the industry. This isthe first study that investigates these issues. Our work should help guidetrade show organizers diagnose and improve the appropriate mix of showsin a given industry and help trade show participants select better shows toattend, depending on their show objectives and selling/buying interests.

KEYWORDS. Trade show, business marketing

Trade shows or industrial exhibitions constitute an important elementof the communication mix in the business marketplace. In terms ofexpenditure allocations, they make up the third largest component of themarketing communication mix behind direct marketing and businessmagazine advertising (Pinar, Rogers, and Baack 2002). In the UnitedStates alone, the trade show industry annually hosts over 12,000 shows,involves over two million exhibitors and over 100 million attendees(CEIR 2001), and generates over 50 billion U.S. dollars (Marken 2004).While the industry has seen some consolidation in recent years, growth ininterest and participation remains high (Donberg 2006; The IndustrialRobot 2006). The widespread prevalence of this form of market/meetingplace in the United States seems representative of the world at large. Forinstance, Hansen (1996) reports that there were 169 international showsheld in Germany with 124,000 exhibitors and over 10 million attendees,as well as 220 international shows organized in the UK with 61,000exhibitors and 4.6 million attendees.

In business markets, two types of trade shows are formed based onmarket coverage: vertical shows and horizontal shows (Gopalakrishnaand Lilien 1994). The nature of the show audience in terms of productinterest, buying/selling plans, the time horizon of strategic consideration,and so on, can differ greatly by type of show. Typically, a vertical showinvolves a narrow range of products and attracts visitors specifically inter-ested in those products. In contrast, a horizontal show usually involves amuch broader range of products and a more diverse audience. For example,attendees at the Association of Operating Room Nurses show, a verticalshow, are almost all operating room nurses, and exhibitors display products


Wu, Lilien, and Dasgupta 399

that are used almost exclusively in the operating room. A typical horizontalshow is the National Design Engineering Show, in which firms demonstrateproducts ranging from mechanical components, electrical and electroniccomponents, plastics, elastomers, to CAD/CAM systems (Gopalakrishnaand Williams 1992).

It seems curious that vertical and horizontal shows coexist in manyindustries and that the proportion of horizontal shows formed varieswidely across industries. Table 1 details the number and proportion ofvertical shows and horizontal shows for different industries in a seven-year span (1985–1991). There appears to be no simple pattern here: whyare 319 of the 322 shows in the communications industry vertical, 61 ofthe 64 shows in the food processing and distribution industry horizontalwhile the chemical industry splits almost evenly with 20 vertical and 23horizontal shows?

Even the status of existing shows evolve, with some merging andbecoming more horizontal and others splitting into more narrow shows.For instance, Conexpo-Con/Agg ’96, the largest U.S. trade show in theconstruction industry, resulted from a merger of Conexpo (owned by theConstruction Industry Manufacturers Association), the largest construc-tion show, and Con/Agg (owned by the National Aggregates Associationand the National Ready Mixed Concrete Association), one of the premiervertical shows specializing in concrete and aggregates in the UnitedStates. The new horizontal show attracted more than 1,250 exhibitors and100,000 attendees, and covered 1.25 million square feet of exhibitingarea, making it bigger than the last Conexpo and Con/Agg combined(Show Daily of Conexpo and Con/Agg, ’96, 18). On the other hand, Com-dex, the largest (horizontal) computer show in the United States, has seenthreats by key exhibitors to split and form more narrowly focused shows(Calton 1994, A1). The consolidation and separation of trade shows isoften strategic. For example, in 2005 the merge of Waste Expo andWastecon continued to make news in the waste industry. Yet the twoshows remain separate entities (Johnson 2005).

This diversity and structural evolution of show type has seen little orno academic attention to date. Most trade show studies emphasize sales-lead generation for exhibitors as the key goal, providing little rationale forthe formation of horizontal shows (Gopalakrishna and Williams 1992;Gopalakrishna and Lilien, 1994; Dekimpe et al. 1997). These empiricalstudies assume that lead generation is a dominant objective for exhibitorsand measure the performance of the show by how efficiently trade showsgenerate such leads. That performance measure suggests that because


400

TA

BLE

1. T

he d

istr

ibut

ion

of tr

ade

show

type

s fo

r to

p 21

indu

strie

s (1

985–

1991

)

Fre

quen

cyIn

dust

ryN

umbe

r of

V

ertic

al S

how

sN

umbe

r of

H

oriz

onta

l Sho

ws

Tot

al N

umbe

r of

Sho

ws

Pro

port

ion

of

Hor

izon

tal S

how

s

Ove

r 10

0 sh

ows

Com

pute

rs a

nd c

ompu

ter

appl

icat

ions

379

7845

717

%C

omm

unic

atio

ns31

93

322

1%

50–1

00 s

how

sE

ngin

eerin

g3

9396

97%

Med

ical

and

hea

lth c

are

904

944%

Hou

sing

800

800%

Foo

d pr

oces

sing

and

dis

trib

utio

n3

6164

95%

Ele

ctric

al a

nd e

lect

roni

cs52

759

12%

15–5

0 sh

ows

Pai

nt49

049

0%N

ursi

ng48

048

0%A

utom

otiv

e an

d tr

ucki

ng3

4447

94%

Che

mic

al20

2343

53%

Pla

stic

s0

3333

100%

Rad

io, T

V, a

nd c

able

321

333%

Ene

rgy

030

3010

0%B

uild

ing

and

cons

truc

tion

1510

2540

%R

esta

uran

ts a

nd fo

od s

ervi

ce0

2424

100%

Man

ufac

turin

g6

1521

71%

Pho

togr

aphi

c21

021

0%W

eldi

ng17

017

0%P

acka

ging

015

1510

0%E

duca

tion

150

150%

Tot

al1,

152

441

1,59

328

%

Sou

rce:

Exh

ibit

Sur

veys

Inc

. T

he in

dust

ry c

lass

ifica

tion

used

her

e ar

e in

kee

ping

with

the

Tra

de S

how

Bur

eau’

s ow

n cl

assi

ficat

ion

sche

me.

Indu

strie

s w

ith fe

wer

than

15

trad

e sh

ows

have

bee

n om

itted

from

this

tabl

e be

caus

e of

the

inst

abili

ty o

f the

pro

port

ions

whe

n th

e to

tal n

umbe

rof

sho

ws

in th

e in

dust

ry is

that

low

.



vertical shows outperform horizontal shows in terms of selling efficiency,there should be little economic rationale for the formation of horizontalshows. Research to date has not studied why the proportion of show typesvary across industries, which is the key question we explore here. To seeka coherent answer to this question requires an investigation of the actorsand factors involved in the trade show formation process.

Our exploratory analysis suggests that trade show participants’ selling/buying propensity and breadth of product interests are closely correlatedwith the type of trade show formed: the higher the selling/buying propen-sity the more vertical shows we observe whereas the higher the breadth ofproduct interests the more horizontal shows. Our analysis provides amechanism to identify those product or market areas that appear eitherunderserved or overserved by one or another type of show.

We proceed as follows. In the next section we identify some importantvariables that theory and logic suggest should be linked to the diversity oftrade shows across industries. Next we use data collected from a leadingtrade show research firm to empirically investigate our propositions. Weconclude by discussing the theoretical and managerial implications of ourresults and the potential for future research in this area.

CORRELATES OF TRADE SHOW DIVERSITY

In this section we identify the variables that are most likely to correlatewith the formation and type of trade shows. We focus on two sets of cor-relates: (1) trade show–specific correlates on Selling/Buying Propensityand Breadth of Product Interests, and (2) an important industry-specificcorrelate—Technology Innovativeness.

Trade Show–Specific Correlates

Most of the research on trade show focuses on understanding the char-acteristics within a show or trade show–specific determinants. Someresearch (Gopalakrishna and Williams 1992; Gopalakrishna and Lilien1994, 1995; Dekimpe et al. 1997) has focused on what might be calledtransactional goals (sales-lead generation, supplier selection, and the like)while other research (Borghini et al. 2006; Blythe 2000; Sharland and Balogh1996; and Shoham 1999) has also looked at nontransactional goals (e.g.,intelligence gathering and general awareness generation). If we call the ratioof the importance of transactional to nontransactional goals buying or selling



propensity (for attendees and exhibitors respectively), we should expectthese ratios to vary from industry to industry. And in some industries suchas energy that rely on numerous technologies and see broad applications,participants may be interested in a broad range of products while in others(e.g., paint) those interests are likely to be much narrower. Variation inboth these sets of participant goals may be expected to affect the type oftrade shows that are formed.

Selling or Buying Propensity

Exhibitors may have different selling propensities when exhibiting at atrade show. Kerin and Cron (1987) examined exhibitors’ performanceusing factor analysis and found two categories of roles that trade showsplay in the exhibition industry: some exhibitors have objectives that areprimarily sales oriented, such as generating leads, while others have non-sales-oriented objectives such as gathering competitive intelligence andenhancing company image. Following Kerin and Cron’s approach,Hansen (1999) proposed a two-dimensional framework on exhibitors’ tradeshow performance: an outcome-based dimension that includes the sales-related activities, and a behavior-based dimension that includes nonsales-related activities, such as information-gathering activities, image-buildingactivities, motivation activities, and relationship-building activities. Hansen(1999) also proposed a dual-motive theory in which selling and nonsellingfunctions may exist not only for exhibitors but also for attendees. Usingdata collected from the seafood exporting industry of Norway, Hansen(1999) empirically demonstrated general support for his dual-motivetheory.

Consistent with Hansen’s (1999) findings, attendees have differentbuying propensities as well. The results from a Trade Show Bureau(1994) survey demonstrates that some attendees have immediate purchaseneeds while others may use trade shows to gather new product informa-tion that may be useful in guiding future purchases. Godar and O’Connor(2001) find further support for this perspective, segmenting trade showattendees’ motives into two groups: short term (e.g., to make a buyingdecision or to confirm a prior decision) and long term (e.g., to reinforcecontacts, develop contacts, and gather information for future purchasedecisions).

Exhibitors and attendees prioritize the importance of their informationneeds when attending a show, and should look for the best return on thatinformation-exchange investment. Exhibitors with low selling propensity



should seek to spread the word broadly; attendees with low buying pro-pensity should seek the broadest base of information collection possible.Hence, situations with low selling/buying intensity by exhibitors or sellersshould favor horizontal shows over vertical shows as efficient informa-tion exchange venues. Therefore, we have:

Proposition 1. A lower selling or buying propensity for exhibitors orattendees will be positively correlated with the number of horizontalshows.

Breadth of Product Interests

There are several reasons why exhibitors and attendees may have inter-ests in different product categories, which we define as breadth of productinterests. For example, small firms produce fewer products (as exhibitors)and are likely to need fewer products (as attendees). And some industriesare linked to fewer other industries or technologies than others, leading toa lower breadth of product interest. The same reasoning holds for verti-cally organized firms, whose show participants are likely to represent anarrow set of interests. Industries characterized by vertical integrationwill see lower breadth of product interests while highly integrated indus-tries should see a higher breadth of product interest (Williams et al. 1993).

Although both exhibitors and attendees may demonstrate signifi-cant breadth of product interests, there may be differences between thetwo parties because the costs incurred by exhibitors (in dealing withmultiple products) are much higher than those for attendees. In addi-tion to the costs that attendees incur, exhibitors also incur space rental,booth setup, and entertainment costs. The more offerings exhibited,the more (expensive) show floor footage required, and exhibitors’setup costs increase significantly as the number of products exhibitedincreases. Attendees see no such cost increases as their product inter-ests broaden.

If both exhibitors and attendees have a large breadth of product inter-ests, a horizontal show should suit their needs better than a vertical show.This is analogous to a consumer going to a shopping mall or a departmentstore versus a specialty shop. Therefore, we have:

Proposition 2. A greater breadth of product interests for exhibitors orattendees will be positively correlated with the number of horizontalshows.



Although it seems intuitive that the higher breadth of product interestsby either exhibitor or attendee should make horizontal shows moreappealing, the size and possible congestion at a large, heavily attendedand diverse show can make such a show inefficient. Conexpo and Con/Agg ’96, a horizontal show formed by merging two vertical shows, illus-trates this issue. In the merger decision a major issue for the Con/Aggteam was their concern that exhibitors of specialized concrete and aggre-gates equipment might get lost in a large general construction forum.Robert A. Gale, representing the Con/Agg team, said, “The decision onproduct concentration area was a key. That helped us [on the Con/Aggside] make the decision about 1999. We realized product concentrationareas were essential . . . not just for concrete and aggregates suppliers butfor many other product categories too” (Show Daily of Conexpo and Con/Agg, ’96, 18).

This inefficiency of horizontal shows can be critical when the dispar-ity of breadth of product interests between exhibitors and attendees ishigh. To see the rationale, assume there are two products X and Y.Assume that while all attendees are interested in both products (i.e., ahigh breadth of product interest for attendees), some attendees (denotedby AX) are more interested in X while others (denoted by AY) are moreinterested in product Y. Assume also that the exhibitors have a narrowbreadth of product interest, leading to a high degree of disparity ofbreadth of product interests between attendees and exhibitors. Supposealso that some exhibitors (denoted by EX) are only interested in exhibit-ing X while others (denoted by EY) are only interested in exhibiting Y.Other things being equal, it will be more efficient if EX (or EY) canidentify AX (or AY) with less effort (e.g., saving valuable time at atrade show). In a horizontal show with both products X and Y, all EX,EY, AX, and AY will attend. In two vertical shows with products X andY separated, EX and AX will attend one show while EY and AY willattend the other. Exhibitors (EX or EY) at the horizontal show mustincur extra effort in screening out their marginal attendees (AY or AX),effort avoided in vertical shows through self-selection. This inefficiencyleads to the separation of a horizontal show into one or more verticalshows where the scanning costs for exhibitors and attendees will besmaller. Therefore, we have:

Proposition 3. A greater difference in breadth of product interestsbetween exhibitors and attendees will be negatively correlated with thenumber of horizontal shows.



Industry-Specific Correlates

We discussed trade show specific correlates in trade show formation.However, there are industry-specific correlates in trade show formation aswell, a critical one being technological innovativeness.

Technological Innovativeness

One of the fundamental industry characteristics is technologicalinnovativeness. The degree of technological innovativeness affectsneeds and wants of exhibitors and attendees. When the technologylifecycle is short, it is critical for all actors to stay abreast of the latestdevelopments not only within the core industry but also in adjacentindustries into which they can expand or from where future competi-tors may emerge (Zook and Allen 2001). A cross section of industries(see Table 1) will allow us to examine the effect of technologicalinnovativeness. Specifically, for industries with higher levels of tech-nological innovativeness (which we refer to as technological turbu-lence), more frequent and broader environmental scanning isnecessary to stay abreast of current and future product developments.Therefore, we have:

Proposition 4. Higher levels of industry technological innovativenesswill be positively correlated with the number of horizontal shows.

These ideas and our propositions are summarized graphically in Figure 1.(We discuss the possible effect of other industry-specific correlates fur-ther on in the discussion section.)

EMPIRICAL ANALYSIS

Data

We use data obtained from Exhibit Surveys, Inc., a leading researchfirm in the exhibition industry, and supplement it with two other datacollection efforts described below. Our Propositions require data at theshow-participant level (within show) and data at the show level acrossindustries (cross show). We first conduct an empirical analysis on theobjectives and product interests of trade show participants at a givenshow to establish the validity of the assumptions underlying our conceptual



framework. This analysis, while preliminary in nature, allows us tosee if the conceptual constructs on selling/buying propensity andbreadth of product interests are empirically supported and may also beused to support further theory development. We next test our Propositionson the drivers of trade show diversity derived from our conceptualframework.

Trade Show Participant Data

These data come from a single large show in the building and con-struction industry held in Las Vegas in 1996 with more than 100,000attendees and more than 1,000 exhibitors. The data come from twosources: (1) questionnaire data collected via two mail surveys (1–7Likert scale), one from a sample of exhibitors and the other from asample of attendees; and (2) scanner-type data obtained from attendeeswhen they have their electronic swipe card scanned at the booth theyvisited.

FIGURE 1. A conceptual framework showing the drivers of trade showformation and diversity.

Note: This figure suggests that two sets of parties, exhibitors and attendees, can vary intheir trade show participation objectives along two sets of objective dimensions (i.e., breadthof product interest and buying and selling intensity) when influencing the formation of thetype of trade show. The impact of that variation depends on industry structure (i.e., technologyinnovativeness).

Technological Innovativeness

Technological InnovativenessT

echn

olog

ical

Inn

ovat

iven

ess

Tec

hnol

ogic

al I

nnov

ativ

enes

s

ExhibitorSelling Propensity

AttendeeBreadth of Product Interest

Exhibitor Breadth

ofProduct Interest

hgiH woLHigh

Low Low High

High

Low

AttendeeBuying Propensity

Trade Show Formation, Diversity



Cross Show Industry Data

These data also come from two sources. (1) the number of verticalshows and horizontal shows in the United States in a seven-year span(1985–1991). For purposes of data reliability, we retained only thoseindustries that had at least 15 shows (i.e., at least two shows per year),resulting in 1,539 shows in 21 industries, which represents 91% of thetrade shows in this time period (Figure 2). (2) the assessed breadth ofproduct interest and selling/buying propensity (for both exhibitors andattendees in an industry), and technological innovativeness for each of the21 industries in Figure 2. To generate these latter measures required thatwe obtain data from industry experts. We were able to identify and enlistthe cooperation of two experts with broad cross-industry experience toassess breadth of product interest and selling/buying propensity for bothexhibitors and attendees (using a 1–7 Likert scale). Using a Delphi-likeprocedure, the experts converged on the assessments found in Table 2after two rounds. We also identified 17 well-known experts in new prod-uct development and technological innovation to assess, in each of the 21industries, the degree of innovativeness, defined as “the relative amountof new product development and new technical information that occursper year within the industry” (1–7 Likert scale).

FIGURE 2. Distribution of breadth of product interests for exhibitors andattendees

(A) shows the distribution of number of product types exhibited by exhibitors within the16 product types and (B) shows the number of product types of interest to attendees, indi-cating significant cross-product interests for both exhibitors and attendees at this show, butwith the breadth of product interests for exhibitors smaller than those for attendees.

Exhibitors (A)

(Sample size = 351, Mean = 1.23)

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Breadth of Product Interest

Perc

enta

ge

Attendees (Sample size = 281, Mea n = 5.79 )

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

1 2 3 4 5 6 7 8 9 1 01 1 1 2 1 3 1 41 5 1 6 Breadth of Product Interest

Perc

enta

ge

(B)


408

TA

BLE

2. M

ean

expe

rt e

valu

atio

ns o

f var

ious

cha

ract

eris

tics

of b

oth

the

trad

e sh

ows

and

the

asso

ciat

ed in

dust

ries

note

d in

Tab

le 1

Indu

stry

Tec

hnol

ogic

al

Inno

vativ

enes

s(I

ndus

try)

(T

EC

H)

Buy

ing

Inte

rest

(Atte

ndee

)(A

TB

UY

)

Sel

ling

Inte

rest

(Exh

ibito

r)(A

TS

ELL

)

Bre

adth

of P

rodu

ctIn

tere

st (

Atte

ndee

)(A

TP

RO

D)

Bre

adth

of P

rodu

ctIn

tere

st (

Exh

ibito

r)(E

XP

RO

D)

Com

pute

rs a

nd c

ompu

ter

appl

icat

ions

6.58

6.00

6.00

3.00

4.00

Com

mun

icat

ions

6.32

5.00

6.00

2.00

3.00

Eng

inee

ring

5.00

2.00

3.00

6.00

5.50

Med

ical

and

hea

lth c

are

5.08

5.00

4.00

2.00

2.00

Hou

sing

2.66

5.00

4.50

2.50

3.50

Foo

d pr

oces

sing

and

dis

trib

utio

n3.

112.

504.

004.

504.

50E

lect

rical

and

ele

ctro

nics

5.58

3.00

5.00

2.00

2.00

Pai

nt2.

266.

003.

002.

003.

00N

ursi

ng2.

373.

005.

002.

004.

00A

utom

otiv

e an

d tr

ucki

ng3.

262.

003.

004.

004.

50C

hem

ical

3.71

3.50

4.00

3.50

4.00

Pla

stic

s3.

823.

004.

004.

504.

50R

adio

, TV

, and

Cab

le3.

536.

004.

002.

004.

00E

nerg

y2.

613.

003.

005.

005.

50B

uild

ing

cons

truc

tion

2.53

3.00

5.00

5.00

3.00

Res

taur

ants

and

food

ser

vice

2.45

2.00

2.00

4.50

5.00

Man

ufac

turin

g3.

793.

005.

004.

004.

00P

hoto

grap

hic

4.26

6.00

5.00

2.50

2.50

Wel

ding

2.32

5.00

6.00

3.00

3.00

Pac

kagi

ng3.

035.

004.

504.

504.

00E

duca

tion

2.76

6.00

4.00

4.00

2.00

Mea

n3.

674.

054.

293.

453.

69S

D1.

331.

511.

091.

251.

07

Not

e: C

olum

n 1

(TE

CH

) is

the

mea

n fr

om 1

7 ex

pert

eva

luat

ions

whi

le t

he o

ther

col

umns

rep

rese

nt t

he c

onse

nsus

judg

men

t af

ter

two

roun

dsfr

om o

ur tw

o tr

ade

show

res

earc

h ex

pert

s.



Correlates of Trade Show Formation

Participants’ Objectives

In the exhibitor survey, exhibitors were asked to rate (on a 1–7 Likertscale) the importance they placed on 12 attendance objectives and howwell the show performed on those objectives. Those objectives were:

• Number of leads• Quality of leads• Introducing new products• Meeting current customers• Selling at the show• Generating awareness for specific products• Delivering a specific message• Improving company awareness/image• Enhancing corporate morale• Identifying new prospects• Market testing new products• Gathering competitive information

We conducted an exploratory factor analysis for the exhibitors’ objec-tives and report the results in Table 3. Our results are largely consistentwith previous studies (e.g., Kerin and Cron 1987). The factor analysisreveals a two-factor structure of the exhibitors’ objectives: selling factorand nonselling factor. Objectives such as Number of Leads, Quality ofLeads, Introducing New Products, and Selling at the Show are moreheavily loaded on the selling factor while objectives such as ImprovingCompany Awareness/Images, Enhancing Corporate Morale, IdentifyingNew Prospects, Market Testing New Products, and Gathering Competi-tive Information are more heavily loaded on the nonselling factor. Theother three objectives are more evenly loaded on both factors. The two-factor model explains 52% of the total variance.

In the attendee survey, attendees were asked which of the followingreasons explained why they attended.

• See new products/developments• Fact finding for future purchases• Make a purchase• Attend seminars/association meetings



• To see specific company(s)/product(s)• To find products to represent• Solve a problem• Network with peers• Obtain technical or product specifications

As these are binary (Yes-No) data, a standard approach is inapplicable.An appropriate approach is to assume that there is a latent continuousvariable and a threshold constant (e.g., Bock and Lieberman 1970;Christofersson 1975; Muthen 1978) for each observed index , in which

if and only . Then, the latent factor analysis model can bewritten as , which has the same mathematical form as theusual common factor analysis model, and is assumed to have a multi-variate normal distribution with mean zero. Because the latent variable is unobservable and the observed variable x is dichotomous, the estimatedcorrelation matrix will be the tetrachoric correlation matrix. There areseveral estimators developed for this model. Bock and Lieberman (1970)used an MLE estimator and Christofersson (1975) and Muthen (1978)developed a GLS estimator. We use Christofersson’s approach via PRELISof LISREL and report the tetrachoric correlation matrix in Table 4a.

TABLE 3. Exploratory factor analysis of exhibitors’ objectives

Variable Factor Loadings

Factor 1 Selling Factor 2 Nonselling

Number of leads 0.78 0.01Quality of leads 0.79 0.09Introducing new products 0.79 0.13Meeting current customers 0.43 0.33Selling at the show 0.69 0.21Generating awareness for specific products 0.61 0.40Delivering a specific message 0.56 0.54Improving company awareness/image 0.16 0.61Enhancing corporate morale 0.13 0.68Identifying new prospects 0.30 0.68Market testing new products 0.09 0.79Gathering competitive information 0.05 0.62

Eigen value 4.68 1.61Cumulative variance explained 39.00% 52.42%

Note: The Eigen values of the factor analysis are 4.68, 1.61, 0.97, 0.84, 0.73, 0.65, 0.58, 0.45,0.44, 0.37, 0.35, 0.29. The minimum Eigen value criterion suggests a two-factor pattern.

xi* t

xixi = 1 xi

* > tx F* = +l e

ex*



Similarly, we conducted an explanatory factor analysis for attendees’objectives. We use the tetrachoric correlation matrix as the input for thefactor analysis and report the results in Table 4b. The factor pattern forattendees’ objectives is not as clear as that for exhibitors’ objectives. Theminimum eigenvalue criterion would suggest a four-factor structure ofour data. However, the third and fourth largest eigenvalues are 1.17 and

TABLE 4. Exploratory factor analysis of attendee’s objectives

Variable Tetrachoric Correlations

a. The tetrachoric correlation matrix of attendee’s objectives

See new products/developments 1.00Fact finding for future purchases 0.19 1.00Make a purchase 0.11 0.12 1.00Attend seminars/association

meetings0.16 0.19 0.22 1.00

To see specific companies(s)/product(s)

–0.09 0.10 0.09 0.18 1.00

To find products to represent –0.06 –0.10 0.11 –0.10 0.24 1.00Solve a problem 0.09 0.22 0.44 –0.07 0.39 –0.02 1.00Network with peers –0.12 –0.07 –0.06 0.13 0.19 0.14 0.07 1.00Obtain technical or product

specifications0.33 –0.03 0.11 0.05 0.30 0.03 0.21 0.17 1.00

Variable Factor Loadings

Factor 1 Buying Factor 2 Nonbuying

b. The exploratory factor analysis based on the tetrachoric correlation matrix of attendee’s objectives

See new products/developments 0.07 –0.30Fact finding for future purchases 0.41 –0.20Make a purchase 0.73 –0.01Attend seminars/association meetings –0.02 0.21To see specific companie(s)/product(s) 0.40 0.68To find products to represent 0.12 0.49Solve a problem 0.84 0.15Network with peers –0.20 0.73Obtain technical or product specifications 0.11 0.33Eigen value 2.02 1.47Cumulative variance explained 22.40% 38.79

Note: The Eigen values of the factor analysis are 2.02, 1.47, 1.17, 1.10, 0.94, 0.85, 0.73,0.47, 0.25. The minimum Eigen value criterion suggests a four-factor pattern but weretained a two-factor solution for better exposition.



1.10, very close to the traditional cut-off threshold of the minimum eigen-value criterion. For ease of interpretation we retained a two-factor structure.Consistent with previous research (e.g., Kerin and Cron 1987), we findthat these two factors are best described by Buying factor and Non-Buyingfactor. Objectives such as Fact Finding for Future Purchases, Make aPurchase, and Solve a Problem are more heavily loaded on the buyingfactor while objectives such as See New Products/Developments, AttendSeminars/Association Meetings, Network with Peers, and Obtain Technicalor Product Specifications are more heavily loaded on the Non-Buyingfactor. Other objectives split across factors. The two-factor modelexplains 38.79% of the total variance.

Participants’ Product Interests

At the construction trade show, the products are classified into 16 types.

• Earthmoving equipment• Compaction equipment• Power equipment• Paving equipment• Lifting/hoisting equipment• Concrete/asphalt plants• Crushing/grinding/screening equipment• Drilling equipment• Trucking equipment• Mixer• Auxiliary equipment• Parts• Devices• Concrete additives• Computer hardware/software• Others

In the exhibitors’ survey, respondents reported what product types theyexhibit. In the attendees’ survey, respondents reported what role they playin the purchase of each product type, using the same product classification.We define breadth of product interests for an exhibitor or attendee as thenumber of product types in which the exhibitor or attendee shows interest.

Figure 2 reports the distributions of breadth of product interests forboth exhibitors and attendees. We find that the means (1.23 for exhibitors



and 2.20 for attendees) of the breadth of product interests for both exhibi-tors and attendees are significantly greater than 1 (at p = 0.01), indicatingsignificant cross product interest for both exhibitors and attendees.

Determinants of Trade Show Diversity

A simple, integrated way to investigate our Propositions regarding thecorrelates of trade show diversity is the binary logit model. Such a modelwould use the data from Table 2 to predict the likelihood that a show inone of the industries in Table 1 would be, say, horizontal. Note that fromthis perspective, the first row of Table 1 reports 379 vertical shows and 78horizontal shows (totaling 457 in the industry) and the formation of thoseshows should be driven by the characteristics identified in the first row ofTable 2. There characteristics are measured using a 7-point Likert scale(1 = Low and 7 = High) and are defined as follows.

• EXSELL: Exhibitors’ selling propensity• ATBUY: Attendees’ buying propensity• EXPROD: Exhibitors’ breadth of product interest• ATPROD: Attendees’ breadth of product interest• TECH: The industrial technology innovativeness• DIFFPROD = |ATPROD – EXPROD|: The disparity of the breadth

of product interests between exhibitors and attendees

The first column of Table 5 describes the results from running thismodel, with the observations being weighted by the total number ofshows in the industry. As can be seen from the table, all variables are sig-nificant at 5% and beyond with the signs consistent with Propositions 1–4.

The negative coefficients of EXSELL and ATBUY show that thelower the selling/buying propensity from exhibitors or attendees thehigher the likelihood of horizontal shows to be formed, supporting Proposi-tion 1. The positive coefficients of EXPROD and ATPROD show that thehigher the breadth of product interests from exhibitors or attendees, themore likely horizontal shows are to be formed, supporting Proposition 2.

The results on the disparity of breadth of product interests betweenexhibitors and attendees are interesting. First, the negative coefficient ofDIFFPROD on horizontal shows in equation (2) suggests that the greaterthe difference in the breadth of product interest between exhibitors andattendees, the less likely horizontal shows are to be formed, directlysupporting Proposition 3. Second, assuming other things being equal,



TABLE 5. Results of simple logit, sequential two-stage and simultaneous equations models

Model Type

Simple Logit Sequential Two-Stage Simultaneous Equations

Coefficients for proportion of horizontal showsConstant –7.166*** –7.561*** –6.881***

(1.023) (1.093) (1.652)EXSELL –0.586** –0.624** –1.188***

(0.238) (0.250) (0.380)ATBUY –0.283** –0.255* –0.558**

(0.143) (0.143) (0.261)EXPROD 0.625*** 0.682*** 1.737***

(0.212) (0.217) (0.342)ATPROD 1.938*** 1.934*** 1.906***

(0.272) (0.275) (0.397)DIFFPROD –1.256*** –1.110*** –1.718***

(0.277) (0.282) (0.397)TECH 0.563*** 0.617** 0.570**

(0.175) (0.192) (0.252)Residual –0.510

(0.341)

Coefficients for total number of showsConstant 1.593 1.593

(1.284) (1.011)EXSELL –0.004 –0.004

(.153) (0.113)ATBUY –0.047 –0.047

(0.121) (0.099)EXPROD 0.329* 0.329**

(0.178) (0.144)ATPROD –0.275* –0.275**

(0.148) (0.121)DIFFPROD 0.274 0.274*

(0.198) (0.161)TECH 0.537*** 0.537***

(0.115) (0.274)

Variance-Covariance of errorsSIGMA1 0.485***

(0.075)SIGMA2 4.033***

(0.926)RHO –0.105**

(0.052)Log Likelihoods –388.29 –14.58 (stage 1) –400.08† –387.21 (stage 2)

Note: Standard errors are in parenthesis. ***, **, and *denote significance at 1%, 5%, and 10%,respectively. †Reported as general fit indices; not comparable across models.



forming smaller vertical shows over a larger horizontal show would leadto an increase in the total number of shows organized when DIFFPROD ishigh. Our empirical analysis further supports this argument: the coeffi-cient of DIFFPROD on total number of shows in equation (1) is positiveand significant.

Finally, the positive coefficient of TECH shows that the higher level ofindustry technology innovativeness, the more likely horizontal shows areto be formed, supporting Proposition 4.

Robustness Checks

We took several steps to assess the robustness of these results. We cal-ibrated standard OLS and probit models and found consistent results, withthe logit model demonstrating both superior fit and predictive validity.However, a concern with the simple binary logit model is the exogenoustreatment of the number of shows (n) as the weighting variable becausethe determinants of diversity may also affect n, which makes the numberof shows endogenous and the estimates biased. To assess our Propositionsmore rigorously, we ran an OLS model on ln n (rather than on n as n isrestricted to be nonnegative), took the residual as a control of the endoge-neity of n, and used it as an additional explanatory variable in the logitmodel.1 The result of this two-stage model is reported in the second col-umn of Table 5. The upper half of the column reports the logit coeffi-cients. As before, all the determinants turn out to be significant and areconsistent with our Propositions.

A problem with the two-stage estimation procedure is that while such aprocedure produces consistent estimates, those estimates are not efficient ingeneral, compromising our confidence in the significance of the estimates.Estimation efficiency may not change the signs of the estimates, but doeschange the significances, which consequently changes the supports of ourPropositions. The most rigorous way to assess our Propositions involvesestimating the nonlinear simultaneous equation system jointly using anMLE

where Xi are the exogenous determinants of trade show formation(i.e., Constant, EXSELL, ATBUY, EXPROD, ATPROD, DIFFPROD,

ln n X ui i i= +g (1)

n Binomial n F X vih

i i i~ ( , ( ))b + (2)



TECH), F (.) is the logistic cdf,2 (ui, vi ) represent error terms that are

jointly normally distributed as ,

and ni and represent the total number of shows and the number of

horizontal shows in industry i. This approach,3 although methodologicallymore complex, provides a more rigorous test of our Propositions.

The results from this simultaneous two-equation estimation approachare described in the third column of Table 5. These results are consistentwith those we obtained from the sequential two-stage estimation. Thecoefficients determining the log of n (reported in the middle of the col-umn) coincide with those obtained in the two-stage model4. The correla-tion coefficient (reported along with the standard deviations of the errorterms at the bottom part of the column) is negative and also significant at5%. More importantly, the beta coefficients (reported on the top part ofthe column) are significant at 5% and beyond, and appear with signs thatsupport our Propositions.

In summary, all three approaches support our Propositions on the cor-relates of trade show diversity. Notice that the error correlation betweenthe number of shows and the diversity of shows is significant (at 0.05),suggesting the need for a simultaneous approach to testing our Proposi-tions. That correlation also shows that several of the drivers we have iden-tified that affect show diversity—breadth of product interest forexhibitors and attendees and technological innovativeness of the indus-try—appear to affect the number of shows in an industry as well.

DISCUSSION AND CONCLUSIONS

We have studied the diversity of trade show types across industries andhave found that horizontal shows are more likely to be formed if (1) par-ticipants have lower selling/buying propensity and/or higher breadth ofproduct interest (short-term oriented) or (2) if the industry is more innova-tive in technology (long-term oriented). We also found that these keydrivers are closely related to the number of shows formed in an industry.

This exploratory, empirical examination of trade show formationshould help researchers better conceptualize and develop coherent theo-ries on trade shows, an important need in trade show research. The lack oftheory in this area has produced empirical results to date that are hard to

u

vNormali

i

⎛⎝⎜

⎞⎠⎟

⎛⎝⎜

⎞⎠⎟

⎛

⎝⎜

⎞

⎠⎟

⎛

⎝⎜

⎞

⎠⎟~ ,

0

012

1 2

1 2 22

s rs s

rs s sni

h



reconcile; empirical explorations on the correlates of trade show diversityshould help shed light on understanding those results in particular and onthe trade show generation process in general.

Managerial Implications

One implication of our research is that it provides a framework to pre-dict trade show diversity based upon an understanding of the level and thecongruence between exhibitor and attendee objectives and breadth ofproduct interests. To illustrate the managerial implications of that frame-work, we use the simultaneous equation model with a jackknife approach:for each industry we used the data from the other 20 industries as thecalibration sample and predicted the proportion of horizontal shows forthe single hold-out industry. The residual is the difference between theempirical proportion of horizontal shows in that industry and the pre-dicted proportion, based on the estimated coefficients of the model fromthe calibration sample. Table 6 gives the results (along with full sample-fitted values and residuals). From a managerial perspective, it appears thatthe packaging and the building and construction industries may be over-served by horizontal shows (large positive residuals) while the photographicindustry appears to be underserved by those types of shows (large nega-tive residuals).5 Thus there may be opportunities for organizers in theseindustries to profitably redefine the focus of some of the shows there.

Hence our framework, which links trade show participants’ objectivesand industry characteristics to the diversity of trade shows, has clear man-agerial uses: it can help trade show organizers diagnose and improve theappropriate mix of shows in a given industry and help trade show partici-pants select better shows to attend, depending on their show objectivesand buying/selling interests.

Limitations and Future Research

Our research has been exploratory; it suggests some possible theoreti-cal results that should be assessed more rigorously with broader anddeeper data sets. Our explorations relied on expert judgments for severalof the variables. While our use of multiple estimation models and ourjackknife analysis suggests that our results are relatively robust; replica-tions with other experts and perhaps triangulation with more objectivedata would be desirable.

Our simultaneous model estimates showed that the total number andtypes of trade shows formed in an industry is related to two trade show


418

TA

BLE

6. P

redi

ctiv

e va

lidity

ana

lysi

s of

the

sim

ulta

neou

s eq

uatio

n m

odel

Indu

stry

Act

ual H

oriz

onta

lS

how

Pro

port

ion

Ful

l Sam

ple

Jack

knife

Est

imat

ed H

oriz

onta

lS

how

Pro

port

ion

Res

idua

lE

stim

ated

Hor

izon

tal

Sho

w P

ropo

rtio

nR

esid

ual

Com

pute

rs a

nd c

ompu

ter

appl

icat

ions

17.0

0%23

.62%

–6.5

5%23

.71%

–6.6

4%C

omm

unic

atio

ns1.

00%

7.87

%–6

.94%

7.84

%–6

.91%

Eng

inee

ring

97.0

0%99

.48%

–2.6

0%10

0.00

%–3

.12%

Med

ical

and

hea

lth c

are

4.00

%15

.05%

–10.

80%

19.2

9%–1

5.03

%H

ousi

ng0.

00%

12.0

8%–1

2.08

%12

.32%

–12.

32%

Foo

d pr

oces

sing

and

dis

trib

utio

n95

.00%

84.3

1%11

.00%

82.9

8%12

.34%

Ele

ctric

al a

nd e

lect

roni

cs12

.00%

20.4

6%–8

.60%

16.7

7%–4

.91%

Pai

nt0.

00%

9.30

%–9

.30%

9.63

%–9

.63%

Nur

sing

0.00

%9.

67%

–9.6

7%10

.47%

–10.

47%

Aut

omot

ive

and

truc

king

94.0

0%86

.17%

7.45

%87

.29%

6.33

%C

hem

ical

s53

.00%

62.4

0%–8

.91%

59.1

2%–5

.63%

Pla

stic

s10

0.00

%89

.40%

10.6

0%89

.41%

10.5

9%R

adio

, TV

, and

cab

le3.

00%

13.9

8%–1

0.95

%15

.61%

–12.

58%

Ene

rgy

100.

00%

95.7

6%4.

24%

95.7

4%4.

26%

Bui

ldin

g an

d co

nstr

uctio

n40

.00%

32.4

2%7.

58%

5.06

%34

.94%

Res

taur

ants

and

food

ser

vice

s10

0.00

%95

.70%

4.30

%95

.68%

4.32

%M

anuf

actu

ring

71.0

0%74

.01%

–2.5

8%82

.08%

–10.

66%

Pho

togr

aphi

c0.

00%

17.9

5%–1

7.95

%18

.57%

–18.

57%

Wel

ding

0.00

%14

.70%

–14.

70%

15.4

5%–1

5.45

%P

acka

ging

100.

00%

65.8

6%34

.14%

64.7

2%35

.28%

Edu

catio

n0.

00%

10.0

6%–1

0.06

%10

.35%

–10.

35%

Unw

eigh

ted

MA

D10

.05%

11.9

2%

MA

D =

Mea

n ab

solu

te d

evia

tion.



specific correlates: breadth of product interest—by attendees and exhibi-tors and an industry-specific correlate—and the technological innovative-ness of an industry. A more complete theory of trade show formation willrequire the identification of economic and other drivers beyond thoseidentified here, such as other industry-specific correlates including indus-try competitiveness, concentration, product profitability, growth rate, andproduct life-cycle stage.

Trade shows remain an important and understudied element of thecommunications mix, demanding more research to better understand howthey can best be used and how to better assess their effectiveness. Wehope research such as this and the research directions noted above willadd to our understanding and improve our use of this important marketingcommunications vehicle.

NOTES

1. We also tried using a count data model to explain n instead of running an OLSmodel (with ln n as the left hand side variable). Both Poisson and negative binomialregression models score much worse in terms of log likelihoods (the values are –113.98and –94.94 as opposed to –14.58 for the OLS model) and hence, AIC. Because of thisfinding and also because of the difficulty of extending such a model to the two-stageapproach outlined next, we followed the simple OLS approach.

2. We used the cdf for the normal distribution in estimation and prediction as well andfound that the cdf of the logistic distribution has better goodness-of-fit statistics and per-forms better on tests of holdout predictive validity.

3. See the Appendix for the model estimation procedure.4. This is not a surprising result. OLS coefficient estimates and MLE estimates are

identical for linear regression models with normally distributed errors. A close look atequation (3) in the Appendix reveals that the stage one likelihood value dominates thejoint likelihood value, ensuring close coefficient estimates for stage-one for both thesequential 2-stage model and the simultaneous model.

5. From Table 6 we note that, from a predictive validity perspective, our model doesquite well. More than half our sample comes from the first two categories, yet our predic-tion error for computers is −6.64% while it is −6.91% for communications. These compareto an average mean absolute deviation (MAD) of 11.92% for the sample as a whole (on anunweighted basis). Apart from one industry, building and construction, the full sample andjackknife predictions correspond quite well, which makes that industry a possible candi-date for an outlier. (Indeed, we reran the simultaneous equation analysis after deleting thisindustry and found a substantial effect on our assessment of our Propositions). The out-come of the jackknife procedure demonstrates the robustness of these results, as thoseindustries that feature disproportionately large numbers of shows do not determine theresults obtained.



REFERENCES

Blythe, J. (2000). Objectives and measures at UK trade exhibitions. Journal of MarketingManagement 16:203–222.

Bock, D. and M. Lieberman (1970). Fitting a response model for n dichotomously scoreditems. Psychometrika 135:179–197.

Borghini, S. F. Golfetto, and D. Rinallo (2006). Ongoing search among industrial buyers.Journal of Business Research 59(10/11):1151–1159.

Calton, J. (1994). Comdex complaint as big computer show opens, key exhibitors grumbleabout costs. Wall Street Journal, November 14, A1.

CEIR (2001). CEIR direct: The newsletter of the Center for Exhibition Industry Research(January–February).

Christofersson, A. (1975). Factor analysis of dichotomized variables. Psychometrika40:5–32.

Dekimpe, G. M., P. Francois, S. Gopalakrishna, G. Lilien, and C. Van den Bulte (1997).Generalizing about trade show effectiveness: A cross-national comparison. Journal ofMarketing 61(October):55–64.

Donberg, D. (2006). Plastics show visitors increase. Paper, Film and Foil Converter80(8):18.

Godar, S. H. and P. J. O’Connor (2001). Same time next year—Buyer trade show motives.Industrial Marketing Management 30(1):77–86.

Gopalakrishna, S. and J. D. Williams (1992). Planning and performance assessment ofindustrial trade shows: An exploratory study. International Journal of Research inMarketing 9:207–224.

Gopalakrishna, S. and G. L. Lilien (1994). A dynamic model of business trade showeffectiveness. ISBM Report: 3–1994.

Gopalakrishna, S. and G. L. Lilien (1995). A three stage model of industrial trade showperformance. Marketing Science 14:22–42.

Gopalakrishna, S. G. Lilien, J. Williams, and I. K. Sequeira (1995). Do trade shows pay-off?Journal of Marketing 59:75–83.

Hansen, K. (1996). The dual motives of participants at international trade shows: Anempirical investigation of exhibitors and visitors with selling motives. InternationalMarketing Review 13(2): 39–53.

Hansen, K. (1999). Trade show performance: A conceptual framework and its implicationsfor future research. Academy of Marketing Science Review 99(08). Available at: http://www.amsreview.org/amsrev/theory/hansen08–99.html.

The Industrial Robot (2006). World’s largest robots and vision show increases attendanceby 55 percent. The Industrial Robot 33(2):140–141.

Johnson, J. (2005). Trade shows stay separate. Waste News, December 19.Kerin, R. A. and W. L. Cron (1987). Assessing trade show functions and performance: An

exploratory study. Journal of Marketing 51:87–94.Marken, G. A. (2004). Trade show PR . . . It’s time to focus on goals, substance, results.

Public Relations Quarterly 49(3):30–33.Muthen, B. (1978). Contributions to factor analysis of dichotomous variables. Psychometrika

43:551–560.



Pinar, M., J. D. Rogers, and D. Baack (2002). An examination of trade show participationin a developing country: An exploratory study in Turkey. Journal of Euromarketing11(3):33–52.

Sharland, A. and P. Balogh (1996). The value of non-selling activities at internationaltrade shows. Industrial Marketing Management 25:49–56.

Shoham, A. (1999). Performance in trade shows and exhibitions: A synthesis and directionsfor future research. Journal of Global Marketing 12(3):41–57.

Show Daily of Conexpo and Con/Agg, ’96 (1996). Program of the Conexpo and Con/Agg’96 trade show.

Trade Show Bureau (1994). A guide to the US exposition industry. Denver, CO: TradeShow Bureau Resource Center.

Williams, J. D., S. Gopalakrishna, and J. M. Cox (1993). Trade show guidelines forsmaller firms. Industrial Marketing Management 22:265–275.

Zook, C. and J. Allen (2001). Profit from the core: Growth strategy in an era of turbulence.Boston: Harvard Business School Press.



APPENDIX: ESTIMATION APPROACH FOR SIMULTANEOUS MODEL

Because , the log-likelihood function for

our simultaneous model equations (1) and (2) is

where f denotes the standard normal density function. Because the inte-gral above does not admit a closed-form solution, we calculate it usingnumerical integration (via Gauss quadrature). Our estimates are obtainedby maximizing equation (3) using Newton-Raphson algorithm.

v u N ui i i| ~ ( )rss

s r2

122 21, −

⎛⎝⎜

⎞⎠⎟

LLi

i in X( ), , lnlnb g

sF

gs

Σ = +−⎛

⎝⎜⎞⎠⎟

⎛

⎝⎜⎞

⎠⎟⎧⎨⎪

⎩⎪

+

=∑Contant

1

21

1 1

1

lln ( ( )) ( ( ))

(ln

F X v F X v

ni Xi

ini

h

ini

vb b

s rF

rss

+ − +

−

− −

−∞

∞

∫ 1

1

122

2

1v gg

s r

)

221−

⎛

⎝

⎜⎜⎜

⎞

⎠

⎟⎟⎟

⎫

⎬⎪

⎭⎪

dv (3)



IMPLICATIONS FOR BUSINESS MARKETING PRACTICE

Trade shows represent a major element of the communication mix inbusiness-to-business marketing: they make up the third largest componentof the marketing commutation mix behind direct marketing and businessmagazine advertising. In the United States alone, the trade show industryannually hosts over 12,000 shows, involves over two million exhibitorsand over 100 million attendees, and generates over $50 billion.

Yet, very little is known on how a trade show is formed and why twodifferent types of shows (i.e., horizontal and vertical) coexist in the exhi-bition industry. For example, some industries are dominated by horizontalshows (e.g., engineering, food processing and distribution, etc.) whilesome are dominated by vertical shows (e.g., medical and health care,housing, etc.). This article is the first to investigate the key issues in tradeshow formation and diversity. It can help trade show organizers diagnoseand improve the appropriate mix of shows in a given industry. It helpstrade show participants select better shows to attend, depending on theirshow objectives and selling or buying interests.

We analyzed the trade show industry within industry (trade show–specific factors) and across industry (industry-specific factors). Withinindustry, we find that both exhibitors and attendees of trade shows vary intheir objectives. Exhibitors may seek immediate sales at the show or seekto fulfill nonsales objectives such as obtaining competitive marketingintelligence or both. On the other hand, attendees may seek to fulfillimmediate buying needs or nonbuying objectives such as new productinformation. We find that for a market with higher selling or buying pro-pensity, vertical shows are more prevalent while for a market with ahigher breadth of product interests, horizontal shows are likely to be moreprevalent. In addition, we find that a highly innovative industry has morehorizontal shows.

This article provides a methodology to analyze the nature of theobjectives of trade show participants and the linkage on how theseobjectives are related to trade show formation and diversity. For exam-ple, the jackknife approach illustrated in this article provides a frame-work to predict trade show diversity based on an understanding ofselling or buying propensity levels and matches and likewise for breadthof product interests. For a given industry, this approach can forecastwhether the industry has the most appropriate balance of vertical or hor-izontal shows, providing guidelines for show organizers on what typesof shows to organize.



Trade show participants (exhibitors and attendees) can use these resultsto determine, given their goals (breadth of product interest and buying orselling intensity), which are the most appropriate shows to attend. Andshow organizers, as noted previously, can use our results to identifyindustries that are underserved by one show type or another and uncoverpotentially lucrative business opportunities

Although the nature of our study is exploratory, our framework offersan initial step in building up a theory of trade show formation and diver-sity that will ultimately benefit all participants in the exhibition industry.


An Exploratory Study of Trade Show Formation and Diversity.pdf

Documents

romanian ministry

louisiana

pennsylvania

tetrachoric

massachusetts

trade show

show generating

minimum eigenvalue