Review of Industrial Organization 9: 1-23, 1994
`© 1994 Kluwer Academic Publishers. Printed in the Netherlands.
`
`First-Mover Advantages from Pioneering New
`Markets: A Survey of Empirical Evidence
`
`WILLIAM T. ROBINSON
`School of Business Administration, University of Michigan, Ann Arbor. MI 48109, U.S.A.
`
`GURUMURTHY KALYANARAM
`
`School of Management, 2601 North Floyd, University of Texas at Dallas, Richardson, TX 75083, U.S.A.
`
`and
`
`GLEN L. URBAN
`Sloan School of Management, MIT, SO Memorial Drive, Cambridge, MA 02139, U.S.A.
`
`Abstract. Market pioneers can develop iirst-mover advantages that span decades. The most general
`first-mover advantage that helps explain higher pioneer market share levels is a broad product line or
`brand proliferation. In markets for experience goods, pioneers tend to shape consumer tastes and
`preferences in favor of the pioneering brand. While the preliminary results vary by industry, they
`indicate that market pioneers do not tend to perish more often than later entrants. Accounting profits
`for market pioneers generally are lower in the first four years of operation, but higher thereafter.
`Overall, market pioneers follow innovative strategies that have high initial costs and risks, but yield
`high potential returns.
`
`Key words: First-mover advantages, market share, market pioneers.
`
`I. Introduction
`
`Being the market's first entrant is generally more costly than being an early
`follower or a late entrant. The reason is that product innovation tends to be more
`costly than product imitation (Mansfield, Schwartz and Wagner. 1981; Levin et
`al., 1987). Will the first entrant's higher costs be offset by revenue gains? If so,
`innovation is encouraged, which is a key source of economic growth.
`Though the empirical evidence in economics generally supports first-entrant
`revenue gains, it is based on a few industries, such as pharmaceuticals (Bond and
`Lean, 1977; Gorecki, 1986; Hurwitz and Caves, 1988; Grabowski and Vemon,
`1992) and cigarettes (Whitten, 1979). More extensive empirical work on first-
`mover advantages comes from research in business schools.
`We survey empirical evidence of first-mover advantages from pioneering new
`markets. (Also, see Scherer and Ross, 1990, pp. 582-592.) One key research
`stream examines the impact of order of market entry on market share. The survey
`
`AstraZeneca Exhibit 2139
`Mylan v. AstraZeneca
`IPR2015-01340
`
`Page 1 of 24
`
`

`
`2
`
`WILLIAM T. ROBINSON ET AL.
`
`results consistently show that market pioneers tend to maintain market share
`advantages over later entrants. For example, across broad samples of business
`units, Robinson and FomeU (1985) and Robinson (1988a) estimate the empirical
`association between order of market entry and market share. It is almost as strong
`as the empirical association between market share and profitability (Shepherd,
`1972; Ravenscraft, 1983). Note that Schmalensee (1989) classifies the association
`between market share and profitability as one of "the main empirical regularities
`that have been uncovered in inter-industry research" (p. 953). These two general
`tendencies suggest that market pioneers also tend to have higher profitability.
`How have so many pioneers maintained market share advantages for literally
`decades? We address this question as well as several others:
`• How is a market pioneer identified?
`• Can the market pioneer shape consumer tastes and preferences?
`• Are market pioneer advantages often sustained by crushing later entrants with
`aggressive reactions?
`• Do market pioneers typically start with superior skills and resources?
`• How long does it take later entrants to reach their long-term (asymptotic) share
`level?
`• How important is the pioneer's leadtime in developing sustainable market share
`advantages?
`• Do market pioneers have higher accounting profits than later entrants?
`These questions are typically addressed by examining surviving market pioneers
`and later entrants. Such research provides insights into the rewards for market
`pioneering, conditional on survival. Because market pioneering is both costly and
`risky, the survival issue is also examined. Finally, public policy implications are
`addressed.
`
`n. Identifying the Market Pioneer
`
`A fundamental consideration is how to identify the market pioneer. Though
`various definitions have been proposed (Lieberman and Montgomery, 1988), the
`market pioneer is typically defined as "the market's first entrant." Implementing
`this definition requires that (1) entrants be distinguished from firms that attempt
`to enter a new market, but fail and (2) market boundaries be established.
`To be classified as an entrant, a business should reach a competitive scale of
`commercialization. A competitive scale gives the market pioneer an opportunity
`to capture first-mover advantages. If competitive scale is not reached, a firm should
`be classified as having failed in its attempt to enter a new market. For example,
`Whitten (1979) requires that an entrant in a national cigarette market be supported
`by national advertising. Urban et al. (1986) require that an entrant in a national
`market for consumer packaged goods achieve national distribution. Hence, if a
`firm does not have national advertising or national distribution in a national
`market, it does not qualify as an entrant.
`
`Page 2 of 24
`
`

`
`A SURVEY OF EMPIRICAL EVIDENCE
`
`3
`
`At what point in the market's evolution should market boundaries be deter-
`mined? Market boundaries are typically determined after customer acceptance for
`a new product has been established. The reason is that it is very difficult to assess
`the commercial potential of a dramatic innovation at the time of its introduction.
`For example, when Xerox entered the photocopying market in 1958, the commer-
`cial potential for plain-paper copying was not recogtiized (Washington Post, 1985).
`Even industry experts did not foresee plain-paper copying as a new market, but
`it tumed out to be one because it generated many more photocopying applications.
`Plain-paper and coated-paper copying were different markets because they were
`not viewed by customers as close substitutes. That fact could not be determined in
`1958 however, and Xerox was not immediately recognized as the market pioneer.
`Once customer acceptance has been established, how should market boundaries
`be determined? Market boundaries typically reflect customer substitution in use.
`In practice, they have been based on accepted industry practice (Whitten, 1979),
`consumer evaluations of products that are close substitutes (Urban et al., 1986),
`and managers' evaluations of their company's target customers (PIMS Data Man-
`ual, 1979). As a whole, these market boundaries tend to be narrower than four-
`digit SIC codes. The general insights into market pioneer advantages described
`below are robust to these varied definitions of market boundaries.
`In contrast, Golder and Tellis (1993) identify market pioneers using historical
`analysis. This innovative approach relies on objective information from multiple
`sources at the time the market originated, thus avoiding reliance on industry
`representatives to identify the market pioneer many years after the market's
`beginning. Though industry representatives are experts, Golder and Tellis argue
`their response is often based on personal recall or on the firm's oral tradition. As
`a result, high market share firms can be misidentified as market pioneers. The
`main strength of historical analysis is that it overcomes this problem.
`One weakness of historical analysis is that it can miss the start of new markets
`that are based on dramatic innovations. For example, Golder and Tellis identify
`Xerox as a later entrant, not as a market pioneer. Also, historical analysis requires
`a great amount of time to implement. After gathering usable information from
`more than 450 articles, Golder and Tellis piesent detailed insights on only 36 mar-
`kets.
`Table I highlights the importance of these market pioneer definition issues.
`Across Golder and Teilis' 36 markets, only four market pioneers remain the
`current market leaders. Note that Golder and Tellis do not require the pioneer to
`reach a competitive scale of commercialization. Hence, their definition can identify
`a very low market share firm as a market pioneer. For example, Golder and Tellis
`identify Trommer's Red Letter as the market pioneer of light beers. Trommer's
`was introduced by Piels Brewing Company of Brooklyn New York and was only
`on the market for about six weeks. Prior to Miller Lite, Trommer's and other
`brewers of diet beers were described as "an aberration in the minds of floundering,
`small-time brewers" (Marketing and Media Decisions, 1983). By more conven-
`
`Page 3 of 24
`
`

`
`WILLIAM T. ROBINSON ET AL.
`
`a — __ „ ™
`
`nil
`
`U
`
`*
`
`f^
`
`^
`
`.caa
`
`00
`
`photo
`Instar
`
`3u
`
`'«
`
`paper
`using
`aB Copy
`
`IE
`
`E«a.
`
`^ I
`
`r-^ E
`
`Page 4 of 24
`
`

`
`A SURVEY OF EMPIRICAL EVIDENCE
`
`5
`
`tional definitions. Miller Lite and seven other current market leaders in Table I
`are classified as market pioneers. This increases the total niunber of pioneers that
`are current market leaders from four to 12.
`In summary, historical analysis helps avoid the problem of high market share
`firms being misidentified as market pioneers. Though time consuming, historical
`analysis in conjunction with current industry insights should yield the most accurate
`market pioneer definitions. Either approach used in isolation can generate mislead-
`ing results.
`
`m. Industry Studies
`
`The pharmaceutical industry has received the most attention. Bond and Lean
`(1977) analyzed two United States ethical drug markets. After more than 10 years
`in the market, both first entrants maintained market share leadership. Hurwitz
`and Caves (1988) examined market share levels for 29 original patent holders in
`United States pharmaceutical markets. At varying numbers of years after patent
`expiration, average pioneer market share was 63%. Two years after patent expir-
`ation, the average pioneer market share was 51%. These results point to powerful
`first-mover advantages in the United States pharmaceutical industry. Gorecki
`(1986) provides similar conclusions for the Canadian pharmaceutical industry.
`What was the main first-mover advantage? Bond and Lean conclude the main
`advantage was physician preference for the established and familiar pioneering
`brand names. Because patent protection has numerous limitations, "trademark
`protection . . . appears to be far more powerful than patent protection" (p. 77).
`Market share leadership was not driven by the pioneer's superior product quality
`or lower prices. For example, though quality differences can arise across ethical
`drugs, the first movers in Hurwitz and Caves' sample all faced generic substitutes.
`Even so, these first movers charged an average price premium of 127%. Hurwitz
`and Caves estimate that "a 10 percent increase in the leader's price premium loses
`it only three- to four-tenths of a percentage point of market share" (p. 314).^
`Market share leadership does not appear to be driven by promotional spending.
`Bond and Lean conclude (p. vi), "the data appear to reveal that sales and pro-
`motional dominance go hand-in-hand. Nonetheless, the data also show that the
`opportunities for gaining sales via promotion are decidedly limited. Qualitative
`characteristics such as the timing of entry and therapeutic novelty appear to
`determine both the profit-maximizing level of promotion and the sales associated
`with that promotion". Though first-entrant promotional spending was relatively
`large, it was smaller than competitors' as a percentage of sales.
`When facing strong first-mover advantages, what can a later entrant do to btiild
`market share? Bond and Lean say "physicians can be persuaded to prescribe late-
`entering brands if those brands offer some therapeutic gain useful to a subset of
`patients" (p. 76). For Canadian provinces, Gorecki recommends certifying brands
`as therapeutically equivalent and allowing pharmacists to select a lower priced
`
`Page 5 of 24
`
`

`
`6
`
`WILLIAM T. ROBINSON ET AL.
`
`brand. Hence, one approach is to offer a better product. Another is to match
`product quality, cut price, and encourage price competition.
`
`Whitten (1979) studied seven cigarette submarkets that started between 1913 and
`1958. In 1974, all seven pioneering brands were still in the market with an average
`market share of 51%. Though it is not clear why these pioneering brands were so
`successful, lower prices can be eliminated because price competition was rare. As
`found by Bond and Lean, first entrants' advertising spending was large, but
`relatively small as a percentage of sales. For the three pioneers that lost their
`number one rank, better flavor from later entrants contributed to their decline.
`Mascarenhas (1992) found that market pioneers had higher market shares in
`national markets for semisubmersible oil-drilling rigs. He speculates these market
`share advantages arose from pre-empting prime drilling sites and from buyer
`switching costs.
`For 58 financial innovations, Tufano (1989) reports that investment banks that
`created the financial innovations did not charge relatively high prices during their
`brief period of monopoly. After entry, prices typically remained below those of
`rival imitative products. By being first and charging prices that were no higher
`than rivals', the first movers were rewarded with a higher long-run market share.
`The financial innovators were also rewarded by an influx of new customers to
`whom they could sell additional services.
`What conclusions can be drawn from these industry studies? First-mover advan-
`tages as reflected in higher long-term market share levels are important in national
`pharmaceutical, cigarette, semisubmersible oil-drilling rig, and investment banking
`industries. The sources of these first-mover advantages are varied, but customer
`familiarity and brand loyalty are important. Overall, a market pioneer's enduring
`trademark protection is more important than patent protection.
`Later entrants who want to overtake the pioneer often give customers either a
`product or price incentive to switch brands. Lower prices can increase market
`share, but opportunities for price competition are often limited. Promotional
`spending is most effective when coordinated with a unique and meaningful product
`benefit.
`Are the market pioneer advantages unique to the industries cited or do they
`typically arise across markets? This is a key issue because later entrants can benefit
`from free-rider effects, resolution of technological and market uncertainties,
`changing technologies, changing consumer needs, and incumbent inertia (Lieber-
`man and Montgomery, 1988). Examples of market pioneers that have been quickly
`overtaken by later entrants are Reynolds Intemational Pen (ballpoint pens). Bow-
`mar Instruments (hand-held electronic calculators), Osbome (portable computer
`market), and Royal Crown Cola (diet and caffeine-free colas). In contrast to
`industry-specific examples and counterexamples, cross-sectional research yields
`insights into general tendencies.
`
`Page 6 of 24
`
`

`
`A SURVEY OF EMPIRICAL EVIDENCE
`
`Table IL Order of market entry and market share for consumer packaged goods*
`
`4th
`
`5th
`
`6th
`
`_ -__- 1
`
`1.2
`
`_ ---1
`
`3.9
`12.4
`
`_ -_1
`
`8.1
`15.5
`13.8
`
`Market
`
`share forecasts (%)
`
`3rd
`
`- -2
`
`5.4
`20.8
`17.9
`15.9
`
`2nd
`
`- 4
`
`1.5
`31.0
`25.4
`21.9
`19.4
`
`1st
`
`100.0
`58.5
`43.6
`35.7
`30.8
`27.3
`
`Entry order
`
`Share relative to
`pioneering brand
`
`First
`Second
`Third
`Fourth
`Fifth
`Sixth
`
`1.00
`0.71
`0.58
`0.51
`0.45
`0.41
`
`* From Urban et al. (1986. p. 654). Positioning quality and advertising spending are held constant in
`these maricet share forecasts.
`
`IV. Cross-Sectional Studies
`
`Urban et al. (1986) estimate the relationship between order of market entry and
`market share across consumer packaged goods. The sample includes 129 major
`brands in 36 product categories. The first and second entrant both averaged more
`than 25 years in the market. Product category examples are chewing gum, freeze-
`dried coffee, fabric softener, and furniture polish.
`Market share perfonnance is measured by the later entrant's share relative
`to the first entrant's share. Holding positioning quality and dollar advertising
`expenditures constant. Table II shows that the «,h entrant's share relative to the
`pioneer's roughly equals 1 divided by the square root of its order of entry. For
`example, the fourth entrant has roughly one half of the pioneer's share. In a
`market with four competitors, these results predict a market share of roughly 36%
`for the first entrant and roughly 18% for the fourth entrant.
`How are these pioneer share advantages maintained for literally decades? Vari-
`ous aspects of consumer buyer behavior should benefit the pioneering brand. For
`packaged goods, it is important to separate consumer trial from repeat purchase.
`Kalyanaram and Urban (1992) define trial as the percentage of consumers who
`have tried tbe brand. Repeat purchase is the cumulative percentage of triers who
`repeat by period t.
`Trial should be higher for market pioneers than for later entrants. Schmalensee
`(1982) argues that even for identical products, the risk of an unfavorable experi-
`ence motivates rational constmiers to continue buying the pioneering brand. Along
`sitnilar lines, Hauser and Wemerfelt (1990) say, "if two brands enter the market
`with the same distribution of perceived utility, the brand that enters earlier will
`be considered more often. If it is considered more often, it should have a higher
`market share" (p. 400). Kardes et al. (1993) also find that pioneers are included
`more often in the consideration set. All of these forces should lead to greater
`consumer thai for pioneering brands.
`Repeat purchase should also be higher for pioneering brands. One reason is
`that consumers learn more about pioneering brands because of their longer time
`
`Page 7 of 24
`
`

`
`8
`
`WILLIAM T. ROBINSON ET AL.
`
`on the market. Kardes and Kalyanaram (1992) conclude, "The learning advantage
`conferred to the pioneering brand translates into more extreme and confidently
`held judgments of the pioneer. Judgments held with conviction are persistent over
`time and resistant to competitors' activities. Together, these judgmental processes
`lead to long-run pioneering advantage" (p. 356).
`Another reason for higher repeat purchase is that for experience goods, tastes
`and product attribute importance actually shift toward the first entrant. Carpenter
`and Nakamoto (1989) report "order of entry affects the stnicture of consumer
`preferences for brands in the category, yielding pioneers a superior position and
`a substantially higher share of buyers' choices" (p. 294). Thus, for experience
`goods, the pioneering brand can shape consumer tastes along the lines of its own
`characteristics. This helps the pioneer to be recognized as the industry standard.
`Though behavioral research has not yet linked these various hypotheses to
`market share, Kalyanaram and Urban (1992) estimate the impact of order of entry
`on both trial and repeat purchase. Their consumer packaged goods sample covers
`eight first entrant and 18 later-entrant brands. Early entry increases both trial and
`repeat purchase, but the estimated trial impact is substantially larger than the
`impact on first repeat purchase (0.48 vs. 0.26). Hence, a late entrant typically
`appears to face more problems in gaining trial than in gaining repeat purchase.
`Market pioneers should also benefit from favorable distribution, which is espe-
`cially important for consumer packaged goods. In Alpert, Kamins, and Graham's
`(1992) survey of 145 resellers of supermarket products, resellers had more favor-
`able attitudes toward pioneering brands because they tend to be more effective
`at "meeting an unmet need, contributing incremental sales, generating excitement
`about going shopping, and having potential to achieve high volume" (p. 36).
`When the pioneering brand generates favorable attitudes among resellers, has
`a higher consumer trial, and higher repeat purchase, the pioneer should typically
`spend less on advertising and promotion as a percentage of sales. Favorable
`attitudes reduce the need for advertising. Higher consumer trial and repeat pur-
`chase reduce the need for promotions targeted to consumers, such as coupons.
`These forces should also help reduce promotions targeted to retailers, such as
`discounts to gain favorable shelf space.
`Across established consumer goods businesses, Buzzell and Farris (1977) con-
`clude that "the combination of early entry and high market share almost certainly
`leads to lower marketing costs" as a percentage of sales (p. 136). Fomell, Robin-
`son, and Wemerfelt (1985) reach a similar conclusion. Thus, market pioneers,
`and especially ones that have maintained a high market share, tend to have lower
`advertising and promotional costs as a percentage of sales.
`Given these pioneering advantages, how can a late entrant catch the pioneer?
`One option is to increase the advertising budget. To offset the order-of-entry
`disadvantage. Urban et al. (1986) estimate the second entrant's dollar advertising
`budget should continually be 3.4 times the pioneer's budget. If the pioneer reacts
`with a budget increase, the second entrant's dollar budget must be increased even
`
`Page 8 of 24
`
`

`
`A SURVEY OF EMPIRICAL EVIDENCE
`
`9
`
`Table in. Order of market entiy and market share for mature consumer and industrial goods'
`
`Order of market entry
`
`Average market share
`Consumer
`(n = 371)
`29%
`17%
`12%
`18%
`' From Robinson and Fomell (1985. p. 310) and Robinson (1988a. p. 89).
`
`Market pioneer
`Early follower
`Late entrant
`R^
`
`Industrial
`
`(n = 1209)
`29%
`21%
`15%
`9%
`
`further to maintain the 3.4-to-l.O advantage. Hence, it is unlikely that a typical
`second entrant can profitably catch the pioneer by using only heavier advertising.
`Introducing a superior product affords more opportunity. The second entrant is
`estimated to catch the pioneer if its product continually yields 36% more prefer-
`ence after consideration.^ Though 36% does not seem out of reach, it can be
`difficult to achieve. This is because as a market matures, introducing a brand with
`a major product advantage becomes increasingly difficult (e.g., see Robinson,
`1990).
`
`A. CONSUMER AND INDUSTRIAL GOODS
`
`The cross-sectional evidence cited applies to consumer packaged brands. Is market
`pioneering important for consumer goods in general? For industrial goods? Robin-
`son and Fomell (1985) and Robinson (1988a) address these questions using the
`PIMS (Profit Impact of Market Strategies) data.
`A PIMS business sells one kind of products or services, to an identified target
`market, and faces clearly defined competitors. A market pioneer is defined as
`"one of the pioneers in first developing such products or services". A later entrant
`is either an early follower or a late entrant. An early follower entered a "still-
`growing, dynamic market". A late entrant entered a "more established market
`situation".
`Grouping businesses in three broad categories typically yields more than one
`pioneer per market. This approach increases random measurement error, which
`tends to weaken the empirical results. Even so, market pioneers in PIMS typically
`receive long-term market share rewards. Table III shows that across 371 mature
`consumer goods business, market pioneers had an average market share of 29%,
`early followers 17%, and late entrants 12%. More than 80% of these pioneers
`had been in the-market for 20 years or more. The empirical results are similar,
`though not quite as strong for pioneers in mature industrial goods markets.
`The descriptive strength of this empiricai association can also be measured by
`the percentage of variation in market share explained by order of entry alone.
`Order of entry alone explains 18% of the variation in market share for consumer
`goods and 9% for industrial goods. A useful point of comparison is the empirical
`
`Page 9 of 24
`
`

`
`10
`
`WILLIAM T. ROBINSON ET AL.
`
`association between market share and return on investment (ROI). In the PIMS
`sample, market share alone explains 21% of the variation in ROI for consumer
`goods and 8% for industrial goods.^ Hence, the empirical association between
`order of entry and market share is almost as strong as the association between
`market share and ROI.
`The PIMS studies also help explain how so many pioneers maintain important
`market share advantages for long f)eriods of time. Two important dimensions of
`a business's offerings are perceived product quality and product line breadth.
`Perceived quality is based on customer perceptions of quality. Though such infor-
`mation is subjective, managers who report perceived quality to PIMS are encour-
`aged to use input from market research, the salesforce, and distributors. Perceived
`quality equals the percentage of sales that are superior to competitors' less the
`percentage inferior.
`In both consumer and industrial markets, pioneers in the market less than 20
`years tend to have higher perceived quality. Pioneer perceived quality advantages
`could arise from proprietary experience, by defining industry standards, or from
`the consumer-leaming-based advantages described previously. The pioneer prod-
`uct quality advantages deteriorate sharply after 20 years or more.
`During their initial monopoly and near-monopoly period, market pioneers have
`an opportunity to introduce numerous products for the biggest and best market
`segments. This can force later entrants to target smaller market niches with more
`focused product lines, a product line strategy consistent with brand proHferation
`models (e.g., Schmalensee, 1978). In both consumer and industrial markets, mar-
`ket pioneers tend to have broader product lines than late entrants. Product line
`advantages show limited deterioration after 20 years or more. Thus, product line
`breadth advantages tend to be more sustainable than product quality advantages.
`Because of this sustainability, pioneers often pre-empt competition by using
`rapid product line extensions. For example, when Quaker Oats introduced its
`Ouaker Chewy Granola Bar, it did not introduce a single flavor but four flavors
`simultaneously. In the short run this effort strained Quaker's limited capacity, but
`in the long-run it helped the company maintain leadership of the granola snack
`category (Quaker Quarterly, 1984).
`In both PIMS samples, market pioneer advantages do not appear to be influ-
`enced directly by product patents or trade secrets. Price differences do not gen-
`erally play a key role in explaining market share. These results are similar to those
`of the industry studies described previously. In contrast to the pharmaceutical
`industry, average pioneer prices are within 2% of late-entrant prices. Pioneers
`had average total cost savings from purchasing, manufacturing, and physical distri-
`bution of roughly 1 to 2%. A portion of these cost savings helps build market
`share, but the associated market share impact is estimated to be less than one
`share point. These results are consistent with the modest importance of manufac-
`turing scale economies in United States markets (Scherer and Ross, 1990).
`Market pioneer advantages are also influenced by industry charaaeristics. In
`
`Page 10 of 24
`
`

`
`A SURVEY OF EMPIRICAL EVIDENCE
`
`H
`
`consumer markets, pioneers tend to have higher shares when the product's pur-
`chase price is less than $10. This observation is consistent with Schmalensee's
`(1982) model of informational advantages for first movers in markets where con-
`sumers tend to buy from habit.
`In industrial markets, pioneer share levels tend to increase as the product's
`purchase price increases. When big-ticket items carry a high level of perceived
`risk, a switching-cost advantage accrues because customers tend to rely on known
`suppliers and brands. Also, if experience advantages are more important for big-
`ticket items, the market pioneer should benefit.
`Seller concentration is also associated with pioneer market share advantages.
`Using the PIMS data. Parry and Bass (1989) report higher pioneer share advan-
`tages in oligopolistic than in perfectly competitive markets. This can arise because
`entry barriers that increase concentration also give the pioneer first-mover advan-
`tages.
`
`B. SAMPLE HETEROGENEITY
`
`First-mover advantages are often conditional on various business- and industry-
`specific factors, such as firm proficiency, luck, and environmental changes.(Kerin,
`Varadarajan and Peterson, 1992). However, the cross-sectional empirical results
`combine many different product categories. Does this type of heterogeneity bias
`the empirical results?
`Kalyanaram and Wittink (1993) examine five categories of consumer packaged
`goods. Within individual categories, they find wide differences in the impact of
`order of market entry on market share. Even so, the average impact of order of
`entry on market share is more than 50% stronger when estimated within individual
`categories. This indicates that because of aggregation bias, broad cross-sectional
`regression studies tend to underestimate the true pioneer market share advantages.
`Ramaswamy et al. (1993) develop an empirical pooling approach that uses the
`data to detennine relatively homogeneous groups of PIMS businesses. Market
`pioneers are more likely to serve national or intemational markets, whereas later
`entrants are more likely to serve regional markets. This supports the product
`proliferation observation above that market pioneers often pick off the biggest
`and best market segments for themselves. Later entrants are the most likely to
`serve customers that value product quality or sales force support. This type of
`product differentiation should help later entrants offset the modest scale economy
`cost disadvantages they face versus market pioneers.
`
`C. ALTERNATIVE EXPLANATIONS
`
`Are the market pioneer advantages described due to the pioneer's intrinsically
`superior strength rather than to first-mover advantages? This issue was raised by
`Vanhonacker and Day (1987), Keck and Rao (1987) and more recently by Moore,
`
`Page 11 of 24
`
`

`
`12
`
`WILLIAM T. ROBINSON ET AL.
`
`Boulding and Goodstein (1991). For example, Moore and his co-authors used
`Hausman's specification test to evaluate the exogenous versus the endogenous
`nature of market pioneering. Their results indicate that market pioneering is
`endogenous.
`If market pioneering is endogenous, a critical bias in the empirical results can
`arise if market pioneers are intrinsically stronger than later entrants. The bias is
`critical if long-term market pioneer advantages are based on superior skills and
`resources at the time of entry rather than on first-mover advantages.
`Though the literature predicts skill and resource differences for market pioneers,
`it does not predict intrinsic superiority for market pioneers. For example, Lieber-
`man and Montgomery (1988, p. 54) say, "Firms whose entrepreneurial vision and
`new product R&D are excellent will tend to find first-movership attractive, whereas
`finns having relative skill bases in manufacturing and marketing tnay not". Lamb-
`kin and Day's (1989) ecology model as well as empirical results of Lilien and Yoon
`(1990) and Robinson, Fomell, and Sullivan (1992) support differences in skill and
`resource pattems. They do not support intrinsic superiority for market pioneers.
`Kalyanaram and Urban (1993) take this research a step further by linking
`pioneer skills and resources to market share. Order of market entry is specified
`as a function of the firm's skills and resources and the maximum anticipated
`market share for the entrant. Order of market entry is then linked to market
`share. In the empirical results, the market share penalty for later entrants increases
`when entry is specified as endogenous. This result suggests that when market
`pioneering is treated as exogenous, the estimated market share rewards are too
`low.
`Are market pioneer advantages sustained by crushing later entrants with aggress-
`ive reactions? In a sample dominated by industrial goods businesses, Robinson
`(1988b) finds that the most likely product, distribution, marketing expenditure, or
`pricing reaction to entry is no reaction. For consumer packaged goods, Kalyanaram
`and Urban (1993) examine reactions to entry in terms of price, promotion expendi-
`tures, advertising expenditures, and distribution. Only the advertising expenditure
`reaction is significant.
`Across 18 pharmaceutical markets, Grabowski and Vemon (1992, p. 347) find
`that lower priced generic products challenged higher priced branded products.
`They conclude, "The pioneering firms did not attempt to deter entry through their
`pricing strategies. Rather, in most cases, the firms continued to increase their
`prices at the same rate at which existed prior to entry". Kim (1992) reports similar
`findings