A Run Del

8/2/2019 A Run Del

1/23

First Mover (Dis-)Advantage and Real Options

Tom Cottrell

Gordon Sick

University of Calgary

Faculty of Management

Finance Area

July 30, 2001

8/2/2019 A Run Del

2/23

Abstract

We compare first mover advantages against the real option arisingfrom delay and flexibility. The real options model recognizes the valueof delaying projects until risk can be resolved. This value to delayis offset by the convenience value of possessing an operating project.Sometimes this convenience value is in the form of a first mover advan-tage. We believe that fear of losing first mover advantages has causedmany managers to ignore real options analysis completely and simplygo ahead with any project that they think has a positive net presentvalue. In this paper, we investigate first mover advantage to find thatit usually isnt all that it is cracked up to be. By considering the mer-its of a delayed-entry follower strategy, we show that value enhancingmanagers will want to be suitably cautious before ignoring the realoption analysis.

1

8/2/2019 A Run Del

3/23

1 Introduction

Strategists often ignore real option valuation because they believe there aresignificant first-mover advantages to quickly entering a market. Rather thanfollow the real options methodology, they argue that if entry is delayed,a competitor will reap the rewards of moving first. But there can alsobe significant disadvantages to early entry: most obvious, early entrantsspoil the value of their real option to delay. Beyond that, there are oftenadvantages to observing how the market responds to competitor productsbefore putting ones name brand and quality reputation on the line. Inthis paper, we consider the trade-off between real options and first-moveradvantages.

The real option approach has signifcant advantages over traditional cap-ital budgeting techniques when there is significant uncertainty in the upsidepotential and the investment is sunk. In this case, the challenge to the in-vestment decision is to determine the conditions for exactly when to committo the investment. But in order to have the ability to choose the optimaltiming of exercise, the firm must be able to delay the investment decision.If a competitor threatens to enter the market ahead of the firm, the value ofthe real option can be significantly reduced, or spoiled entirely. The benefitsaccruing to a firm that is first into a market are termed first mover advan-tages, and corporate strategy identifies a number of important benefits tothe firm that is first to enter a market.

In our view, a first mover advantage can be understood as a convenience

dividend on a real option. Imagine two competitors with options to expandinto a new market. The firms in this example are otherwise identical, withthe exception that one is able to enter the market ahead of the other. Ifthe first entrant into the market gains advantages, with the result that itcaptures the ability to preclude the second mover from gaining significantmarket share, there are significant advantages to moving first. In this case,the second-mover is constrained in its strategy because of a threat of rapidexpansion by the first-mover. When this occurs, there are waiting costs tothe second mover, and this results in a convenience dividend that is capturedby the entrant who moves early. The competitive effect of a potential entrantpushes the first mover to enter the market more quickly than would be

desirable. Recent academic work on this concept is found in [3] [4],[1], [5]and [6]. See also [14] which examines real options for entry in the contextof networks.

But first mover advantages are often ephemeral, and occasionally failto materialize even when they are expected to be significant. For example,

1

8/2/2019 A Run Del

4/23

microcomputer software and hardware markets are expected to be charac-

terized by benefits from early entry. It is widely recognized that gaininga large installed base is an important aspect of strategy in these markets.Yet early entrants were rarely able to retain a dominant position. In thehardware market, for example, the Apple II gained significant early marketshare only to be superceded by the IBM PC. According to some accountsof the nascent personal computer industry, there were important learningadvantages to IBM for watching Apple enter the market. One history in-dicates that IBM learned two critical lessons from the Apple II: an openarchitecture (e.g. slots for 3rd party cards such as video or pointer devices)and pricing. It wasnt that IBM had not developed a personal computerby 1979, but it was over-priced and lacked certain compatibility features.

IBMs market share slipped as Compaq and later Dell introduced compati-ble machines with more competitive prices. These later entrants identifieda follower strategy that would circumvent the dominant market share thatIBM had established.

A similar story describes several components of the operating systems en-vironment. Although Xerox pioneered the graphical user interace, for exam-ple, competitors Apple (with the Macintosh) and Microsoft (with Windows)reaped the benefits. Word processors WordStar and WordPerfect were eachdominant in the early 1980s, but eventually Microsoft Word dominated themarket. Similarly, in the spreadsheet market, VisiCalc was superceded byLotus 1-2-3 and eventually by Microsoft Excel. While each of these marketswas expected to provide important first mover advantages, early entrants

failed to retain a dominant position.These software markets are characterized by network standards. Strate-

gic maneuvering to capture these standards can provide incentives for eitherearly or late entry into the market excessively rapid adoption of stan-dards, so that more advanced standards are not adopted, or excessivelyprotracted adoption rates, so that the adoption of promising standards isdelayed. The important trade-off between rapid adoption of a current stan-dard and deferred adoption for expected technological improvements turnson the expected rate of improvement in the technology. Farrell and Saloner[2] describe these alternatives in Competition, Compatibility and Stan-dards: The Economics of Horses, Penguins and Lemmings. The standards

competition for lemmings is that after one market participant adopts acertain technology, new adopters blindly follow in order to be compatible.The widely cited example is the success of the IBM PC standard. In contrast,penguins reflect a greater reluctance in market entry. Hungry penguins siton the edge of the ice flow, hoping to enter the water and hunt for food.

2

8/2/2019 A Run Del

5/23

Table 1: Product markets and successful followers

Product First Mover Replacement Follower

Global SatelliteCommmunication Iridium Globalstar

Computing Apple II PC IBM, Dell

GUI Xerox Alto/Star Macintosh Apple

Spreadsheet VisiCalc 1-2-3 Lotus

Spreadsheet Lotus 1-2-3 Excel Microsoft

Word Processor Wordstar Wordperfect WordPerfect Corp.

Word Processor Wordperfect MS-Word Microsoft

European Ariel, Triumph, Japanese Honda,Motorcycle BSA, Ensign Motorcycle Yamaha, Suzuki

Electric motors Electricaland lights Edison systems GE

Electronic NewsRetrieval Reuters Bloomberg

Concern over predators lurking in the depths make it risky to enter the wa-ter. Eventually, one or two penguins are pushed into the water, and thoseremaining on the ice flow watch to see whether they survive or not. These

are markets where early entry can be disastrous, and competitors watch andlearn. We believe that competition in these markets is hazardous for earlyentrants, as illustrated in the following table.

A host of examples demonstrate the advantages to second or third en-trants in markets outside of computing or standards competition. There arestudies of imitation strategies in both the consumer and industrial products.For the former, Steven Schnaars discusses 28 cases in [13] where imitationstrategies succeeded over first movers. Table 2 lists the product markets.Schnaars identifies a number of sources of first mover advantages: imageand reputation, brand loyalty, market positioning, technological leadership,set product standards, access to distribution, experience effects, patents,

and switching costs. It is not our position that first-mover effects are notimportant. We only argue that they should be considered in the contextof the trade-off with loss of the option to delay, especially when first-movereffects can be difficult to bring to realization. For several of the advantagesto first-movers, we now consider counter-examples.

3

8/2/2019 A Run Del

6/23

Table 2: Product markets where imitators succeeded pioneers (Schnaars)

35 mm cameras Money-market mutual fundsAutomated Teller Machines Magnetic Resonance ImagingBallpoint pens Nonalcoholic beerCaffeine-free soft drinks Personal computer operating systemsCAT Scanners Paperback booksCommercial jet aircraft Personal computersComputerized ticketing services Pocket calculatorsCredit/Charge cards Pro jection TVDiet Soft drinks SpreadsheetsDry beer Telephone answering machines

Food processors Video Cassette RecordersLight beer VideogamesMainframe computers Warehouse clubsMicrowave ovens Word-processing software

1.1 Image and reputation

One important motivation for innovation is that consumers will identify thefirm with creative innovative ideas. Rather than a pejorative copycatimage, the firm may be seen as a cutting edge innovator. It is hard to

argue with the positive benefits of a strong reputation, but we do questionthe meaningfulness of this reputation effect in a wide variety of markets.Consider the first commercial jet aircraft, the Comet. With the introductionof the very first jet, consumers attributed technological and innovative savvyto the company. Eventually, however, the world discovered that the airframewas unsafe. Steel used in the pressurized cabin became brittle over time andseveral planes eventually exploded at high altitudes. The Comet companyeventually went bankrupt, despite its innovative design and early marketshare. Associating ones name with such a colossal failure is a dangerousrisk, and second-movers in the market have fared well. While less dramatic,the failure of Iridium satellite network faces a similar criticism. While earlyentry surely afforded important advantages, the downside risk of bankruptcymight have been reduced with a more cautious roll-out.

4

8/2/2019 A Run Del

7/23

1.2 Brand loyalty and switching costs

Early entrants may gain advantages in loyal customers who are unlikely totry competing follower products. If switching costs are involved in the use ofa new techology, the first entrant to a market benefit because consumers whoadopt the product would have to incur a cost to use a competing product.This means that later entrants must offer a discount to the first moversinstalled base or else expect customers to eventually switch to the entrantsnewer technology. Either way, entrant profits are eroded by the differencebetween what customers would be willing to pay and the switching costto induce the change. Yet we find several examples where switching costsare not important. Apple was quick to adopt and promulgate a graphicaluser interface, yet Microsoft Windows, the later entrant, has a dominant

position. While the loyalty of Apple users is legendary, the growth in themarket meant that brand loyalty was less important than market position.

1.3 Experience effects

Learning curves play a role in competition in many industries, and the abil-ity to enter early and move quickly down the cost curve is often consideredimportant in corporate strategy. Yet it is sometimes difficult to realizethe benefits of lower costs as higher profits. The semiconductor industryis famous for the benefits of learning curves, and Texas Instruments oncedominated the industry. The cost curve was important at TI, where unitcosts for semiconductors were found to decline to 73% with each doubling ofproduction. Throughout the 1960s and 70s, TI aggressively pursued growthand output in order to gain a low-cost producer position. By 1982, however,TIs market position began to erode. TIs single-minded focus on push-ing down the experience curve eventually led to standardized commodityproducts, and the return on investment declined. In addition, the semi-conductor manufacturing that TI specialized in became obsolete with theadvent of newer metal-oxide technologies. As a result, competitors Intel andMotorola, who had been working in the new semiconductors, found them-selves in a stronger position. TIs focus on early entry and rapid marketexpansion, and effort to realize experience curve benefits, lead to a myopiccorporate strategy.

1.4 Patents

Biotech industry giant Genentech is widely known as an innovative and ag-gressive competitor. As an early entrant, Genentech was quick to build

5

8/2/2019 A Run Del

8/23

Table 3: Sample means of responses on patent effectiveness

Method of appropriation Processes Products

Patents to prevent duplication 3.52 4.33

Patents to secure royalty income 3.31 3.31

Secrecy 4.31 3.57

Lead time 5.11 5.41

Learning curve 5.02 5.09

Sales or service 4.55 5.59

a strong patent base and to develop a thicket of patents around not onlymolecules but also the processes by which useful products were manufac-tured. Founded by noted scientist Herbert Boyer with Robert Swanson in1976, Genentech pioneered and patented innovative biotech products andprocesses. However, defense of those patents came at a price. Intellectualproperty law for the technology required legal precedent for litigation. Whilethe patent prima facie protects the product, it is left for litigators and courtchallenges to determine the true strength of the protection. Throughout itshistory, Genentech has invested in litigating alleged patent infringements.Industry followers could play a far less risky wait-and-see game to find outhow the courts would interpret intellectual property protection in the in-dustry. Some have argued that the expense and uncertainty associated with

the litigation process resulted in a cash-starved Genentech that was an easytarget for the $2.1B merger with Roche Holdings in 1990.

In the Yale Appropriability Study (see [8] and the discussion below),patents were found to be one of the weakest approaches to capture thereturns from innovation. The effectiveness of patents in securing competitiveadvantage was rated by 650 respondents on a Likert scale from 1-7. Theresponses related to innovation in either new products or the manufacturingprocesses itself. The results, abstracted from Table 1 of that report inour Table 3, illustrate the relative position of patents to other means ofappropriating the returns from invention. Clearly, being first to invent, andgaining a patent from first-mover advantages, is generally over-rated.

The open source software movement, such as Gnu Public License orLinux, demonstrate that product learning may be more important thanintellectual property protection. Software developers are willing to puttheir software into an open source licence because they believe it will helptheir product to become more standardized and more bullet-proof. With-

6

8/2/2019 A Run Del

9/23

out patent protection, the developer may lose first-mover advantage, but

benefits with greater product knowledge. In the open source software com-munity, the advantage to moving first may be the opportunity to influencethe standards upon which later versions of the software is based.

2 Imitator Strategies

Clearly a wide array of tactics are available for firms to pursue an innovativecompetitor. Schnaars lists several advantages to being a follower:

avoiding products that have no potential (as demonstrated by thefailed first mover),

lower R&D expenditures,

lower costs to imitating than innovating,

ability to catch innovator with heavy marketing expenditures,

lower costs of educating customers on innovative product ideas,

stranding the innovator with an obsolete product standard, and

ability to avoid from or benefit from the pioneers mistakes.

Of these, perhaps most widely understood is the fact that it is cheaper tocopy an innovation than to create it in the first place. Mansfield, Schwartz

and Wagner ([10]) empirically demonstrate that lower costs and reducedproduct development time favor an imitator. After examination of 48 inno-vations in chemicals, drugs, electronics and electrical machinery, they esti-mate that imitation costs averaged 65% of innovator costs. A more compre-hensive and extensive survey of imitation cost and time occured in the YaleAppropriability study ([8]). Survey respondents were asked to estimate thecosts of duplicating an innovation as a percentage of the innovators R&Dcost. While patenting increased both the time and the cost of duplication,most innovations in the survey were found to be duplicated at considerablecost savings to the imitator, as shown in Figure 1 for products. The costsavings are not so significant if there is patent protection, and indeed, for a

major product innovation with patent protection, the cost savings are min-imal, perhaps because the innovator is more inclined to litigate to protectthe patent. Similary results hold for processes.

Perhaps more striking are the reponses on the time it takes to dupli-cate product innovation, as in Figure 2. Usually, innovations (even major

7

8/2/2019 A Run Del

10/23

New Product CostsTypical Innovation

0%

10%

20%

30%

40%

50%

60%

< 25% 26-50% 51-75% 76-100% >100% Cannot

Costs as % of innovator

%o

fResponses

Patented

Unpatented

New Product CostsMajor Innovation

0%

10%

20%

30%

40%

50%

60%

< 25% 26-50% 51-75% 76-100% >100% Cannot

Costs as % of innovator

%o

fResponses

PatentedUnpatented

Figure 1: Estimated cost to duplicate products, as a proportion of costrequired by innovator.

8

8/2/2019 A Run Del

11/23

innovations) can be duplicated within 1 to 3 years. The results for process

innovation are similar. Of course, these estimates vary by line of businessand the follower firms capabilities, but they serve as an important bench-mark in considering the real impact of imitation strategies.

3 Examples of Winning Late Movers

We next review in some detail winning second movers: the Betamax versusVHS battle in VCRs and Microsofts remarkable late entries.

3.1 VHS versus Beta

The video cassette recorder (VCR) adoption and standards war has beenwidely1 discussed. Commercial video tape recorders (VTRs) had been devel-oped by the Ampex corporation of Redwood California, using an innovativetechnology involving a helical scanning system that allowed a high-frequencytelevision signal to be stored on a relatively slow-moving magnetic tape.Ampex products were the standard equipment in television studios, but theproduct was difficult to thread and the company could not market its prod-uct to homes, schools or even industrial users because of the complexity andcost. Moreover, Ampex didnt have experience with transistorized circuits,so it negotiated a deal with Sony in 1960 whereby Sony would supply it withtransistorized circuits for its VTR products and Sony would have the rightto make VTRs with the helical scanning technology for non-broadcast users.

This sharing arrangement led to disputes, which were settled out of courtin 1968. Sony proceeded to develop products for the home and commercialmarket.

In 1970 Sony showed its new U-matic helical scan cassette tape systemsystem to Matsushita and its affiliate JVC. The cassette simplified loadingand was more compact. The three companies agreed to share this commonstandard, subject to incorporating some JVC technology for color. Althoughthey were equal partners, Sony made the most revenue and profit from thispartnership. The product never found a significant market for home use,but became popular in professional settings and with TV networks. It useda 3/4 tape, but was limited to one hour of recording time.

The Japanese companies experimented with various VCR products forthe home consumer, but none caught on. In December of 1974, Sony demon-strated its new Betamax (then with a one-hour recording time) to JVC and

1See, for example, [7] and [9]

9

8/2/2019 A Run Del

12/23

New Product Duplication TimeTypical Innovation

0%

10%

20%

30%

40%

50%

60%

< 6Mo 1/2-1Yr 1-3Yrs 3-5Yrs >5 Yrs Cannot

Time to duplicate innovation

%o

fResponses

Patented

Unpatented

New Product Duplication TimeMajor Innovation

0%

10%

20%

30%

40%

50%

60%

70%

80%

< 6Mo 1/2-1Yr 1-3Yrs 3-5Yrs >5 Yrs Cannot

Time to duplicate innovation

%o

fResponses

Patented

Unpatented

Figure 2: Estimated time to duplicate products, as a proportion of timerequired by innovator.

10

8/2/2019 A Run Del

13/23

Betamax and VHS VCR Sales to Dealers

1

10

100

1,000

10,000

100,000

1976 1978 1980 1982 1984 1986 1988

BetamaxVHS

Figure 3: VCR Sales (000s), Logarithmic Scale

invited them to share in a standard, along the lines of the standard of theU-matic cassette system. But, JVC was convinced that a 2-hour recordingtime was more valuable to consumers and was also concerned that Sony

had benefitted disporportionately from the U-matic standard. Thus, JVCdeveloped the Video Home System (VHS) to compete with Sonys Betamax.Sony entered the market in 1975 and rallied Zenith, Emerson, Sanyo and

Toshiba around its Betamax standard. JVC started selling its VHS systemin 1977, and had the other consumer electronics manufacturers, includingMatsushita (with its Panasonic brand), Hitachi, Sharp and RCA ralliedbehind the VHS standard. Sony shipped 30,000 units to dealers in 1976,and 140,000 in 1977, when total VHS sales were started at 80,000 units2

In each subsequent year, VHS outsold Betamax, with Betamax peaking at1,538,400 units in 1985 and VHS peaking at 11,582,000 in 1987.3.

Figure 3 shows the sales from 1976 to 1992. VHS won the battle on twofronts, and it won the battle quickly.

1. It had 2-hour capability when Betamax had 1-hour. Then it had 4-hour when Betamax had 2-hour capability. Subsequently Sony caught

2Fortune, July 16, 1979, pp. 110-116

3EIA Consumer Electronics U.S. Sales, 1985-1992, as cited in [11] [12]

11

8/2/2019 A Run Del

14/23

up with the recording length.

2. VHS established liasons with the nascent video rental industry anddeveloped a network effect because the consmers who had VCRs thatwere compatible with most other VCRs had a wider selection of videotitles at rental shops. Meanwhile, the rental shops knew that more con-sumers had VHS-compatible machines and stocked even more videosfor that format. This serves to highlight the importance of marketingand network externalities, and the relative unimportance of being firstto market.

Betamax was first to market and had a reputation of higher quality. Theuser interface was better: it was easier to set-up the timer for recording on a

Sony VCR. A variation of the Betamax system, called Betacam, became thestandard recording system for TV newscasters and networks. However, Sonymissed the importance of the length of recording times for consumers andcame up against a group of competitors that were upset that they hadntbeen treated fairly in the earlier partnership. In the end, it lost out onthe positive feedback of the network effect. Sony quickly lost its first-moveradvantage and subsequently started to sell VHS VCRs to consumers.

3.2 Microsoft: A Frequent Follower that Wins

Many industry observers note that Microsoft has rarely been an innovatorin the software market Instead, they have learned from watching consumers

discover what features are valued. Then they move to either acquire a firmwith appropriate technology or develop their own.

The software business has two sources of significant economies of scale4,leading to the potential of a monopoly in the market. For a software devel-oper, the programming cost is largely independent of the number of users,so it is fixed. Distributing software and supporting users is a variable cost.Thus, a company with deep pockets can gain entry to an existing marketby absorbing the fixed costs and pricing the product close just above vari-able cost on the assumption that it can gain market share and move to aprofitable price point. This enables one company to be a second successfulmover into a market, particularly if it can reduce development costs by ob-

4When there are economies of scale, the average cost of producing a unit decreaseswith increases in output: large fixed costs and network effects. If there are large fixedcosts of production and small variable costs, there are often important strategic effets ofscale economies.

12

8/2/2019 A Run Del

15/23

serving the successes and failures of others. It also illustrates the concept5

of a serial monopolist.A serial monopoly is characterized by a series of different monopolists

over time in a particular product space. For example, WordStar had adominant market share in the word processing market, and was followed byWordPerfect, which was finally displaced by Microsoft Word. Dan Brick-lin and Bob Frankston invented the spreadsheet in 1979 and called theirproduct VisiCalc. It dominated the market for several years and was su-perceded by the competing product Lotus 1-2-3 with the introduction ofthe IBM PC. In turn, Lotus 1-2-3 was finally displaced by Microsoft Excelwith the introduction of Microsoft Windows. The operating system marketfor microcomputers was originally dominated by CPM, then by Apple II,

Microsoft DOS and Microsoft Windows. All but the last of these transitionswas accompanied by a change in the hardware standard.With a serial monopoly, the initial monopolist is displaced by a well-

funded6 competitor who covers the fixed costs of development on the as-sumption of gaining dominant market share. The story of how they gainthe market share varies from product to product and according to whom istelling the story. Some argue that Microsoft is just a better software de-veloper.7 Others suggest that Microsoft ties in its end-user software to theoperating system market that it dominates. This is the position advancedby the Department of Justice in the recent Microsoft anti-trust trial.8 Evenin a market with a dominant player, it is possible to enter as a second-moverand accomplish a significant return on investment.

The other source of economies of scale in the software market is thenetwork effect that we have seen in the VHS vesus Betamax case. Usersshare files and thus would like their co-workers to have software compatibleto their own. If their co-workers have selected their software, then the valueof compatible software is higher to them. Once they also adopt that softwarestandard, they generate the same network effect with other users.

5See, for example [9].6E.g. Microsoft.7This is the p osition often cited in [9]. For example, their chapters 8 and 9 point out

how the successful software developer typically had high reviews from computer magazinesprior to winning market share.

8Others, such as [9] are quick to point out that Microsoft dominated the spreadsheetand word processor categories of the Macintosh market without having control of theoperating system. Some within the Macintosh community point out that Apple has hadto develop its operating system under the constraint that it wouldnt break the dominantMicrosoft software. The reasoning is that users would drift away from the Macintosh ifthey couldnt share word processing and spreadsheet files with users of Microsoft products.

13

8/2/2019 A Run Del

16/23

In the experience of one of the authors (Sick), Microsoft software has

often been involved in a special variation of the network effect that allows itto force upgrades on users that they dont really need. This happens when weor our students are working in groups. One member of the group, who we willcall the geek is often trying out the latest software to see whether or not itis an improvement over previous software. This person is often assigned thetask of developing a draft spreadsheet or word processing document for thegroup. When they do this with the latest version of Microsoft Word or Excel,they often discover that the software revision creates a new file format andis incapable of saving to the earlier file formats, at least in its earliest days.This makes sense if the new file format allows improved functionality. But,in many cases, a simple word processing document or simple spreadsheet

could successfully exist in the earlier file format, but the Microsoft softwareis incapable of doing the conversion. Or, the software offers to save as anearlier version, but warns the user that formatting or information maybe lost by saving to an earlier format. Not knowing what might be lost,the fearful geek lobbies his colleagues to upgrade to the latest version of thesoftware in order to be able to read his files. Note that the new versionof the software does not have to offer new functionality to the geek. It ismerely his curiosity that causes him to investigate the software. It is hissubsequent fear that makes him ask his colleagues to upgrade. Of course,once his colleagues upgrade, they can become party to the same upgradecycle on other colleagues. Before long, a whole cohort will be using theupgraded software and nobody will know or even needs to believe that it

provides any enhanced functionality for anyone in the group.9,10

9Indeed, the early releases of new versions of Microsoft software products often havelimited functionality to save to earlier versions, with the claim that this feature will beprovided later. Sometimes, the functionality is never provided. For example, one of theauthors has a moderately complicated Excel spreadsheet that was created in 1996 withMicrosoft Excel 5.0 on a Macintosh. It has never been modified by a subsequent version ofExcel. When the file is opened in a later version of Excel, such as Excel 8.0, and saved as anExcel 5.0/95 workbook, a dialogue box says This workbook contains features that are notsupported by the file format that you selected. To avoid the possible loss of formatting orinformation, be sure to save a copy in the most recent Microsoft Excel Workbook format.The workbook was unchanged, yet Excel 6.0 wouldnt save it in the original format. Ifthis is an important document, someone merely opening it in a colleagues copy of Excel8.0 has significant incentives to acquire that version of Excel for themselves in case the

file has been modified or damaged so that it cant be reliably used any longer with theirolder software.

10This is an example of Fear, Uncertainty and Distrust or FUD, which is the popularcharacterization of one way of a monopolist maintaining or developing a dominant position.The monopolist must make users fearful of the uncertainties of using other companys

14

8/2/2019 A Run Del

17/23

Microsoft initially thought that the Internet was not going to be an

important product for consumers. Thus, it avoided involvement in develop-ing Internet software and Internet standards. The first movers in the Webbrowser market were scientists and other academics. The University of Illi-nois developed the Mosaic browser that worked on multiple platforms andquickly became the graphical user interface of choice on the World WideWeb. This software was freely distributed. Mosaic project members likeMarc Andreesen left to create Netscape which became the dominant browseron multiple platforms. When Microsoft Chairman Bill Gates eventually re-alized the importance of the Internet for commercial computing, Microsoftlicenced the Mosaic code from the University of Illinois in order to developa competing browser, Internet Explorer (IE). IE was often smaller, faster

and more reliable than Netscape, depending on the version number and theplatform. Microsoft agressively entered the broswer market by bundling IEwith new releases of Windows and freely distributing the product over theWWW.

This approach quickly established market share. By distributing up-grades with each new operating system release, Microsoft continued to pro-tect its dominant position in operating systems, and made it extremelyunlikely that competitors could enter and survive. The claim of other par-ties was that Microsoft could expand a dominance in the browser marketto a dominance in forming other standards that suited it better than itscompetitors. For example, it distributed software for editing Web pagesthat did not follow the open Internet standards, but followed Microsofts

proprietary standards. A web surfer would find that many websites wouldnot work on other browsers (if the website was developed to proprietaryMicrosoft standards), and thus would settle on Internet Explorer as her ownstandard browser. There are faster web browsers than Internet Explorer,such as Opera, but they dont access some sites that IE accesses, so theuser has a choice of forgoing those sites, or tolerating the midling speed ofInternet Explorer.

To make a first mover advantage really valuable, there must be costs toa customer for switching the product of a new entrant. These could be

1. Costs when the entrants product will not read and write the legacyfiles of the customer.

standards or older versions of the software. If a competitor promises to introduce a newand better software revision, it can promise software to be delivered in the future thatwill outperform the competitors. If the company has some credibility, it may deter itscustomers from migrating to the new software. This is called vaporware if the softwaredoesnt materialize quickly.

15

8/2/2019 A Run Del

18/23

2. Training costs for the customer or its employees.

3. Incompatibilities with other software of the customer or its businesspartners.

These costs need only be apparent to the user and need not be out-of-pocket costs. If switching costs are low or the benefits of switching are veryhigh, a new supplier can enter the market. For example, if the customer ischanging one piece of hardware or system software, the cost of introducingnew ancillary software becomes lower. Thus, if the customer upgrades to anew computer and that computer has new system and ancillary productivitysoftware installed as an optional bundle, the costs of switching to the newsoftware become lower and the new entrant can gain a toehold in the market.

If the new entrant does not know when, where or how new hardware orsystem software will come into being, it is harder to overcome these switchingcosts.

It is interesting to see how quickly an encumbent can be overthrown whenthey dont pay attention to the competition. A 1987 Wall Street Journal[15] article interviewed Microsofts Jeff Raikes about its Excel spreadsheet,which had been successful in the Macintosh market, but hadnt entered theWindows market. He said Were going to be the dominant spreadsheet.The encumbent software on the PC was Lotus 1-2-3. When asked aboutthe competition, Jim Manzi, chairman of Cambridge, Mass.-based Lotus,said If theres a Rock of Gibraltar in this marketplace, its 1-2-3. InitiallyManzi seemed to be right, for in 1988, Lotus had almost 70% unit share of

the spreadsheet market while Excel was under 10%. By 1992, Lotus sharedipped to less than 50%, while Microsoft Excel was at 30%. By 1997, thereverse was complete, with Excel at almost 70% and Lotus at under 30%.11

Part of the reason for the switch from Lotus 1-2-3 to Excel was the superiorquality of Excel. The other reason was that the switching costs were reducedas the consumers also switched from DOS to Windows at the same time.Lotus chose to protect its legacy DOS user interface, while Microsoft wasfree to develop a graphical user interface consistent with Windows. Sinceusers were migrating to Windows, the switch in spreadsheets didnt costthem so much in personal training as if they had already been in Windowswith Lotus 1-2-3 or had stayed in DOS.

A similar transition occured with word processing software. In 1986,there were over 8 word processors on PCs running DOS and Windows. The

11These figures are from [9] Figure 8.6, page 175. In Figure 8.7, they show that Microsoftwas able to increase its price while Lotus had to decrease its price by 1997, since thedisparity in revenue was even greater than that in unit share.

16

8/2/2019 A Run Del

19/23

largest revenue market share was WordPerfect with almost 20%, followed by

Ashton-Tates Multimate at about 15% and Microsoft Word (Windows andDOS) at less than 10%.12 By 1990, consolidation started, with Wordperfectpeaking at about 45% revenue share, and Word for DOS and Windows atapproximately 30%. By 1997, the Windows version of Word rose to over90%, while the others shared the rest. Again, WordPerfect worked well in aDOS market, but Word had a friendly user interface and worked well withWindows. The switching costs were smaller as users switched to Windows.

It is also important to remember that the overall microcomputer marketwas growing strongly in this period and new consumers had few legacy coststo switching. Thus, for Excel and Word, Microsoft was able to be a successfulsecond or third mover for several reasons:

1. The products were freshly designed and of superior quality, buildingon the mistakes and successors of predecessors.

2. The products had little legacy to protect in terms of user training.

3. Switching costs were lowered because users were also switching oper-ating systems.

4. The market was growing sufficiently fast that legacy users werent tooimportant and future users were more lucrative.

Of course, second movers dont always win. In the Macintosh mar-

ket, the first major word processor was Microsoft Word and the first majorspreadsheet was Microsoft Excel. The mid-1980s versions of these productsbecame prototypes for the Windows versions. In 1988, Microsoft Excel hadover 60% of the Macintosh spreadsheet market. Informix (a large Unix soft-ware vendor) entered the market with Wingz in 1989 and captured almosta 20% market share, but that declined thereafter even when it was takenover by Apples software company Clairs and renamed Resolve.13 Lotus hadentered the Macintosh market in 1985 with an integrated software productcalled Jazz, that included a spreadsheet and word processor. However, itwas pre-announced and very late in coming. It also needed more memorythan was readily available in early Macintosh computers, so it failed.

In the Macintosh word processor market, Apple bundled MacWrite in1984 and Microsoft introduced Word in 1985. When competitors from Word-Perfect and an independent company WriteNow entered in 1988, all started

12See Figure 8.8 of [9].13See Figure 8.16, [9].

17

8/2/2019 A Run Del

20/23

from market shares of 20% to 30%. By 1991, Microsoft held over 60%, and

the others were under 20%.14

4 Discussion

Our purpose in this paper was to investigate situations where first moveradvantage is likely to be large and hence to motivate the manager to exercisereal options early. If strong first-mover advantage is ubiquitous, then per-haps a careful real options analysis is not really needed. Managers shouldadopt any positive NPV project that they find in order to preempt thecompetition and achieve the first mover advantage. However, as we haveexamined a variety of product markets, we have found reason to be careful

about the power of first mover advantage. In contrast, it is very common tofind the second mover taking over the market after learning from the actionsof the first mover.

A useful exercise for the reader is to try to find examples where the firstmover in a new market actually was able to sustain a long-term competitiveadvantage that created shareholder value. One of the best examples of astrong first mover advantage may be General Electric. Its founder, ThomasAlva Edison, developed the light bulb and the electric generation industry.General Electric continues to dominate these markets today. But even withthis example, Edison was forced out of General Electric before he couldcapture much value for himself and his heirs.

On the other hand, there are many examples of innovators developingstrong markets but spending so much time protecting the original marketthat they missed the evolution to a competing market. Xerox and Polaroidreadily come to mind. Both companies are currently in distressed conditiondespite creating and dominating their respective markets for decades.

We never mean to imply that first-mover advantages dont exist or arenot important for strategy. We mean instead to illustrate that there are cer-tain constraints on first-mover advantages, and that these should be carefullyweighed against successful imitator strategies that allow an option for delay.

5 Summary and Conclusions

With significant advantages to imitation, firms should be wary of optimisticassessments of first mover advantages, particularly in the face the loss of the

14The data for 1988 and subsequently are from Figure 8.17, [9], while the discussion ofthe earlier market is from the one of the authors personal experience.

18

8/2/2019 A Run Del

21/23

option value of waiting. Reluctance to delay because of concern over the loss

of first-mover advantages can be evaluated directly against the estimatedvalue of the option. While formal models of strategic interaction can bequite complex, and require careful applications of game theoretic reasoning,we advocate a simpler approach.

First, estimate the value of the option to delay. This depends on thecurrently expected value of the project and its current costs, the volatilityof the underlying business opportunity, and the expected time over whichthis option retains its value (ignoring for the moment the threat of entry).There is an extensive literature on the practicalities of real option valuationwhere the methods assume a monopoly position in the opportunity.

Next, evaluate the likely impact of the entrant on the stand-alone value

of the project. In some cases, this can be done with a degree of precision.For example, in the oil industry, the cost of delay where a competitor beginsto drain an oil field in which both parties have development rights can beestimated based on the price of oil and the rate at which the reservoir isdrained. In durable goods markets, it is the same principle but substitutethe market share of the competitor for the rate at which the reservoir isdrained. Based on our reasoning above, estimates of reduced follower costsshould also be added into the cost function in determining the right strikeprice. If costs are falling over time because of the benefits of moving second,the optimal adoption hurdle price as a follower may be quite different fromthe hurdle price as a first mover. These can easily be adapted to the realoption valuation models in the literature. It is also possible to introduce

the threat of a new entrant into a real options valuation model. Based onthe rationale provided above, it seems clear that if the cost savings to beinga follower are significant, it will also be important to model the benefit todelay that derives from implementing an imitation strategy. If imitationstrategies are a significant benefit, then second-mover advantages inducegreater delay.

Finally, compare an estimate of first-mover advantages to the loss in theoption to delay. If first-mover advantages determine that entry must occur,then aggressively pursue the market. However, weigh this carefully againstthe more conservative strategy that delays the risky investment. After allweve considered, those first-mover advantages may not be that important

after all.

19

8/2/2019 A Run Del

22/23

References

[1] James Brickley, Clifford Smith, and Jerold Zimmerman. An introduc-tion to game thoery and business strategy. Journal of Applied CorporateFinance, 13:8498, 2000.

[2] Joseph Farrell and Garth Saloner. Competition, compatibility and stan-dards: The economics of horses, penguins and lemmings. In H. Lan-dis Gabel, editor, Product Standardization and Competitive Strategy,number 11 in Advanced Series in Management, chapter 1, pages 123.North-Holland, New York, NY, 1987.

[3] Steven Grenadier. The strategic exercise of options: Development cas-

cades and overbuilding in real estate markets. Journal of Finance,51:16531679, 1996.

[4] Steven Grenadier. Option exercise games: The intersection of real op-tions and game theory. Journal of Applied Cororate Finance, 13:99106,2000.

[5] Bart Lambrecht and William Perraudin. Real options and preemption.JIMS Working Papers, 1997.

[6] Bart Lambrecht and William Perraudin. Strategic sequential invest-ments and sleeping patents, chapter 15, pages 297323. Oxford Univer-sity Press, 2000.

[7] James Lardner. Fast Forward: Hollywood, the Japanese, and the On-slaught of the VCR. WW Norton, New York, 1987.

[8] Richard C. Levin, Alvin K. Klevorick, Richard Nelson, and Sydney G.Winter. Appropriating the returns from industrial research and devel-opment. Brookings Papers on Economic Activity, 3:783820, 1987.

[9] Stan J. Liebowitz and Stehen E. Margolis. Winners, Losers & Mi-crosoft: Competition and Antitrust in high technology. The IndependentInstitute, Oakland, 1999.

[10] Edwin Mansfield, Mark Schwartz, and Samuel Wagner. Imitation costsand patents: An empirical study. The Economic Journal, 91:p 90718,December 1981.

[11] Sangin Park. Competition and organization in technology-intensiveindustries. In NBER Universities Research Conference, 1997.

20

8/2/2019 A Run Del

23/23

[12] Sangin Park. Quantitative analysis of network externalities in compet-

ing technologies: The VCR case. SUNY Stony Brook, 2001.

[13] Steven P. Schnaars. Managing imitation strategies: How later entrantsseize markets from pioneers. The Free Press, New York, 1994.

[14] Helen Weeds and Robin Mason. Networks, options and preemption.Real Options Conference, Cambridge, UK, 2000.

[15] David Wessel. Microcomputer software field is narrowing to threeLotus, Microsoft, Ashton-Tate, once niche players, go head to head.Wall Street Journal, Feb. 12 1987.

21