1 Introduction

Sustainable pioneering advantages are hard to spot in service industries, in contrast to manufacturing industries (Song et al. 1999) where empirical research has revealed not only that order of entry advantages exist but also how they decline over time (Kalyanaram et al. 1995). It is generally accepted that innovations are difficult to protect and competitive actions are easy to identify and imitate in service industries, making it hard to hold onto first-mover advantages (Kerin et al. 1992).

While services have characteristics in common, however, they also differ in many ways (Bharadwaj et al. 1993). This may work for or against the appearance and consolidation of different sources of pioneer advantages, such as technological leadership or pre-emption, along with the creation of buyer-switching costs (Lieberman and Montgomery 1988).

The object of our research is the study of first-mover advantages in the mobile telecommunications services industry. This is a regulated industry in which the licence is the scarcest resource; once a licence has been obtained, however, all operators have access to the same technology. It is also an industry where it is extremely easy to copy competitors’ actions (new services, lower prices, etc.). Additionally, the pioneer firm faces a series of risks, such as technological and commercial uncertainty and the need to create the market, that follower firms do not (Lieberman and Montgomery 1988). All this casts doubt on the possibility of establishing sustainable pioneering advantages. On the plus side, though, switching costs may provide the pioneer in the mobile telecommunications market with an advantage that is difficult for followers to cancel out.

In the last decade the mobile telecommunications industry has undergone huge changes both in the services that it offers (data transmission and other services related to the new economy) and the market’s competitive conditions. While the analogue technology markets in most of Europe were monopolies, with digital technology all the European governments opened their markets to competition. This has allowed the mobile telecommunications industry to grow exponentially in Europe, jumping from 6 million users at the end of 1992 (with analogue technology) to more than 400 million by the end of 2005 (with digital technologies).

And yet different models of development of the mobile telecommunications market co-exist in Europe. For example, in some cases two licences were initially granted, while in others only one was granted. After the initial concessions, the authorities in the different countries went on to grant further licences at a faster or slower pace. In addition, only some pioneers had prior experience with analogue technology as the exploitation of this technology had typically been granted as a monopoly. Lastly, number portability was not permitted in any country until 1999.

This may all have helped or hindered the securing of first-mover advantage. It has led us to set out a double objective for this study: (1) To look for the existence of advantages linked to order of entry and time in market in a service industry, as occurs in manufacturing industries; and (2) To study the impact of the main sources of pioneering advantages. As the mobile telecommunications industry is characterised by the accessibility of the technology to all the operators, we shall analyse the impact of pioneer, prior experience as a source of technological leadership. Other characteristics of the industry prompt us to examine switching costs—number portability—and to attempt to estimate the effect of pre-empting scarce resources by considering the number of pioneers.

The empirical study was performed using a sample that includes data on all the mobile telecommunications operators in the European Union (EU), Norway and Switzerland from 1993 to 2005.

The paper is organised as follows. First, we shall provide some background information on the industry and discuss the advantages that a pioneer in the European mobile telecommunications industry may enjoy. We shall then present our hypotheses, along with the supporting theoretical arguments. The next section contains the empirical study. The paper finishes with a discussion of the findings and some general conclusions.

2 The mobile telecommunications service industry

The mobile telecommunications services industry emerged in the early 1980s with the arrival of analogue technology. It did not begin to grow strongly, however, until the introduction of digital technology in the 1990s. Thus, in 1993 the average market penetration in the European Union (EU), Norway and Switzerland was 7.5%, while by the end of 2005 it was 104% (see Fig. 1).

Fig 1
figure 1

Penetration of mobile telephone services in Europe (%)

Full size image

The industry has gone through three phases, each one defined by the technology used. The first began in the 1980s with the commercialization of analogue (or first generation) systems. As the analogue networks of the different operators were incompatible, it was not possible to connect networks within or between countries. During this phase only the state monopolies operating in the conventional wireline industry received licences (with the exceptions of the United Kingdom, Sweden and France, where two licences were granted).Footnote 1

The second phase began in early 1993 when the first digital (or second generation) systems—GSM-900 and then DCS-1800—started to be used in Europe. The main advantage they brought was compatibility among the different operators of the EU, Norway and Switzerland. This meant it was now possible to end a call in a different operator’s network from the one in which it had begun. All European countries decided to open their digital markets to competition.

In some countries only one licence was initially granted, with other operating licences gradually being added. In others, though, competition was allowed right from the beginning by granting two licences, with more added later. In these cases, however, one of the licences was always granted to the incumbent firm operating with analogue technology.

The concession of licences for UMTS (or third generation) digital systems began in 2001. This technology uses a wider bandwidth, which makes it possible to offer a greater range of services such as MP3 sound and MP4 video file downloads and high-speed Internet access. This network is also compatible worldwide. In this case, all European countries opened their markets to competition from day 1 by granting between two and six operating licences. Not all the operators, however, began to exploit their concessions at the same time.

Figure 2 shows the evolution of pioneer and follower firms’ market shares between 1993 and 2005. The figure clearly shows that the pioneer firms were still hanging onto more than half of the market share at the end of the period under study. A comparative analysis (using the mean comparison test) of pioneer and follower market share for each country in the last period of the sample indicates that the market share of pioneers—on average—is significantly higher than that of the follower firms (see Table 1). In summary, then, a priori it appears that first movers in the mobile telecommunications industry may enjoy an advantage. We shall now go on to analyse these advantages in more detail.

Fig. 2
figure 2

Erosion of first-mover advantage

Full size image
Table 1 First-mover advantage in 2005
Full size table

3 Pioneer advantages in the mobile telecommunications services industry

First-mover advantage and their maintenance over time have been studied closely in the last 20 years (VanderWerf and Mahon 1997; Lieberman and Montgomery 1998). In general, though not absolutely conclusively, this empirical research supports the existence of pioneering advantages in industrial and consumer industries (Szymanski et al. 1995) from three types of sources: pre-emption of scarce assets, technological leadership and switching costs (Lieberman and Montgomery 1988).

The characteristics of the mobile telecommunications industry also seem to favour the appearance of some of these sources of pioneering advantages, as Table 2 shows.

Table 2 First-mover advantages in the mobile telecommunications services industry
Full size table

Pioneers can pre-empt scarce resources—such as favourable physical locations, inputs, plants and equipment or distribution and service systems—needed to develop the business. Thus, in the mobile telecommunications industry the location of base stations poses problems for firms that enter the market later as there are fewer sites available, along with opposition from local residents to placing them in towns and cities. Likewise, although scale economies are not important in most service industries because of the decentralisation of the service production process, they are significant in equipment-based services—such as mobile telecommunications—and may even be a source of competitive advantage (Bharadwaj et al. 1993). Pre-emption, then, not only makes it difficult for new competitors to consolidate their positions, it also enables pioneers to lower their costs and target the customers they want to reach.

Technological leadership is gained by experience economies and R&D superiority. In this case, however, all operators, regardless of when they enter the market, have the same technology. Pioneers, though, can take better advantage of experience economies as they have more time to accumulate output and lower unit costs.

Switching costs serve to lock in customers. The chief ones in this industry are non-portability of numbers and network economies.Footnote 2 In practice, non-portability of numbers keeps the customer captive due to the inconvenience of having to change telephone number when switching to another operator. As mobile networks are highly compatible it is likely that network effects will be caused by price discrimination inside and outside the network.

4 Hypotheses

4.1 Order of entry

As more competitors enter and the market stabilises there will be few opportunities for new entrants (Carpenter and Nakamoto 1989; Robinson et al. 1994). First, late entrants will find it more difficult to install their own infrastructure if they need scarce resources such as locations for base stations. Second, pioneers are positioned better to build a wide customer base. The heterogeneous nature of the services may give successive entrants the opportunity to find niches that satisfy the needs of particular groups of customers, but the opportunities presented by specific segments are likely to evaporate as new competitors enter. All this reduces late entrants’ chances of gaining a foothold in the market, as empirical research has found (Urban et al. 1986; Kalyanaram and Urban 1992; Makadok 1998). It leads us to put forward the following hypothesis:

Hypothesis 1:

The order of entry of follower firms negatively affects the disappearance of first-mover advantage.

4.2 Time in market

Along with order of entry, we must take into account the impact of lead time—the time between entries—on the erosion of first-mover advantage (Robinson et al. 1994). If the first follower firm enters the market quickly, the pioneer will have fewer opportunities to develop and exploit the first-mover advantage (Shepherd et al. 2000). Indeed, there is empirical evidence of some quick second entrants outperforming pioneers (Conner 1988; Smith et al. 1992).

In addition, while order of entry produces an effect that is conceptually constant in time (Brown and Lattin 1994), first-mover advantages seem likely to decrease as the number of years in competition increases (Robinson and Fornell 1985; Scherer and Ross, 1990). Empirical research does show that pioneer advantages gradually diminish as followers are given time to develop new resources (Brown and Lattin 1994; Huff and Robinson 1994; Kalyanaram et al. 1995). Our next hypothesis follows on from this situation:

Hypothesis 2:

The longer a follower firm has been operating in a market compared to the pioneer, the smaller the first-mover advantage will be.

4.3 Number of pioneers

First-mover advantage has traditionally been studied when there is only one pioneer. There are cases, however, where two or more firms compete from the beginning; when this occurs the competitive possibilities of followers are likely to be affected. One reason for this is the pre-emption of scarce resources that pioneers can exert. Although, a single pioneer is able to take better advantage of scale economies and thereby consolidate its position against future entrants, two pioneers may be able to control the complementary resources needed for the business. This would give them a better chance of grabbing the lion’s share of the market and locking in customers, and consequently make it tougher for new entrants to establish themselves. We therefore put forward the following hypothesis:

Hypothesis 3:

It is easier for followers to erode first-mover advantage when there is a single pioneer firm than when there are two pioneer firms.

4.4 Experience of the pioneer

There is a big difference between entering a market as a start-up or as an incumbent firm. Firms with prior experience can apply some of the resources they already possess to the new market (Mitchell 1991; Shamsie et al. 2004). Having already worked with a previous technology provides an experience effect that can easily be transferred to the new market, along with core knowledge of the business. The stock of sales and marketing knowledge built up over the years can be particularly crucial for the commercialization of an equivalent product (Cho et al. 1998). Klepper and Simons (2000) have shown how prior experience in the radio industry had a profound effect on the entry and performance of the firms producing television receivers. Pioneers with experience of a previous technology may also already have a recognised brand. This is a vital resource for building competitive advantages in service industries, especially when dealing with experience goods (when consumers have no advance knowledge about their quality) or services with a high degree of intangibility (Bharadwaj et al. 1993). All these resources help to consolidate the position with the new technology. Therefore:

Hypothesis 4:

In markets where two firms competed from the beginning, it is more difficult for the followers to erode the advantage of the first mover with prior experience of the industry.

4.5 Number portability

The lack of portability of a mobile telephone number is one of the main sources of switching costs in the mobile telecommunications industry. After analysing a series of regulatory variables, Grzybowski (2005) concluded that portability increases competition and lowers prices. Non-portability has hampered late-entering operators in their efforts to take customers away from pioneers, regardless of the attractiveness of the deals being offered. Until 1999 it was not possible to change operator and keep the same telephone number in any European country. Portability, then, is likely to have increased competition and contributed to eroding pioneer market share, which leads us to our final hypothesis:

Hypothesis 5:

It is easier for followers to erode first-mover advantage when number portability is allowed.

5 Methodology

5.1 Data

The sample comprises the 61 firms (24 pioneers and 37 followers) that provided mobile telecommunications services in the EU, Norway and Switzerland. Footnote 3 The data were collected in quarterly periods from January 1993 (when firms began commercializing second generation digital technology: GSM-900 and DCS-1800) until December 2005 (when some firms were already commercializing UMTS technology).

The first year of the sample was 1993 because this was when the existing monopolies with analogue technology started to be eliminated and the commercialization of GSM technology began. In the countries where there was a monopoly period with GSM technology, however, data were collected from the moment the market was opened to competition.

Information on each operator’s market share, date of entry, portability, type of licence, etc. was obtained from the trade journal Mobile Communications,Footnote 4 the GSM World website (http://www.gsmworld.com), the reports and websites of the operators and the websites of the different regulatory bodies.

5.2 Variables

The dependent variable in this study—Erosion—is the ratio of the market share of the follower firm “i” (second, third, fourth…) to that of the first mover (Urban et al. 1986, Huff and Robinson 1994, Brown and Lattin 1994). This variable makes it possible to measure how much each new firm has eroded the advantage of the first mover. Thus:

$$ \begin{aligned}{} & {\text{Erosion}}_{{itm}} = \frac{{{\text{Market}}\,{\text{share}}_{{itm}} }} {{{\text{Market}}\,{\text{share}}_{{1tm}} }},\; \\ & {\text{where}}\quad ''i'' = 1,2,3.......\;I\;{\text{follower}}\,{\text{firms}} \\ & \quad \quad \quad \,\,''t'' = 1,2,3.......\;T\;{\text{periods}} \\ & \quad \quad \quad \,\,''m'' = 1,2,3.......M\;{\text{countries}} \\ \end{aligned} $$

In countries with two pioneer firms, the market share of the first mover is calculated by adding the two firms’ market shares together. This is appropriate as this study aims to analyse how followers take market share away from firms that were present in the market from the beginning.

The first independent variable we consider is Order of entry itm (second, third, fourth…), reflecting in what position the follower firm “i” began to operate in country “m”. Time in market itm / 1tm measures how long the follower firm “i” has been operating compared to the pioneer in its country. We calculate this variable as the ratio between the time periods the follower firm “i” has been in the industry to those of the pioneer.

No. Pioneers m , Experience1m and Portability tm are dichotomous variables. The first takes value 1 if there were two pioneers in country “m”; otherwise, it takes value 0. The second reflects whether the pioneer firm “j” in country “m” had previously worked in the industry with analogue technology (value 1) or not (value 0). Lastly, Portability tm takes value 1 if country “m” in period “t” allowed portability; otherwise, it takes value 0.

Different studies have pointed out the importance of two structural variables—growth of the industry and concentration—in the erosion of first-mover advantage (Makadok 1998; Pan et al. 1999; Shamsie et al. 2004). Consequently, we include them in the analysis to control for potential effects on the erosion of the pioneer’s market share. Growth of industry (GrowthInd tm ) measures the variation in the number of customers in country “m” between the periods “t” and “t − 1”. Concentration tm is calculated as the sum squared of the competitors’ market shares in a specific market and period.

5.3 Model specification and estimation

To empirically estimate the effect of order of entry on the erosion of the pioneer’s advantage, we begin with a base model of the form:

$$ {\text{Erosion}}_{{itm}} {\mathbf{ = }}{\text{Order of entry}}^{{{\text{ $ \beta $ 1}}}}_{{itm}} {\mathbf{ + }}{\text{Time in market}}^{{{\text{ $ \beta $ 2}}}}_{{itm{\mathbf{/}}{\text{1}}tm}} {\mathbf{ + }}{\text{GrowthInd}}^{{{\text{ $ \beta $ 3}}}}_{{tm}} {\mathbf{ + }}{\text{Concentration}}^{{{\text{ $ \beta $ 4}}}}_{{tm}} {\mathbf{ + }}\upvarepsilon_{{itm}} {\text{ (where $ \beta $ 1 to $ \beta $ 4 are coefficient parameters)}} $$
(1)

The form of the model is non-linear to reflect the hypothesis that the impact of the second firm to enter will be greater than the third or fourth (Urban et al. 1986; Brown and Lattin 1994; Huff and Robinson 1994). For model estimation, Eq. (2) is a log transform of Eq. (1) with primes denoting neperian logarithms:Footnote 5

$$ {\text{Erosion}}^\prime _{{itm}} {\text{ = $ \beta $ }}_{{\text{1}}} {\text{ $ \times $ Order of entry}}^\prime _{{itm}} {\text{ + $ \beta $ }}_{{\text{2}}} {\text{ $ \times $ Time in market}}^\prime _{{itm/1tm}} {\text{ + $ \beta $ }}_{{\text{3}}} {\text{ $ \times $ GrowthInd}}^\prime _{{tm}} {\text{ + $ \beta $ }}_{{\text{4}}} {\text{ $ \times $ Concentration}}^\prime _{{tm}} {\text{ + e}}_{{itm}}^\prime $$
(2)

In addition to including variables that measure how different factors affect the erosion of the pioneer’s advantage, we tested two other models similar to the one specified in Eq. (1). Model 2 includes the dichotomous variable No. Pioneers m , while Model 3 incorporates the variable Experience1m . The estimate in Model 3 is only made for countries where there was no monopoly situation. To calculate the variable Erosion itm in Model 3, we analyse each pioneer’s market share individually as we want to compare the impact of experience on the erosion of each pioneer’s competitive advantage. The impact of Portability tm is captured in Model 4.Footnote 6

Analysis of longitudinal data is subject to violation of standard OLS assumptions, primarily those concerning homoskedasticity and auto-correlation. Such violations may introduce bias into estimates, including inflated F-statistics (Bergh and Holbein 1997). It is necessary, therefore, to correct the heteroskedasticity caused by the cross-section (Sayrs 1989) and the potential auto-correlation due to the number of time periods included in the sample.

Cook-Weisberg and Wooldridge tests indicated the presence of heteroskedasticity and serial correlation, respectively, when using models of ordinary least squares. To avoid both problems cross-sectional time-series linear models using feasible generalised least squares were employed. These models allow estimation in the presence of AR(1) auto-correlation within panels and specify the heteroskedastic error structure, thus ensuring that the standard errors are robust.

We also checked the level of correlation between the independent variables to detect any serious multi-collinearity problems. In general, correlations were moderate and did not appear to violate the assumption of independence between explanatory variables (see Table 3). In addition, none of the VIF (variance inflation factors) for the variables in the three models exceeded two. Bearing in mind the rule forbidding these values to be above ten (Nester et al. 1985; Chaterjee and Price 1991), we can conclude that multi-collinearity is not a problem in this study.

Table 3 Descriptive statistics and matrix correlation for the sample
Full size table

6 Results

Table 4 displays the regression coefficients of the four models testing the hypotheses on the erosion of pioneer market share. All the models are significant at one per cent level.

Table 4 Empirical model of the erosion of the first mover’s competitive advantage
Full size table

Model 1 tested the effects of order of entry and time in competition on the erosion of pioneer firms’ advantages. The coefficient Order of entry itm is negative and significant (β = −0.2324, P < 0.001). Late entrants, therefore, find it more difficult to seize market share from pioneers than early entrants. Hypothesis 1, then, is confirmed. The coefficient Time in market itm/1tm , measuring the relationship between the time a follower firm has been operating compared to the pioneer, is positive and highly significant (β = 0.9525, P < 0.001). This finding strongly supports the prediction of Hypothesis 2 that the longer a follower firm has been operating in a market compared to the pioneer, the smaller the first-mover advantage will be.

Models 2 and 3 examined the impact of the number and experience of pioneers, respectively, on the erosion of first-mover advantage. The results of Model 2 support (β = 0.2062, P < 0.001) the prediction of Hypothesis 3 that follower firms will find it easier to erode first-mover advantage when there is a single pioneer firm than when there are two pioneers. Model 3 shows (β = −0.1380, P < 0.001) that in countries where there had been no monopoly—and consequently where two firms had been in the market from the beginning—it was more difficult to take market share from pioneers with experience of an earlier technology. This supports Hypothesis 4.

The coefficient Portability tm in Model 4 provides partial support (β = −0.0059, P < 0.1) for the prediction of Hypothesis 5 that it will be easier for follower firms to prize market share from pioneers when users can switch operators without having to change their telephone number.

Of the control variables, the coefficient Growth of industry tm is positive and partially significant (β = 0.0028, P < 0.095), indicating that the erosion of first-mover advantage is greater in periods when the market has grown more. The coefficient Concentration tm is negative and significant (β = −0.8683, P < 0.001), showing that it is harder to take market share from pioneers in more concentrated markets.

7 Discussion and conclusions

At first sight it seems unlikely that there are pioneering advantages in service industries in contrast to the situation in manufacturing industries, where empirical research has been focused. Not all service industries, however, are exactly the same. Our work specifically looks at the mobile telecommunications industry, where there appears to be factors that work both for and against pioneering advantages. The operators, for example, have equal access to digital technology and can all quickly copy and incorporate the services offered by competitors. This situation would seem to make it more difficult to sustain any first-mover advantage. In addition, pioneers have had to face commercial and technological risks, along with using resources to develop the market, that have allowed follower firms to act as free riders. On the other hand, pioneers can benefit from a range of advantages based on switching costs and the pre-emption of scarce resources (e.g., the location of base stations).

Our empirical research with data on European firms from 1993 to 2005 indicates that it seems premature to assume that order of entry does not provide any type of advantage in service industries such as mobile telecommunications. The findings are similar to those from previous research on manufacturing industries and some studies on financial services (Tufano 1989; Makadok 1998; López and Roberts 2002). There is an advantage based on order of entry, but it is an advantage that diminishes with the time in competition. Thus, the erosion of pioneer market share increases in relation to the time follower firms have been in the market compared to the pioneers.

Just as not all service industries are identical, not all pioneers are in the same position. Thus, we found that their advantage is more sustainable when there are two pioneers instead of one. In other words, followers have found it harder to make headway in markets where two licences were granted from the beginning than in markets where there was a single pioneer. It is likely that two pioneers are able to pre-empt scarce resources more efficiently than one, which makes it harder for followers to acquire the assets they need. Of course, we must also bear in mind that many consumers are reluctant to purchase a product when there is only one supplier (Porter 1985). Having two suppliers may give consumers more confidence to buy, which will in turn allow the market to grow more quickly. This will benefit the firms present from the beginning and make it tougher for new entrants to establish themselves. Paradoxically, then, a measure that was supposed to develop the market and create competition has strengthened pioneer advantage and made it more difficult for new operators to find their feet.

Although, the easy accessibility of the technology in this industry seems to block the securing of first-mover advantages based on technological leadership, not all pioneers are equal; some had worked with a previous technology. In fact, pioneers with experience of analogue technology stood up to the entry of new operators better. Despite the small number of customers they had captured with the previous technology, the experience will have schooled them in the market and the technology. This may have enabled them to learn from their earlier mistakes and given them an initial edge over firms that were total newcomers to the market. Of course, as these pioneers were all previously state-run monopolies they may also have had deeper pockets than their competitors and thus found it easier to establish themselves.

The dual clocks model (Mitchell 1991) recommends that operators with prior experience wait before entering the market, yet in no case did this happen. This model suggests that incumbents that want to move into new technical areas should wait for others to test the market for them, as their performance in the new area will depend mainly on their order of entry relative to other incumbents, not to all other entrants. In our sample, however, incumbents entered the market from the beginning and this helped them to strengthen their position against newcomers.

Lastly, we have studied the impact of one of the most important switching costs: number portability. As expected, portability seems to favour the erosion of pioneer advantage as it permits customers to choose the best deal more easily. We were not able to directly examine the effect of the network economies that are likely to be caused by price discrimination (given the compatibility of digital networks) One would expect the operators, however, to have introduced different pricing policies once portability was allowed, as up until that moment non-portability effectively locked in customers.

Apart from studying pioneering advantages in services, this paper takes a fresh approach that attempts to fill in some of the gaps in the literature on first-mover advantages. The vast majority of empirical research during the last 20 years has focused on firms in the US, with no attention given to finding out if the results are generalizable to other countries (Lieberman and Montgomery 1998; Song et al. 1999). Our sample of European firms will help shed light on the question of generalizability by permitting a comparison of the findings on US firms with European firms.

Another limitation of existing empirical research has been the choice of firms for analysis. First, the inclusion of only large firms in samples and the exclusion of pioneers that disappeared from the market have resulted in pioneering advantages being overestimated. Second, identifying pioneer firms has been problematic, as firms have often been responsible for identifying themselves as pioneers or followers. This introduces a bias in the sample by including a high percentage of pioneer firms (Lambkin 1992). This study avoids these two bias effects by incorporating all the operators in our sample and determining order of entry by when the regulatory body in each country granted the licences.

This paper is not free from limitations. First, it would have been advisable to replicate the study with another measure of results. Market share, in fact, has been criticised as a measure because it has been found to inflate the chances of finding pioneering advantages (VanderWerf and Mahon 1997). Using financial measures of results in a multi-country study is a non-starter, however, as they are calculated differently in each country. And during the period under study, the mobile telecommunications businesses were not quoted independently, which prevented using stock market measures. In any case, a report by CapGemini (2004) indicates that the industry itself seems to use market share as a yardstick of competitiveness. Based on data from 27 European operators, the report shows that firms concentrate their efforts on snatching market share from competitors. Second, it would be helpful to have more information on the operators. Knowing what specific resources pioneers and followers have, for example, along with details of their marketing efforts and competitive strategies would be useful.

This paper’s main practical conclusion is to confirm that it is possible to build first-mover advantages in a service industry. On average, more than 70% of employment depends on services in European countries. Every day we discover new business models with advantages that competitors are not able to erode quickly, even in mature industries. It would seem unwise, therefore, to presume that there are no sustainable advantages in these critical industries. This field requires further research both in the analysis of different types of advantage and the ways they are eroded.