Chapter 2

Setting the Scope of Analysis

Defining Your Industry

Introduction

In this chapter you’ll develop techniques for generating a thoughtful definition of the industry in which your firm competes. The idea of defining an industry and its boundaries may seem odd or maybe even pointless. After all, newspapers and television newscasters and politicians and all the rest of us are quite comfortable talking about the pharmaceutical industry or the auto industry or the air carrier industry. What needs to be defined?

In fact, though, there is no such thing as an industry other than what we think it is! Defining an industry is a conscious selection of a set of firms from a larger environment, so if we each have a different mental set that includes some firms and excludes other firms, then we have different definitions of what the industry really is. This will affect every subsequent step of the analyses you’ll do in this book. For example, the U.S. air carrier industry might be defined as the largest carriers, such as Delta, American, United, US Airways, Continental, Southwest, and JetBlue. This list has the advantage of being short and capturing the largest of the competitors. Another approach would use the membership list of an industry association like the Air Transport Association. This is a broader list but the members are self-selected, which raises concerns about who might not have elected to join hence the inclusiveness of the list. The list includes firms of very different sizes and also includes UPS and FedEx as members. If you are interested in airfreight, this makes sense, but if you are interested in the firms that fly people, this is not an intrinsically good set. We might turn to the Bureau of Transportation Statistics, which reports on all passenger carriers with more than $20 million in annual revenue (of which there are 16 for 2009). The complete list of passenger carriers runs to 52. Deciding which is the right list (or, indeed, if another list is better) depends on the answers you derive to questions about the problems buyers are trying to solve when they make a purchase, the products or services that meet those buyers needs, and the firms that can provide the products or services.

So, why should you develop an industry definition? Given what you just read, it is possible to configure the carrier industry in many ways. One of the most common problems I’ve seen managers have as they learn this process is letting the scope of the industry change over the course of the analysis, which renders any conclusions suspect. Committing to one definition of industry at the outset drives a stake in the ground and keeps your analysis consistent. A second reason to develop an industry definition is because the questions you’ll answer in developing a definition are used repeatedly in subsequent stages of the analysis.

Problems in Defining an Industry

In economics, an industry is often defined as the set of firms that provide close substitutes. In real life, managers often generate an industry definition (i.e., name the set of competitors) based on the competitive pressures they face at the time.¹ That is, what and for whom a firm produces now is the basis for determining how its managers will define competitors. For example, managers tend to classify firms as close competitors if they are similar in form and if they share and compete for the same scarce resources, such as customers and suppliers. Typically, this is a subset of a larger set of firms that serve related/similar/connected markets. Managers can also make mistakes in populating their industry definition by focusing on the largest and best known firms while ignoring new or small or evolving firms.² These classifications are considered faulty or inferior because of a bias in judgment that comes about because search costs (i.e., really digging into the definition) are viewed as high. Perceived similarities with other firms make evaluation and assessment relatively easy and cost effective (in the sense that similar firms should respond or act similarly). Also, because industries tend to generate and participate in a commonly held knowledge trajectory,³ firms may regard as competitors those firms that also share in or generate industry level knowledge in pursuit of solving the economically interesting problem. That is, these are firms that know and value the same things.

The rise of Wal-Mart serves as an example of the cognitive problems rivals might have had in deciding if Wal-Mart was really a competitor (at least, early on!). In the discount-retail industry of the 1950s and ’60s, senior managers believed that a population base of 100,000 was needed to support the large scale, low-margin structure of the stores. Therefore, serious competitors sited stores in urban and suburban areas. However, Wal-Mart originated in small towns, serving broad rural areas that could tap 100,000 people who find themselves closer to a Wal-Mart store than a city-based discounter. That is, for firms in the discount retail industry, Wal-Mart might have differed too much in service area, if not in product or service (and, early on, was too small) to be considered as a competitor. Wal-Mart didn’t hit Kmart, Bradlees, Caldor, Shopko, or other discount retailers as an obvious or important threat until years too late—after Sam Walton was well on his way to solidifying the foundations of industry-leading IT and logistic competence.⁴

Similarly, mistakes occur when managers fail to account for evolutionary—or even revolutionary—changes in the competition or product/service. In the book-selling industry, for example, until the 1960s or ’70s, books were sold primarily through independent, local stores. These were typically small and urban. Moreover, they were usually had a narrow focus such as used books or were limited to a specific discipline (like law) or genre (such as mysteries). Managers then might define the relevant industry as bookstores reasonably close to their physical location. Then, in the ’60s, chain bookstores, such as B. Dalton, Waldenbooks, and Crown, emerged as mall-based, suburban stores. They were still small, convenience destinations (on the order of 5,000 sq. ft.), but because shopping patterns migrated away from urban to suburban and from private shops to malls, these firms drove many single stores out of business. Also, the scope of competition changed: The chain stores had national presence and, as we’ll see, some associated strengths.

In the mid-eighties, both Borders Books and Barnes and Noble began to expand retail operations with a new model—the bookstore as a superstore and a destination. These were like other specialty, category-killer stores in that they were large (30,000 sq. ft.), offered a wide selection often at lower prices, and were stand-alone rather than mall-based (or they worked as an anchor). They also came to offer amenities, such as coffee bars, comfortable seating, and a relaxed environment. These damaged and eventually absorbed the mall-based chains. Finally, in the mid-nineties, the emergence of the Internet made new, virtual firms, such as Amazon (and, now, Alibris, which is a sort of meta-Amazon) feasible.

Book retailers who failed to keep up with changes in shopping patterns and preferences, changes in the product or service being delivered or how it was delivered, or the competitors that emerged to take advantage of those changes would be in trouble. Changes in the shape of the industry change the dynamics of competition. So, how can managers develop and maintain a better sense of what the industry is?

A Systematic Approach to Industry Definition

Previously I mentioned that the standard economic definition of industry is the firms that produce close substitutes. In strategy, and particularly in industry analysis, the word substitutes has a very specific meaning that differs from the standard economic definition, so I recommend the following definition of industry:

An industry is the set of firms that solves the same economic problem in the same way.

How do we determine this set of firms? Clearly, just picking competitors without a rationale for inclusion or exclusion can yield very different outcomes. For example, when graduate students are asked to define the U.S. auto industry, several ad hoc approaches usually emerge. One is to name the firms headquartered in the United States (like GM, Ford, and Chrysler). The second is to focus on firms that manufacture in the United States (which adds Toyota, Honda, and BMW, among others). An extension adds all firms that sell cars in the United States. Finally, almost always, a definition pops up that includes not only firms that manufacture cars but also suppliers and firms involved in sales and financing (e.g., identifying seat or steering system manufacturers or dealerships as parts of the industry).

In fact, all are legitimate ways of starting the definition and the latter approach anticipates a key analytic tool: the extended value chain. Still, lumping suppliers and buyers in the auto industry is almost always too large and diffuse a group and focusing on only the largest firms may be shortsighted.

This is why we start with “solving” an “economic problem” and doing so “in the same way” across the industry firms. It helps you more completely define who the customers are (since they have the problem that needs or wants solving). It is also useful to be clear about how the problem is solved with respect to product or service characteristics and how those are produced. This approach can help develop stronger criteria for selection. More formally, the process requires the following actions:

• defining the key economic problem
  • Who are the ultimate consumers?
  • What is their problem or need?
  • How is the problem solved? (Product/service characteristics)
• naming the firms that solve this problem

We can use the U.S. auto industry as an initial example. To begin, we are interested in developing the boundaries of the industry or the firms we’ll include or exclude. As a first pass, we know the industry firms produce cars. Who are the ultimate consumers of cars (i.e., the product of this industry)? The answer is usually individuals usually individuals but also governments and rental fleets. Note that if it is not already specified in the problem, you will want to be clear about the scope of your consumer set. For instance, are you considering only the U.S. market or a more global one?

Second, what is the problem these consumers are trying to solve? This can be difficult to define because there are usually multiple facets; consumers are often trying to solve a number of interrelated problems simultaneously. It is very important and useful to understand these consumer preference profiles for several reasons. First, they will help develop the boundary of potential solutions. Second, they help identify potential (and powerful) substitutes. Third, they might lead you to customers who are not very satisfied with the industry solution but don’t have many better choices, and these can be valuable new markets.

In this case, at the most basic level, consumers of cars want convenient transportation—and convenience might be defined as transportation when and where needed (i.e., personal and private, regardless of weather). We should flesh this out to establish when and why consumers may prefer autos rather than public transportation or other modes (such as bicycles). After that, needs or desires within the solution set vary. For example, if I were in the market for a new vehicle, I would have a number of problems or preferences I would want to address and I would want to maximize my satisfaction (my utility) in doing so: I drive 20 miles one way to school, and since it snows six months of the year where I live, a vehicle with all-wheel drive would be a good idea. On the other hand, 20 miles of driving and high gas prices make efficiency a key concern. I also like to haul material for gardens or fishing; a vehicle that has hauling capacity is a good idea. I also like it to be comfortable, if not a little pampered, as I drive. Finally, I have a limited budget. This is a difficult set of preferences to satisfy (in one vehicle, anyway!) because they are internally not very consistent. For instance, all-wheel drive and fuel efficiency don’t usually go together very well. Nor, usually, does hauling capacity and efficiency or luxury. Obviously, I’ll have to weight some factors more heavily than others. Even if I do, though, changing external conditions can change how the factors are weighted. If gas prices spike or collapse, for example, the importance of efficiency will rise and fall.

What makes this interesting is that almost all consumers share most or all these concerns but to varying degrees, which is why industry offerings can be far more limited than the number of customers. Preferences are not usually driven by a single need but by many needs, and they often shift over time or as conditions change. Preferences are typically multidimensional; it is key to get a grip on what the major elements may be.

Understanding what consumers are trying to solve with their choices helps us determine the firms that can meet those needs. In a general sense, firms in the auto industry meet consumer preferences in the same way by offering one version or another of a self-contained, gasoline fueled, internal combustion driven machine in about the same size. A Yukon is bigger than a Cobalt, but they are within an order of magnitude (factor of 10) in size. This differs from, say, passenger planes where the smallest may carry four or five and the largest may carry a hundred times that many. In other words, the firms in the auto industry have solved the primary need of personal, private, and convenient transportation in a particular fashion. Preferences are met by the range of products offered or option packages.

This definitional process is important because it gives us guidance about the sorts of solutions that fit and those that don’t. Are motorcycles equivalent solutions? Bicycles? Trucks? It would not be useful to include motorcycle-making firms as part of the industry because motorcycles solve the transportation problem in a different way, though not as differently as bicycles solve it. Still, there could be some crossover firms in this because Honda, Suzuki, and BMW make motorcycles and also make cars. We’ll have to include these firms but only as auto manufacturing firms. Trucks are even trickier. Probably most people would include light trucks as part of the solution set (Ford Ranger or F150) but how about larger trucks? Are trucks like the Ford F350 the same sort of solution? Are semitrucks, such as Peterbilts or Freightliners or Macks solutions (and their manufacturers part of the auto industry)? Probably not, but firms like Volvo and Daimler do participate in that large-truck industry. Again, we’ll have to discriminate carefully.

Another part of understanding how the consumer problem is solved entails thinking about how industry products or services are actually produced and delivered. In some industries the same basic technologies are used by all competitors but in others they can range from fundamentally manual processes to highly automated and capital intensive (the furniture industry is a good example where you might include firms ranging from individual craft workers to the largest of the commodity manufacturers). This might make a difference in making sure the same sets of customers are being addressed. On the delivery side, consider the book retailer examples used earlier where products can come from both brick-and-mortar stores or from online retailers.

With needs identified and the product or service clarified, we can build a set of firms that comprises the industry. Table 2.1 lists the top 10 firms in global sales for the first half of 2009,⁵ but since we are interested in the U.S. market, this might not be completely useful, as Citroen and Renault are not sold in the United States. Moreover, the relative sizes of the rest of the firms differ at the national level versus the global level. Table 2.2 lists the top 10 auto manufacturers in U.S. sales for 2009 according to the website Ward’s Auto;⁶ note the differences.

Alternatively Edmunds.com, a well-known car pricing and review site, has its own list of auto labels⁷ (see Table 2.3), and Fiat, Peugeot, or Renault are not listed (though this may change with Fiat’s investment in Chrysler). The list has no rank ordering, which makes it less useful than a top 10 list, but it is more inclusive. On the other hand, this list treats brands as individual entities rather than as components of a larger firm (e.g., Toyota and Lexus or General Motors and its then subsidiaries Saab, Saturn, and Hummer). It will take more research to define firms.

This gives rise to another issue: How many firms are enough to constitute a good definition of the industry? Looking ahead, you’ll want your list to be tractable with members that have enough in common so that general environmental and industry analytics are workable. Too narrow a list means you run the risk of overlooking important small or emerging firms, but too broad a list makes managing the analysis much more complex. One approach is to keep the level of industry concentration in mind. Concentration refers to the extent to which industry revenues are controlled by just a few firms (very concentrated) or are broadly spread among many firms (not concentrated). A short list is more reasonable in highly concentrated industries. For information on how to calculate concentration, see the appendix to chapter 5.

A rule of thumb to finish the list would be to rank the firms by size (unit output or revenues), and include the firms that contribute to the top 80% of sales. If, at the end of the process, you find many firms of the same size, you might include them as well.

As an illustration, we can apply this technique to the brewing industry, which has exhibited significant changes in concentration over time. The list of the top 10 brewers in 1950 (shown in Table 2.4)⁸ indicates that the market was pretty evenly divided among these competitors. (Note that the largest brewer was barely twice the size of the seventh largest. Moreover, the top 10 accounted for 33.3% of the market.) An industry definition would surely have to include all these firms and perhaps even more. Contrast that list with the distribution of sales in 2005 (see Table 2.5).⁹ In the latter case, the industry is clearly dominated by three firms, which account for about 80% of market sales. Under some circumstances, this may be enough to use as a basis; though, note that if you were interested in the craft brewing segment of this industry, you would have a different set of firms to use, which would be led by many of the bottom six here.

As you consider candidate firms, you’ll likely conclude that some are clearly on the list, some are clearly not, and some are difficult to classify. As a starting point, use the “clearly in” firms for the first passes at analysis, but don’t forget the “maybe” firms. Later, you can add them to the analysis to see if they make any difference. The important concept to grasp is that just making a list without understanding the assumptions and limitations can skew and distort your list of industry members and damage your subsequent hard work.

There are a number of sources for information on what firms might constitute an industry. Obviously, for the auto industry, you’ve already seen data from Hoover’s, Inc., a reputable source of industry and business data. Similarly, Dun and Bradstreet and Standard and Poor have industry profiles available at most libraries and online. You can also turn to industry associations for information. Most industries have an organization that stands in as a formal and collective voice on industry issues to others. In the auto industry, for example, there are groups such as the Association of International Automobile Manufacturers (AIAM) or the European Automobile Manufacturers Association (ACEA) or the Automotive Industry Action Group (AIAG), among others. Industry associations are a great source of material on issues beyond membership, too, because there is almost always some discussion of key issues facing industry firms. The brewing industry has its own associations including the Beer Institute, which covers the industry in general, and the Brewers’ Association, which is more focused on the craft brew part of the industry. You might also search government websites, especially those of regulatory agencies. The auto industry, for example, is affected by work from the Department of Transportation (DOT) or the National Highway Traffic Safety Administration (NHTSA).

What is the right list of firms? That is up to you and the problem you are trying to analyze. In a political discussion, the U.S. auto industry might reasonably be construed as just the Big Three of Ford, GM, and Chrysler. As we’ve seen earlier, other lists vary with the question: U.S.-based firms, firms that manufacture in the United States, or firms that sell in the United States? The right list depends on what you want to accomplish. What makes it right is that it is consistent with what you want to analyze, it is well defined, and it is inclusive.

Summary

Defining the industry is a useful, if not essential, beginning to competitive analysis. It compels us to think about customers, their needs or problems, and how some firms have elected to meet those needs. Being thoughtful about definition forces us to decide which firms are in and which firms are out. As you go forward through the analytic process, you’ll likely find that your set of competitors (the industry) changes. That’s usually because some decisions you make here in definition were explicitly or implicitly revisited and reassessed. Just make your sure you are aware of what changed in your thinking and why.