CHAPTER 4

Information Economics

In the late 1990s, there was talk of a new economy in which old rules like supply and demand no longer applied; key aspects of the economy would be characterized by abundance, by the problems that come from having to allocate surplus.1 As we all know, whether with iron ore, rice, or water, those scenarios did not come to pass; “too much” is seldom the problem. Throughout the hype, traditionalists held the line: “Durable economic principles can guide you in today's frenetic business environment. Technology changes. Economic laws do not.”2 Information goods include encyclopedias, compact discs, cell phone minutes, or equity analysis, all of which are valued not by what they are but what they convey. The characteristics involved in information economics—particularly network effects, lock-in, and pricing behavior—underlie many of the phenomena in this book.

Very few private-sector companies have chief economists. Investment banks, publicly traded homebuilders, and large healthcare firms all might have need for such a position, but generally few companies hire such individuals. In 1998, two economists from the University of California at Berkeley published Information Rules,3 a neoclassical approach to such issues as technology lock-in (why does everyone still use the demonstrably inferior QWERTY keyboard?), information pricing, and standards wars. Unless otherwise noted, all foundation concepts and some examples noted here come from this source.

In 2002, one of the authors, Hal Varian, began consulting at Google. Shortly thereafter, he took a leave of absence (and eventually retired) from his academic post, moving full time to become Google's chief economist. In that position, Varian's tenure has coincided with the rapid growth of Google's breakthrough AdWords/AdSense pricing mechanism—which is based on complex auction mechanics—as well as the firm's foray into bidding on wireless spectrum.4 Google has begun to lobby world governments and enter often-complicated agreements with competitors and other entities (in regard to mobile phones or book publishing for two examples). At both the macro/policy and micro/strategy levels, information economics is sufficiently nonintuitive and, at Google's scale, so complex that having such expertise in-house appears to have been well worth the hire.

Information Goods

Goods whose primary value lies in the information they convey (rather than in their physical essence, as with wheat, or their functional use, such as a comb) can behave in unexpected ways. Classic examples of information goods include books, newspapers, stock tickers, and movies/DVDs. The peculiar properties of digital information goods combined with the nature of the Internet and related technologies have led to upheaval in some old industries. In brief, four behaviors of information goods have become relevant in music, news, software, financial analysis, and other sectors, often changing business models or economic viability in the process. These four behaviors are:

  1. Sampling
  2. First-copy costs
  3. Price based on value, not cost
  4. Network effects

Sampling

Information goods are experienced and thus need somehow to be sampled before deciding to buy.

The only way to know whether you like a song or movie is to “consume” the information before purchase. Trailers serve this purpose for movies; radio airplay was perfect sampling for recorded music. Third parties become important: Word of mouth, about which we will hear more later, and reviewers are important pieces of the puzzle. Compare news to a car, in which statistics, crash-test results, and other metrics can guide customer choices. Information goods, in other words, cannot easily be understood through standardized representations or objective measurements: Knowing how many actors (or jokes) were in Airplane tells the viewer nothing. Counting pages for a book is similarly worthless.

Reputation, whether conveyed by brand, word of mouth, or otherwise, helps overcome the sampling problem. People buy the Wall Street Journal a year at a time not because they looked at every day's edition before they bought it but because past performance suggests they will benefit from the kinds of information that will be delivered.

First-Copy Costs

Information goods are expensive to produce and very inexpensive to reproduce and, over the Internet, to distribute.

The shorthand way of saying this relates to commercial software: The first copy of Windows 7 cost billions of dollars in research and development costs, advertising, and testing. The second copy cost under a dollar if physical media are involved, more with logistics costs; for online downloads with essentially free distribution, that second-copy cost drops toward zero. With newspapers, it's a similar model, except that timely physical distribution at scale can be very expensive, particularly as the prices of electricity, ink, and diesel fuel for delivery trucks all typically rise with the cost of crude oil.

Price Based on Value, Not Cost

Because replication and distribution costs for the copies after the original are essentially zero online, the price of information goods is not based on cost but on the value attached to the good by individual customers.

Supply and demand curves multiply: At the extreme, there are as many optimal prices as there are people with value preferences. One important task, then, is to find ways to differentiate the product and capture as much of the market's price tolerance as possible. Economists call this price discrimination, and it is often illegal and usually unpopular. What businesses can do is to offer subtle variations on the basic offering. In addition, consumers have become accustomed to free being a viable price for many information goods.

Network Effects

Externalities (outcomes of trade not accounted for in the transaction) are frequently negative: Traffic and pollution can be used as examples. Network externalities, also called network effects,5 can also be positive: The behavior of someone else in the network makes me better off without my doing anything. Flu shots are a positive network externality: The more people in my office who get inoculations, the lower my risk of exposure in that sphere of my life. If someone else signs up with my cellular carrier, in-network calls and texts to that subscriber are often free. If more people buy Word after I do, the odds of being able to exchange documents are better the more people are on the platform.

Pricing Information: Versioning and Bundling

Unlike most examples in commercial history, free is a viable price point for digital information goods: Linux is a classic example. Subsidies are often common: Advertisers provide a revenue stream so newspaper readers, television viewers, and Google searchers can enjoy lower priced or no-cost media. Other revenue streams are possible: Some bands allow taping of concerts to increase fan loyalty or to encourage the sale of T-shirts, commercial-quality recordings, or future concert tickets. (See “Software Copying” later in the chapter.)

Pricing information goods to extract as much profit as possible from this cacophony of demand curves, and often in networked scenarios, is a fascinating exercise. In particular, the notion of “versioning” is very appealing. This practice relates to the ways that producers can reconfigure information goods so as to present different market segments with attractively priced versions of the same core product. Basic and deluxe versions of software are frequently very similar under the hood, but the “professional” features included in the platinum version allow the seller to extract premium prices from people who value the good more highly.

Movies are a convenient example. Depending on one's personal preferences, a customer can watch a movie in dozens of ways, each with its own set of costs and benefits, starting with time from original release:

  • Opening-night premiere in Hollywood: thousands of dollars if at all available
  • Opening week, New York, Friday evening: $20 per seat
  • Opening week, Topeka, afternoon matinee: $5 per seat
  • IMAX 3D version: $25 per seat
  • Second-run theater: $1 with coupon
  • Comcast pay-per-view: $5 for a roomful of viewers
  • Hotel pay-per-view: $13.95
  • Pirated copy of DVD in China: $0.25
  • Over-the-air free TV two years after release: commercial interruption
  • Netflix: physical DVD rental price varies by plan and activity level
  • Director's cut Blu-ray disc with commentary, bloopers, and other extras: $40

For movies, the multitude of distribution and delivery channels makes versioning extremely effective: Blockbuster hits like Avatar stay in theaters as long as they can command an audience while duds are pulled after a week or go straight to cable television and/or DVD. Few other offerings can exploit so much of an audience's willingness to spend, although until the ascendency of the Kindle, hardcover books also had longer runs before the paperback release if they sold well.

Versioning can be based on attributes of the good: Higher-quality (Blu-ray versus VHS), better documentation, extra content such as directors' commentaries on DVDs, access to relevant archives and data sets, real-time stock quotes versus 20-second delayed. Versioning can also create multiple pricing options based on characteristics of the buyer: student discounts, group pricing, or differential pricing by geography.

This practice can backfire, as when Lexicon, a high-end home theater company, repackaged an inexpensive Blu-ray player inside a fancier case and tried to charge a 600% premium.6 The attributes of a more expensive version (whether in a Lexus compared to a Toyota or in a software package or online dating service) need at once to deliver perceptible value to some segment of the customer base and leverage the investment in the base platform. As customers get savvier and share observations online, maintaining the value perception for the different versions can be difficult, as Lexicon discovered.

Based on some classic observations of human behavior, the Information Rules authors recommend adding a third, extremely expensive version of many products that will sell very few copies. The reason for this is the anecdotally familiar notion of extremeness aversion: The best-selling wine in most restaurants is the second cheapest, and buyers typically find comfort in medium sizes, intermediate variants, and other middle options. Having a “diamond” version with extra-special tech support or another apparently worthwhile attribute makes the formerly most expensive “deluxe” version look like prudent middle ground.

Another practice related to addressing the many demand curves for an information good is bundling. While travel presents a classic example in the cruise plus airfare plus hotel plus on-ship amenity credit for a single price, many information goods are bundled:

  • Newspapers are bundled by content (sports, entertainment, news, food, classified ads, etc.) and by time: Subscriptions are temporal bundles. As we will see, both of these bundles are being disaggregated by the Internet, with significant implications for news gatherers, citizens, and investors.
  • Software is bundled; office software with a word processor, spreadsheet, presentation graphics, and potentially other elements is the typical product.
  • Cable television service is bundled, not only in service tiers with access to different groups of channels but also in “triple plays” of TV, Internet, and wired telephony.
  • Encyclopaedia Britannica was a classic bundled offer, including thousands of entries that the owner would never read.
  • Music used to be sold in bundles called albums. Once file-sharing and, later, legal download services made it possible to obtain single songs, the music industry faced the same issues as the news sector. By contrast, few people want only “the best” five minutes of a movie or television show: Dramas and comedies and documentaries do not unbundle in the same way as music. A TV series is a bundle of episodes, but a movie is not a bundle of scenes.

Network Effects

Physical goods can exhibit network effects: Owning a Lotus sports car in an isolated region is made difficult by the lack of trained mechanics, spare parts, or a liquid secondary market for buying and selling these unusual cars. Compare Lotus to Toyota: The more people who own Toyotas in a given geography, the more options an owner has for service. As demand expands, additional dealers move in, providing opportunities for comparison shopping and potentially better service in the presence of competition.

Information goods exhibit many network effects beyond the file-sharing compatibility noted earlier. The English language is an information good with strong network effects, for example: The more people who learn it, the more communications options English speakers have and the easier global travel becomes. VHS versus Betamax and HD-DVD versus Blu-Ray illustrate how standards wars between competing technical systems (and potentially business models), are, at base, stories of network effects. Buyers did not want to own a soon-to-be-obsolete video player/recorder, in part because of the poor selection of prerecorded titles for the losing ecosystem.

Electronic games are perfect examples of network effects. This applies to game platforms (Xbox, Wii, PlayStation) as well as online games and online social networks: As more people flock to a platform, the more valuable and attractive it becomes compared to competing alternatives. Online dating is full of network externalities: The more people who sign up, the better tuned the matching algorithms and the better the selection of potential partners.

The fax machine is a commonly cited example: The first fax machine on earth could have cost $1 million, been gold-plated and encrusted in diamonds, but it would have been worthless for lack of partners. As the technology took hold, every time another fax machine was sold any-where in the world, the option value of any existing machine rose by a small amount.7 The same pattern of thought can be applied to Facebook, LinkedIn, and other social networking services. It is not clear that network effects do not have a ceiling, however: The 2 billionth Facebook user might not be as valuable as the 100 millionth.*

As the fax became available in offices all over the world, its network effect helped drive another phenomenon: positive feedback. Because of the importance of standards, defined both by market success (Microsoft Windows) and industry committees (IEEE† 802.11 or Wi-Fi), success in a technology can often build future success. The same holds true for gaming platforms: I can't play against you if our controllers and software don't agree. With GPS units, in contrast, brand proliferation is a possibility because of weak network effects: Your unit and my unit have little to do with each other. In some situations, search potentially being one, strong positive feedback leads to winner-take-all markets, or markets with a very small number of entrants protected by high barriers to entry.

Technical standards play a critical role in information economics. The choice of mobile phone standard for a given geography involved billions of dollars in revenue for the suppliers as well as intense frustration when a Verizon CDMA‡ phone from Louisville would not work in London or Lisbon. Railroad gauges (nonconforming geographies quickly got isolated), alternating current power supply (electric plugs in the United Kingdom are still not uniform), and the locking device used in container shipping8 (a truly transformative, if invisible, technology standard that helped enable globalization) provide other examples of the long-lasting importance of winning standards battles.

Lock-in

The technology phenomenon known as lock-in can be related to both network effects and winner-take-all positive feedback. As a comparison, consider an automobile. I can drive a Dodge and you a Toyota, but those choices are generally independent. If we were to switch vehicles for whatever reason, I would have access to the same roads and destinations, my driving skills would transfer quickly (except perhaps my ability to find the windshield wiper or headlight controls), and the experience would be generally unremarkable: Switching costs between cars, while not zero—my floor mats and snow tires might not fit the new vehicle—are extremely low.

Now consider software. At the personal level, your documents from Pages on the Mac should be read by my copy of Word on an HP PC. There might be some hiccups, but Mac–PC compatibility is reasonably high (even though that has not always been true). At the enterprise level, things get more complicated: Switching software can be a $100 million proposition as data stored to support billing in Oracle, for example, would likely not be ready to support sales force management in SAP. Users trained on one system face long and sometimes difficult transitions to new software and, sometimes, hardware. Investments in computing hardware, networking equipment, and other infrastructure may or may not be able to support a short-notice switchover.

Switching costs are part of lock-in, but there are other aspects as well. Just as Gillette blades won't fit Schick razors, neither will Dell printer cartridges fit a Xerox unit. Uniqueness in consumables and other modules (ink, video game cartridges, batteries) can impose switching costs or otherwise lock a customer to a product. Training, support, documentation, and other “soft” elements of a system can enforce lock-in; contracts are still another vehicle to achieve the same objective.

Closely related to lock-in (keeping a customer a customer) is lockout: preventing a competitor from entering a market or from converting an existing customer. The same proprietary inkjet cartridge that delivers high profit margins to HP after the printer was sold below cost or even given away in a bundle serves both purposes, lock-in and lock-out, as it prevents customers from moving their supply of ink to a Dell or Xerox machine.

Looking Ahead

The economics of information goods is sufficiently different from other kinds of goods that a wide variety of business decisions, personal behaviors, and legal issues have been affected. We will see the importance of lock-in, of free copy costs, and of the need for sampling in many examples in the forthcoming chapters.

Software Copying

Software copying is truly a problem without precedent. While the Xerox machine did not duplicate the printing (and particularly the binding) process, photocopying did have a major economic impact on at least two subsectors: textbook publishing and sheet music. Even so, these markets are small relative to software, music, movies, and proprietary research. Thus, the business model changes, court cases, and other signs of market adaptation to photocopying do not compare to the issues we face today.

There is surprisingly little literature on the historical arc of this issue. Nevertheless, given the prominence of Napster and its torrenting successors, the world's intellectual property interactions with the Chinese government and market, and the rise to economic power of the gaming sector, technologies of copying and its control have become central issues in a digital economy.

Duplicating a physical artifact requires, at the minimum, having access to both raw materials and skills. Whether that's steel and metalworking, food and cooking, or wood and carpentry, the issue of copying has been nonexistent in some markets and rampant in others. Elsewhere, particularly in branded consumer products such as watches and purses, copying—imitation rather than replication—can be a significant concern. Physicality also implies a moderate barrier to movement: Successful counterfeiters still face the problem of getting merchandise to customers.

The Internet removes the barrier of getting raw materials because code is malleable, easily transported, and closer to idea than infrastructure. The skill involved in copying a file, whether of executable code or data, is minuscule in contrast to what was needed to create either the software artifact or a physical original. Unlike physical counterfeits, which typically lack material quality and/or craftsmanship compared to the original, or analog copies in which successive generations of cassette tapes, photocopies, or faxes rapidly degrade, digital copies are nearly perfect. Furthermore, the means of production (a PC) is inexpensive and ubiquitous, which makes tracing the origin of copies harder than locating activities with heavier infrastructure, such as radio broadcasts and LP record pressing. Finally, the digital distribution channel is not only faster than a physical counterpart, it is instantaneously global.

Owners of digital content have relied on three tactics to combat copying.

First, there has been a series of attempts to make computer discs hard or impossible to copy by hiding files, using proprietary formats (such as game cartridges) or doing something called nibbilizing that rearranged the bit sequence of a copy. (Similarly, Macrovision enforced copy protection in analog VHS recorders.) As software distribution goes increasingly online, such measures still have their place, but they have not slowed the spread of copying by a significant margin. An exception is the video DVD (particularly Blu-ray), which can be copied by knowledgeable users but not casual ones: Proprietary protection of the digital bitstream in a DVD player or PC is enforced in hardware.

Second, software publishers can make it hard or impossible to use a copy. Some companies required users to consult a paper manual (“What is the last word on page 67?”) to generate crude authentication. One manual used symbols, and the company printed the manual in a color scheme that was impossible to photocopy. Still others relied on a hardware device called a dongle to activate a program in conjunction with the software and a generic PC (which quickly raised the problem of getting multiple dongles to interact gracefully on the same machine). More recently, a program can “phone home” via the Internet to see if software with a given serial number is in use on multiple machines. This approach can be made relatively robust for application software, and a variant called FairPlay prevents unauthorized copying of Apple's iTunes music files. Adobe is including auditing and monitoring of print materials in its LiveCycle Policy Server: If a user forwards an e-mail or file, prints it, or otherwise interacts with it, the originator of the document can be informed. How this extensive reach will affect task design and business processes remains to be determined.

Finally, software owners can lobby legislatures to change laws relating to copyright. The doctrine of fair use has been dramatically altered by both the duplication technologies of the past 100 years and the lobbying of content industries. There have been many unintended consequences: Copying application software off a 5¼-inch floppy onto a USB stick generally would be illegal, but with rapidly outmoded storage technologies, what is the owner of the application to do if she owns a PC without the appropriate outmoded drive? At the enterprise and government level, archiving digital assets often turns into an exercise in curatorship of a technology museum: Successive generations of outmoded hardware and software need to be maintained in the event that a given file or storage format needs to be read. Some estimates put the number of digital formats in the tens of thousands.

The content industry currently tends to reject copying as backup: If I buy an iTunes song (or 500 of them) and my host PC's hard drive dies, until cloud music lockers, I'd generally be out of luck even though all software was purchased and used under the terms of the license. Another example is DVDs: If I have two places of residence and want to watch a movie where I am, why must I buy a second copy of the same software rather than make a single copy for personal use? Once again, copyright law tends to prohibit any copying under blanket provisions. Such a move blurs the distinction between copying and counterfeiting, which are overlapping but not identical concepts. As processor speeds, graphics capability, and bandwidth all improve, content owners have lobbied to engineer copy protection deeper into the computing platform under the guise of anticounter-feiting but also to preclude many activities potentially protected by fair use.

This degree of restriction on customer behavior would be unprecedented. If I want to weld a Ferrari nose onto the front end of a Caterpillar dump truck, Ferrari (or Caterpillar) can't control what is done either with the purchased asset or, more important, an oxy-acetylene torch. Governments have engineered protection into color copiers, for example, and it's hard to argue against some degree of action in the public interest to protect the integrity of the money supply. But being able to use small clips of published text in scholarly works, for example, is standard practice—and essential to the expansion of knowledge in law or literary criticism. The parallel action of copying any portion of a movie for personal or scholarly use, however, might be illegal, depending on jurisdiction. Similarly, the study of cryptography is highly regulated: Scholars who decode copyprotection algorithms run the risk of prosecution if they publish their findings.

Herein lies the conundrum. The digital asset copying problem is unprecedented, so new kinds and degrees of measures will be required. At the same time, the legitimacy of certain forms of copying—for preservation, backup, or fair use—means that broad prohibitions, enforced in a general-purpose computing platform, come at an extremely high price to the purchasers and users of software and other digital media. No single answer will apply in every market to every application, but there have been some noteworthy efforts:

  • Use copying to build an installed base. Software makers with sufficiently strong cash reserves and long planning horizons can consider letting copies go relatively unpunished to build up a user base. Once a large body of people is trained on the software and file extensions and other conventions are well established, there are high enough switching costs that there may be reason to buy later versions of the product, particularly if the registration process is tightened, the pricing is attractive, and/or competitors have been weakened.
  • Use copying of entertainment to sell other entertainment. The Grateful Dead's support of tape-swappers who were allowed to record concerts is a widely cited example of allowing amateur copies to thrive as an adjunct to, rather than a substitute for, commercially released products. Other artists have used music downloads as an alternative path around the gatekeepers of radio playlists to build live audiences for concerts—where the T-shirt concession is tightly protected against counterfeiters.9
  • Reconsider analog. Several music labels, faced with plummeting CD sales, have turned to high-quality vinyl releases of both new and back catalog. Some high-end financial newsletters never left paper distribution. So-called three-dimensional printing produces analog physical objects based on digital models.
  • Utilize advertising-supported distribution. Archives are a perfect example: While a few newspapers have succeeded in charging subscriptions, most are failing to monetize their back issues with clumsy subscription or registration models that often don't support permanent linking from blogs or other sources of traffic. As paper newspapers continue to decline in circulation, the economic models of hybrid (digital + physical) production and distribution are ripe for reinvention. As a former big-city newspaper editor recently noted, this talk about the sky falling on newspaper ad revenues has happened before: In the late 1960s, political advertising moved overwhelmingly to television almost overnight, and the newspaper industry survived.
  • Think of King Gillette and sell blades after giving away razors.10 Giving away a multiplayer game title free, or allowing users to copy it without restriction, provides software publishers with a powerful distribution channel. (Back in the day it used to be called “viral.”) Recovering the cost can be achieved more effectively by making the proprietary online gaming environment a tightly controlled, for-profit affair, with monthly or annual renewals: Players will pay for access to other players, not for the plastic disc. Several online gaming environments (including Second Life) have spun off real economies based on cash flowing to merchants of virtual assets.

The list is not exhaustive but should suggest that there are enough viable responses to digital copying such that broad prohibition of all software copying will impose social costs that may outweigh proprietary benefits. It's important that there be open public debate to consider all of these potential costs, benefits, and risks of various courses of action. Copying and piracy, meanwhile, are not one and the same, but the rhetorical landscape tends to make this distinction harder and harder to draw. At the same time, true piracy—illicit DVD pressing plants, for example—should be considered and addressed separately rather than being conceptually lumped in with the many gray areas of fair use.

Notes

1. One example is Kevin Kelly, New Rules for the New Economy (New York: Viking, 1998).

2. Carl Shapiro and Hal R. Varian, Information Rules: A Strategic Guide to the Network Economy (Boston: Harvard Business School Press, 1999), p. 1.

3. Ibid.

4. See Steven Levy, “Secret of Googlenomics: Data-Fueled Recipe Brews Profitability,” Wired, May 22, 2009, www.wired.com/culture/culturereviews/magazine/17-06/nep_googlenomics?currentPage=all.

5. For a terminological distinction that argues that if the effect can be internalized by someone (possibly the platform owner), the effect cannot be called an externality, see www.utdallas.edu/~liebowit/palgrave/network.html.

6. Brian Barrett, “Lexicon Charges $3500 for a Repackaged $500 Oppo Blu-ray Player,” January 18, 2010, http://gizmodo.com/5450893/lexicon-charges-3500-for-a-repackaged-500-oppo-blu+ray-player.

7. See, for example, Michael Katz and Carl Shapiro, “Systems Competition and Network Effects,” Journal of Economic Perspectives 8, no. 2 (Spring 1994): 93–115.

8. On railroad gauges, see Varian and Shapiro, Information Rules, p. 208. An excellent source on the shipping container is Marc Levinson, The Box: How the Shipping Container Made the World Smaller and the World Economy Bigger (Princeton, NJ: Princeton University Press, 2008).

9. See, for example, Joshua Green, “Management Secrets of the Grateful Dead,” Atlantic Monthly (March 2010), www.theatlantic.com/magazine/archive/2010/03/management-secrets-of-the-grateful-dead/7918/.

10. The history of this innovation is less straightforward than often reported. See Alan Elliott, “Gillette, The King Of Razors,” Investors Business Daily, October 18, 2010. http://news.investors.com/Article/550711/201010181705/Gillette-The-King-Of-Razors.htm.

* But as my colleague John Parkinson points out, Facebook's 2 billionth member is the fifth person in some individual's social graph. Networks with power-law characteristics (see Chapter 6) behave in strange ways.

†Pronounced “Eye-triple-E”, the IEEE is the Institute of Electrical and Electronics Engineers, a global professional body responsible for, among other things, many technical standards in computing and communications.

‡For a time the predominant set of cell phone standards used in the United States, but rarely elsewhere, where GSM (Global System for Mobile) dominated. CDMA is shorthand for both CDMAOne and CDMA2000 technologies.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
18.224.51.55