purple motes

Prices for ongoing services often are much less clear than prices for retail-store products that have a price tag affixed to them.  Services do not have a natural physical unit, like an item that you pick up from a shelf.  Compared to a one-time purchase of physical products, an ongoing service purchase has much lower transaction costs for selling with a bundle of prices.  More generally, the shift toward a service economy involves significant changes in the definition of prices and the communication of price information.

Consider my telephone bill here in Arlington, Virginia.  In 1994, it contained two local-telephone-company service charges: "Residential line-main-message/measured Service" and "Touch-Tone Service."  In 2010, the number of local-telephone company service charges had increased to four differently described charges:   "Dial Tone Line," "ELS -- Unlimited Usage -- Flat Service," "Residence Local Usage Package Unlimited Flat/EAC Rate Svc" and "Sensible Minute."   I would guess that most telephone customers don't understand the definitions of these charges and don't understand how these charges relate to practical service choices that a telephone customer might make.

Charges semantically related to governments (taxes, fees, and other government-related surcharges) look similar to these service charges.   On my 1994 telephone bill, five separate charges related to governments were associated with local telephone service.  In 2010, the number of such charges increased to seven.[1]  Three of those seven charges semantically map directly onto three from 1994.   The remaining four in 2010 have only vague semantic relation to the other two in 1994.   I would guess that most taxpayers don't understand the definitions of these taxes and fees and don't understand the democratic political process that generated those taxes and fees.

Discussion of service rates tends to focus on changes in average or representative charges.  In inflation-adjusted dollars, my local telephone service charges increased 6% from 1994 to 2010.[2]    Taxes and fees increased 108%.   Moore's Law and astonishing leaps in communications technology offer hopes for a new economy.   The average charges on my telephone bill exemplify an old economy.

But in the semantics and ontology of its charges, my telephone bill testifies to new implications of service pricing.  Service prices are not naturally meaningful to customers.  That creates opportunities for web-based bill monitoring services to provide meaningful price information to customers.   To the extent that particular services are subject to a formal, administrative process of tariff regulation, controlling and communicating service pricing vocabulary could be a valuable regulatory tool.  More generally, concentrating government regulation on wholesale services may not be appropriate for modern service economies.   Balancing appropriate government regulation across wholesale and retail services might create more beneficial competition among commercial service-providers.

*  *  *  *  *

Data: Telephone bill comparison, Arlington, VA, 1994 to 2010 (Excel version); US residential telephone rates in urban areas, 1994 (Excel version).

Notes:

[1]  The "Federal Subscriber Line Charge" is included here with charges semantically related to governments.  However, the local telephone company receives the revenue from this charge.

[2]  The FCC's urban telephone rate survey found that Bell Atlantic's representative local measured service charge in Richmond, Virginia, in 1994 was $13.59 per month.  However, the lowest charge was $5 month.   My charge was $9.23 per month.  This spread in prices highlights the difficulties of surveying telephone rates.

Offer Comments

Telephone companies operated the first network-mediated, mass-market information search service.  It's called directory assistance.   Make a telephone call, voice a query, and get back relevant information.

By some narrow metrics, U.S. telephone companies have quite successfully developed this information search business.  Prior to 1974, directory assistance in the U.S. was a freely offered ancillary service to regular telephone service.  At that time, directory assistance had significant marginal cost:  additional human operators were required for additional information lookup and query response.  Telephone company representatives and economists argued that free directory assistance is economically inefficient and unfairly advantages heavy users of directory assistance.[1]   Those arguments prevailed.  Directory assistance became a pay-per-use information search business.

The price per call for directory assistance has risen substantially since the mid-1970s and the monthly allowance of free directory-assistance calls has fallen.  Most jurisdictions that established charges for directory assistance in the mid-1970s set a five-call-per-month allowance for free directory-assistance calls.   Subsequent calls cost 10 cents or 20 cents per call.   Current allowances for free directory-assistance calls generally are lower than five calls per month.  Allowances continue to be reduced.  Recent telephone-company petitions to state regulators successfully reduced the allowance from three calls to two in Virginia, and from three to one in Tennessee.  In 2007, the average charge for a directory-assistance call from a landline phone was $1.28.   Hence, adjusting for the rise in the consumer price index, real directory-assistance prices are two to three times higher in 2007 than they were in 1975.  That increase in price is particularly impressive given the shift from human operators to automated systems.  Automated systems undoubtedly provide significant cost savings.[2]

Telephone companies have also successfully removed directory-assistance wholesale service from federal telephone service price caps.  In 1992, regional Bell Operating Companies (RBOCs) reported $209 million in federal price cap revenue for offering directory assistance to other telephone companies at a wholesale price averaging 28 cents per call.   Reported directory assistance calls and revenue subsequently fell, with reported directory assistance call volumes falling 5% per year from 1992 to 1996, and  48% per year from 1996 to 2000.  Other industry data indicate that total interlata directory assistance calls were growing about 8% per year from 1992 to 1999, and total directory assistance business revenue was growing about 10% per year in 1994.  By 2009, RBOCs reported almost no directory assistance calls or revenue, while the total value of the retail directory service business for all service providers had grown to roughly $9 billion.[3]  The reported fall in RBOC regulated directory assistance calls and revenue may have come from a shift to automated provision of wholesale directory assistance service.[4]   In any case, RBOCs successfully removed in practice directory assistance wholesale service from federal price caps.

Despite telephone companies successes in establishing new charges, raising prices, and escaping regulation, telephone companies failed to develop vitally important business opportunities in information search.  Information search changed from high-marginal cost operator-provided services to low-marginal-cost automated services.[5]  With low-marginal-cost, network-based services, encouraging use of free services can be a valuable way to develop businesses.  Google and others, not telephone companies, have prospered from the development of the information search business.

*  *  *  *  *

Data: U.S. directory assistance service statistics, 1974 to 2009 (Excel version).

Notes:

[1] See McDonald, J., "The Use of Management Science in Making a Corporate Policy Decision -- Charging for Directory Assistance Service," Interfaces, v. 7, n. 1, pt. 2, Nov. 1976; Daly, George and Thomas Mayor, "Estimating the Value of a Missing Market: The Economics of Directory Assistance," Journal of Law and Economics, v. 23 (1980). pp. 147-166; Zachry, Charles C. Jr, "Usage-Sensitive Pricing for Directory Assistance," Public Utilities Fortnightly, v. 100, n. 12 (Dec. 8, 1977) p. 23.   These works emphasized the value of managerial science and economic analysis.  Introducing charges for directory service reduced the volume of directory assistance calls by about two-thirds.  See Daly and Mayor (1980) p. 155; Zachry (1977) p. 23.

[2] Here are data on the directory assistance prices and free-call allowances established in the mid-1970s and those about 2008.  On June 11, 2008, the Virginia Telecommunications Industry Association petitioned the Virginia State Corporation Commission to eliminate the directory-assistance free call allowance.  The Commission established a docket, issued an order for notice and comment, considered the 21 responding parties' comments, and issued an order on Dec. 12, 2008.  That order reduced the free-call allowance from three to two per month.  On Jan. 8, 2009, the Tennessee Regulatory Authority (TRA) approved a traiff reducing Embarq's directory-assistance free-call allowance from three to one (reconsideration denied in TRA order on Apr. 7, 2009).   TRA approved a reduction in BellSouth's free-call allowance from three to one on Apr. 17, 2007.   See Exhibit 1 in relevant Citizens Telecommunications Co. filing.   In the mid-1970s, unlimited free directory-assistance calls were typically allowed for handicapped persons, users of coin-operated phones, hospitals, and hotels.  They are now typically allowed only for handicapped persons.

[3] For data and sources, see the directory assistance business statistics.   The RBOCs also collected considerable revenue ($84 million in 1992) through an " information surcharge."  The information surcharge is a separate, wholesale per-minute charge applied to all originating and terminating interlata switched access minutes of telephone companies' interconnecting interlata telephone traffic.   According to AT&T in 1998, the information surcharge had been applied since 1985 on the basis of an interim waiver for the recovery of  "costs relating to white pages and other undefined directory assistance costs."   The CALLS access reform, enacted in 2000, eliminated the information surcharge through targeting of price reductions required under price caps (see 47 CFR 61.45 (i)(2)(ii)).

[4] Directory advertising (Yellow Pages) historically has been a highly profitable, non-regulated business of telephone companies.  In recent years, the Yellow Pages business has been under acute competitive pressure.  In an order released on Jan. 23, 2001, the FCC ruled that local telephone exchange providers must offer nondiscriminatory access to their directory listings, at nondiscriminatory and reasonable rates, to competing directory service providers.   See FCC Order, Provision of Directory Listing Information under the Telecommunications Act of 1934, 16 FCC Rcd 2736 (2001) [as MS Word document].  This order cannot account for trends in directory assistance calls prior to 2001.

[5] Based on Bell information, the marginal cost of directory assistance was estimated to be 19 cents per call in 1977.  See Daly and Mayor (1980) p. 157.  Currently many companies, such as Jingle, Bing, and Google, provide free directory-assistance services.

Offer Comments

Persons tend to prefer what is familiar.  For example, political candidates heavily advertise themselves with signs and bumper stickers that typically include just the candidate's name and office sought.  While voters find this same information on the ballot when they go to vote, repeated exposure to a candidate's name evidently induces voters to prefer that candidate.  Brand advertising, which has been highly successful across a wide variety of media ecologies, is oriented toward the same effect.

Preferring the familiar favors survival in a wide range of actual human environments.  Because persons recognize dangers over time and avoid them, familiar surroundings are less likely to be dangerous than unfamiliar surroundings.  Persons who eat familiar foods are less likely to suffering poisoning than persons who eat unrecognized substances.  Familiar persons are more likely to offer help than are strangers.  Preference for the familiar is a simple decision rule that makes sense from evolutionary and ecological perspectives.

Preferring the familiar can produce good decisions on contrived tasks not directly related to familiarity. For example, presented in the laboratory with pairs of cities and told to choose which city is larger, students more often chose as larger a recognized city that was paired with an unrecognized city. Because actual patterns of conversation and media content refer to larger cities more often than smaller cities, choosing the recognized city identifies the larger city with better than random odds. In fact, on the pairwise city-size decision task, American students correctly choose the larger city more often for German city pairs than for American city pairs.  The opposite was true for German students.  This surprising result indicates the merits of the recognition heuristic.  The recognition heuristic can be applied only to city pairs for which one city is recognized, and one isn't.  City pairs from a foreign country provided more scope for the recognition heuristic, and the recognition heuristic produced better decisions than decisions made when information could be recalled about both cities.[1]

Actual human decision processes point to important characteristics of practical decision logic. No formal decision logic can determine the scope of information that it considers. Every decision necessarily does not consider some possible information. An optimal decision is necessarily defined with respect to an assumed structure of information.  Recognition depends on biological capabilities, a wide range of life experiences, and non-problem-specific characteristics of the environment.  Recognition points to the huge scope of possibilities for useful information.[2]

More information, however, can make predictions less accurate.  In the real world, one does not know the data-generating process for the information under consideration.  Nor does one know whether that data-generating process is the same as the data-generating process relevant to the circumstances of the prediction.  Hence over-fitting and non-representative samples are always risks in real-world statistical applications.  More information can lead to a better estimate of the wrong data-generating process and hence worse predictions.[3]  The data-generating process for less information may be implicitly or explicitly better estimated and more consistent over time.  More information makes more known, but does not necessarily provide a better guide to the unknown.

* * * * *

[1] Goldstein, Daniel G. and Gerd Gigerenzer, "Models of Ecological Rationality: The Recognition Heuristic," Psychological Review v. 109, n. 1, pp. 75-90.

[2] Processing fluency at a lower level of sense than recognition is also important in decision-making.

[3] Thus, for example, in some situations the median, which uses only ordinal information, provides better predictions than the mean.  Gigerenzer, Gerd, "Why Heuristics Work," Perspectives on Psychological Sciences v. 3, n. 1, pp. 20-9, provides a nice overview of how biology ("adaptive toolbox") and real-world decision-making circumstances (ecological rationality) support fast and frugal heuristics.  Gigerenzer is an eminent academic and research leader in this field.  While I know much less, it seems to me that, in this short article, Gigerenzer doesn't adequately distinguish between "irrelevant information (or 'noise' )" and model mis-specification / structural change.  If one knows correctly the data-generating process, larger sample sizes typically serve well to increase prediction accuracy in the presence of noise.  That's not true for a mis-specified model.  Moreover, there is no statistical test for the true data-generating process for data not yet known.

Update:  Section 2 of Gerd Gigerenzer and Henry Brighton, "Homo Heuristicus: Why Biased Minds Make Better Inferences," Topics in Cognitive Science 1 (2009) 107–143,  provides a good discussion of model mis-specification.  It uses the terms bias, variance, and noise in a way that might be jarring for someone focused on textbook statistics.  Textbook statistics, however, typically do not adequately recognize the reality that the true data-generating process is always unknown.   Moreover, in practical circumstances the law of large numbers confronts important limitations:

• increasing the sample size is often costly or not feasible
• a larger sample may create greater model mis-specification because different data-generating processes may apply to different subsets of the sample
• a larger sample enables greater over-fitting and increases the importance of correct parametrization

On the other hand, big datasets and complex algorithms have been successful in practical domains.

Offer Comments

What's the difference between a search service and a service perfectly targeting informative advertising to its users?  Susan Wojcicki, Google VP, Product Management, recently stated:

Google's advertising business was founded on the core principle that advertising should deliver the right information to the right person at the right time. This is very similar to our mission in search, and, like our colleagues in search, those of us on the ads team are constantly striving to achieve better results.

Daniel Tunkelang insightfully responded:

Google insists that it maintains a wall between its search and advertising businesses. But Wojcicki’s post–which is on Google’s official blog–suggests otherwise, at least in spirit. If Google believes that both search and advertising aim to “offer relevant content” and “deliver the right information to the right person at the right time”, then why put up a wall at all?

If an advertising platform truly provides only the most relevant, useful, timely information to its users, it gives advertisers no incentive to bid against each other to place ads.  The most relevant, useful, timely ad, not the highest bidding advertiser, would be presented to users.  Ad results wouldn't differ from search results.  The perfect advertising service would provide no incentive for advertising expenses.

For both search and advertising, emphasizing relevant, useful, timely information obscures the importance of emotions in human decision-making.  A vast amount of information is subconsciously processed within the human body.  Only some of this information is brought into the brain's working memory and higher-order cortical regions.  Competing for attention is not just the business of advertising.  Emotions have a major role in the ongoing competition for attention within the human brain.

Actually making a decision requires a motivational push, not just information processing.  In an influential book describing new understanding of emotions, a neuroscientist described the behavior of a patient with damage to brain regions associated with emotional processing.  The patient was offered the choice between two dates for a laboratory appointment:

For the better part of a half-hour, the patient enumerated reasons for and against each of the two dates: previous engagements, proximity to other engagements, possible meteorological conditions, virtually anything that one could reasonably think about concerning a simple date.  Just as calmly as he had driven over the ice, and recounted that episode, he was now walking us through a tiresome cost-benefit analysis,  an endless outlining and fruitless comparison of options and possible consequences.[1]

When the doctors finally halted this process and asked the patient to come on the second of the two dates, the patient was happy to accept that choice.   Human-Computer Information Retrieval services should not be implicitly modeled on the behavior of emotionally damaged humans and emotionless computers.  An impulse to choose, apart from organizing and accessing information, is a key aspect of human search behavior that concludes with choice.

Search and advertising involve much different emotional patterns in users' information processing.  Web search produces a list of choices designed to be informative, not emotive.  Given the importance of search result rankings to users' choices, users seem to view these results with a feeling "tell me right now" and "whatever, good enough."  The emotional drive is not primarily to make a good choice, but to go somewhere, to finish the search.[2]

Advertising has much more specific emotional targeting than search.  Advertisers design brand names and tag lines to be emotive.  Where possible, they use images and audio-visual materials that evoke favorable feelings.  Compared to clicking on search results, users are much more likely to click on an advertisement based on emotions that the specific advertisement generates.

Because emotions are an integral aspect of human information-processing and decision-making, advertising's future is secure.  Even if they perfectly target ads to be relevant, useful, and timely for users, advertising platforms cannot effectively rank the emotional value of ads.  Persons and companies have deep-seated feelings about how good they and their products are.  These feelings are fully sufficient to drive vigorous bidding for advertising opportunities.

Search will not replace advertising.  Information technology can greatly improve human's ability to organize, rank, and make accessible information.  No such technology exists for human emotions.  That fact will keep advertising platforms in business.

Notes:

[1] Damasio, Antonio R. (1994) Descartes' error: emotion, reason, and the human brain (New York: Putnam) pp. 193-4.  Skin conductance responses, which typically are correlated with emotional responses, have been documented to precede a risky choice in normal individuals. Such a response does not occur for individuals with bilateral damage to the ventromedial prefrontal cortices. The later persons' choices systematically differ from normal persons' choices within the same laboratory experiment. See Bechara, Antoine, Hanna Damasio, Daniel Tranel, and Antonio R. Damasio, "Deciding Advantageously Before Knowing the Advantageous Strategy," Science, V. 275, 28 February 1997, pp. 1293-5. Damasio's somatic-marker hypothesis provides a more general framework for understanding such effects.

[2] Image search seems to involve a different emotional pattern than web search.  See the recent talk of Peter Linsley, Google Image Search Product Manager.

1 Comment

While new media tends to be associated with delusions of an effortless, viral-promotional road to success, small, information-sector firms in the U.S. on average have relatively high advertising expenses.   In 2005,  small corporations (corporations with business receipts under $1 million) and nonfarm sole proprietorships had advertising expenses amounting to 1.4% and 1.2% of business receipts, respectively.  In the information sector, the corresponding figures for small corporations and sole proprietorships were 1.9% and 3.8%.  Sole proprietorships, which are likely to be younger and more innovative than small corporations, had much higher advertising expenses.

Small, new-media businesses have much higher advertising expenses than small, traditional information businesses.  Among sole proprietorships, the information-sector subcategory "Internet publishing and broadcasting" had advertising expenses amounting to 10.0% of business receipts.  That contrasts sharply with advertising expenses of 1.1% of business receipts for "Broadcasting (except internet) and telecommunications".

Data note: Based on 2005 IRS data on advertising expenditure reported on business tax returns.

1 Comment

The U.S. Census Bureau's Service Annual Survey is a great source of data on the information and communications industries.  I've put together a dataset based on this source to make the data for 1998 to 2007 easier to use.

The Service Annual Surveys cover various categories of revenue, expenditure, and inventories.  The data are collected mainly from surveys of firms with employees (employer firms).  The firms are assigned to industry categories based on the North American Industry Classification System (NAICS).   NAICS 51, Information Sector Services, includes wireline and wireless telecommunications, radio, broadcast television, cable television, and other communications services.  NAICS 54, Professional, Scientific, and Technical Services, includes display advertising, direct mail advertising, and advertising services such as advertising agencies.

I'll call the dataset I put together the Information and Communication Industries Revenue (ICIR) Dataset.  It contains revenue data for the information sector (which includes communications industries) and advertising and related services from the Service Annual Survey for 2000, 2004, and 2007.  These three report years include annual data spanning 1998 to 2007, and, in most cases, contain the most recent revisions of previous years' data.  Because many reporting categories have changed across this decade, all the data cannot be easily organized into one consistent, detailed scheme spanning the decade.  Having the data in an analytically convenient form facilitates extracting, re-organizing, manipulating, matching, and comparing data of interest.

You should refer to the official Service Annual Survey reports to understand and validate the ICIR Dataset.  The variable named "lid" organizes the ICIR Dataset in the line order of the Service Annual Survey reports.  The ICIR Dataset includes some estimates that are not Census Bureau estimates.  See the summary tab for a description of these additional estimates.   To understand the meaning of blank values in the ICIR Dataset, refer to the original Service Annual Survey reports.  The dataset has a complex category scheme that does not include complete partitions at some levels.  Once again, refer to the original reports to understand the items.

Here's the Information and Communication Industries Revenue (ICIR) Dataset as a browser-friendly Google spreadsheet.  You can also download the ICIR Dataset as an Excel workbook.

Offer Comments

"I don't read fiction, Doug. Fiction's just not true!"

"I don't have time to read the newspaper. I just form my own opinions."

"I'm fed up with all the media concentration. I've decided, from now on, I'm just going to listen to NPR."

2 Comments

Searching video is notoriously difficult. That implies a scaling problem for a large video repository like YouTube. Given relatively fixed amounts of video tagging and category information, more videos imply relatively less information for searching among videos and targeting ads to videos.

YouTube's announcement of new APIs for external use of the YouTube video platform points to better information for video search and ad targeting. Videos uploaded and played through a third-party site are segmented by the specific third-party site. Third-party sites also have differentiated text correlated with the content of associated videos. This additional information allows better video search and better video ad targeting compared to what's possible for videos uploaded and viewed on the YouTube site.

Online video services have many different niches. Decentralizing video uploading and playing provides a more scalable business for YouTube.

1 Comment

Providing network infrastructure need not be limited to a choice between the model of public roads and the model of selling soap. The provision of public roads depends on market transactions for a variety of goods (construction worker services, trucks, asphalt, etc). Selling soap depends on a variety of public services (money supply provision, law enforcement, public right-of-ways, etc.). Whether in India, Ireland, or Silicon Valley, initiatives to provide network infrastructure are making interconnections between different organizational forms more complex. Revenue models are expanding from taxes, subscription, and advertising to include a variety of public and private sponsorships, in-kind contributions, and special benefits for anchor users.

Impurity is a traditional human concern. In some circumstances, another name for public-private partnership is bribery and corruption. Failed and wasteful network infrastructure projects that involve governmental entities undoubtedly exist. For-profit network providers, who cannot fail without serious public effects, have made dire business mistakes and squandered huge amounts of money. Government entities' judgments about the services that users value are not likely to be better than those that for-profit network providers have made.

The history of libraries in the U.S. suggests that organizational diversity can have enduring value in information infrastructure. In the American colonies and the United States prior to 1876, most organizations that founded libraries were neither government bodies nor commercial organizations. A wide variety of organizations established libraries (see Table 1). The most commonly created form of library was a social library:

a library owned by an association formed to establish and operate a library intended for its members' use. Usually, the members subscribed for stock in order to purchase the initial collection, which was general in subject matter. Then they were assessed a smaller sum (a "tax") each year to keep up the collection.[1]

Public libraries, meaning libraries that government bodies owned and made open to all or most citizens without a specific-purpose charge, began to grow only from the mid nineteenth-century. As late as 1900, about as many social libraries existed in the U.S. as did public libraries (see Table 2).

Public libraries had different characteristics than social libraries. Smaller populations and more recently settled areas favored social libraries, while larger populations in cities with a longer history favored public libraries. Social libraries had typical lifespans about thirty-five years, with considerable variance.[2] Public libraries tended to be more permanent organizations that endured in organizational form through jurisdictional consolidations. Public libraries had a more secure base of funding and grew in size relatively rapidly. Across the last twenty-five years of the nineteenth century, public libraries came to predominate among the largest libraries (see Table 3). These historical facts are consistent with general comparative organizational characteristics: compared to social organizations, government organizations are more difficult to establish and require more developed government administrative capabilities, government organizations are more enduring, and government organizations are more favorable for organizational growth. Shifts in library organizational forms were in part a response to changing demographic and political circumstances.

Different organizational forms, however, interacted significantly. Social libraries and public libraries coexisted as important forms of library organization for more than half a century. Through at least 1875 and possibly into the beginning of the twentieth century, social libraries were widely regarded as a valuable form of library organization.[3] Some public libraries evolved from the buildings and collections that social libraries established. In the 1930s, more than a sixth of all "public" libraries in cities with population 30,000 or greater were libraries for which "the library society and the town government shared control in a manner that makes it difficult to know how power was divided between the two bodies."[4]

In a long-run international historical perspective, the U.S. has had a relatively highly developed information economy. New organizational forms for book sharing, network infrastructure, and telephone service are not just necessary entrepreneurial experiments in rapidly changing technological circumstances. Diversity in the organizational forms of its information infrastructure has been an enduring characteristic of the U.S. information economy. Organizational diversity may be a key to growth of the information economy.

* * *

Notes:

[1] From "Definition of Types," American Libraries Before 1876, Davies Project.

[2] McMullen (1985) p. 214.

[3] Id. pp. 218-20.

[4] Id. p. 223.

References:

McMullen, Haynes (1985), "The Very Slow Decline of the American Social Library," Library Quarterly, vol. 55, no. 2, pp. 207-225.

McMullen, Haynes (2000), American Libraries Before 1876 (Westport, Conn.: Greenwood Press).

4 Comments