purple motes

While newspapers clearly are suffering economically, different sources of newspaper advertising statistics show different extents of decline.  The Newspaper Association of America (NAA) website reports a 27.94% year-on-year decline in newspaper advertising revenue for the third quarter of 2009 (2009 Q3).   Data from the U.S. Census Bureau's Quarterly Services Survey indicate a 14% year-on-year decline in newspaper operating revenue for 2009 Q3.   Evidence from the Census Services Survey for 2008 and plausible assumptions suggest a 17% year-on-year 2009 Q3 decline in newspaper advertising revenue.   Hence, compared to NAA statistics, reasonable estimates from the Census Services Survey imply a 10 percentage point smaller decline in newspaper advertising revenue in 2009.

The Census Services Survey data seem to me more credible than the NAA statistics.  The NAA has obvious self-interest in newspaper industry statistics.  The NAA does not report its statistical sources and estimation procedures.  The NAA statistics also include no estimates of the reliability of the reported data.  The Census Bureau, in contrast, is an expert government agency.  The Census Services Survey carefully describes the sampling and estimation procedures.  In addition, the Census Services Survey includes variability estimates.   The coefficient of variation for the reported 2009 Q3 newspaper operating revenue is 5.6%.  The coefficient of variation for general newspapers' advertising revenue in 2008 is 3.1%.  The standard error of the 2008/2007 general newspapers' advertising revenue change is 1.5 percentage points.   These statistics suggest that the true decline in newspaper advertising revenue year-to-year 2009 Q3 is probably between roughly 12% and 22%.[1]   The NAA figure is most likely an over-estimate.  The NAA's reporting of that figure to one-hundredths of a percentage point is misleading.

The NAA newspaper advertising figures have been much larger than Census figures, but the NAA figures are dropping rapidly relative to Census figures.  For 2008, the Census Services Survey showed $29.8 billion in newspaper advertising revenue.  The NAA figures showed $37.8 billion in newspaper advertising revenue.[2]  That's 27% higher than the Census figure.  However, in 2006 the NAA newspaper advertising revenue figure was 45% higher than the Census figure.    Some of the decline in NAA reported newspaper advertising may be an artifact of NAA's advertising accounting / estimation procedure.

*  *  *  *  *

Data: U.S. newspaper advertising trend, with comparison between Census and NAA data (Excel version).  Large dataset of historical advertising expenditure by media.

Notes:

[1] The decline in annual newspaper advertising revenue was 10% from 2007 to 2008.  I've assumed a 2.5 percentage-point standard error in the decline in quarterly newspaper advertising revenue 2008 Q3 to 2009 Q3.  That number seems reasonable given the error bounds for the quarterly revenue figure, the annual revenue figure, and the change in annual revenue 2007 to 2008.  Twice the standard error, or 5 percentage points, defines a 95% confidence interval.  I've formed that interval around the point estimate of a 17% revenue decline.

[2]  That's the NAA figure for print and online advertising . Since the Census figures are establishment-based, newspapers online advertising revenue are included in the Census figures.  Hence the NAA print and online total is the relevant comparison.

Offer Comments

Yellow Pages print directories currently are distributed to all residences in an area unless a resident explicitly opts out of receiving a directory.  Many persons not interested in receiving a print directory probably don't bother to opt out.   The result is wasted money and wasted natural resources in distributing, re-collecting, and recyling unwanted directories.

An alternative would be to distribute print directories only to residences that explicitly request one (opt-in).  Some persons who might use  a print directory would fail to request one.   But customers who find Yellow Pages very useful probably would be willing to request a free directory.

The requesting process could be used to customize directories for requesters.  Suppose persons interested in receiving a Yellow Pages print directory called a number or went to a website and answered some demographic and behavioral questions ( "How frequently do you use coupons?", "How often do you go out to eat?", "Do you like to try new products?").   A customized directory designed for particular interests and behaviors could then be delivered to requesters.   Information is a key currency in new-media advertising.   Even low-frequency interaction between Yellow Pages' users and Yellow Pages producers might create a more valuable product.

Opt-in Yellow Pages print directory distribution probably isn't an attractive business model as long as many small businesses believe that comprehensive distribution of Yellow Pages print directories implies extensive use of Yellow Pages.  More generally, smaller businesses have less incentive to invest the time and skill to analyze advertising-relevant information.  In the U.S. in 2005, businesses with less than a million dollars in annual business receipts spent in total $27 billion on advertising.  However, they had average annual advertising expenditure of only about $1000 per business (small business advertising data).   Successful advertising products for small businesses have to make advertising performance information easy to understand.

Offer Comments

Telephone services for small businesses is a significant segment of the communications industry.  SOHO (small office/home office) is a widely recognized customer segment.  The SOHO segment is large and growing relatively rapidly.  In the U.S. in 2002, non-farm businesses with fewer than ten employees and with more than $10,000 in annual receipts numbered about 14 million and had $2.7 trillion in business receipts (about 12% of total U.S. non-farm business receipts).   Small businesses that rent office space might also be able to contract with the facilities manager/owner for telephone services.  However, home-based small businesses don't have that contracting opportunity.  In the U.S. in 2002, non-farm, home-based firms with fewer than ten employees and with more than $10,000 in annual receipts numbered about 6.8 million and had about $700 billion in business receipts.  These home-based, small businesses are predominately in construction, retail trade, and professional, scientific, technical and other services.[1]

Small businesses value fancy voice telephone services.  Persons running small businesses have to manage communications with suppliers, customers, and contractors.   With respect to voice telephone communication, they benefit from services such as programmed greetings (attendant menus), flexible, programmed call routing across multiple devices and locations, call line hunt groups (if Jasmine's line is busy, automatically route the call to Sasha's line), and voice mail.   They also benefit from being able to manage these services personally from a variety of devices.  Large businesses have acquired such capabilities through private-branch exchanges (PBX) and Centrex systems.  Capital costs, skill requirements, and maintenance requirements probably favor a hosted PBX or Centrex-type system for small businesses.  U.S. businesses with less than 9 Centrex lines purchased about 2.8 million Centrex lines in 2002.   Those customers, who averaged 4.3 Centrex lines each, accounted for about 16% of total Centrex line purchases.[2]   Small businesses have long been significant customers of manageable voice telephone services.

Competition for providing communications services to small businesses is producing cheap, highly capable services.  For example, Junction Network's OnSIP provides hosted PBX service for small businesses.  Its $40 per month SOHO package offers unlimited short-number extensions, free, unlimited intra-extension calling, five voice mailboxes, three attendant menus, three hunt or simultaneous-ring groups, dial-by-name directories, business-hour routing of incoming calls, and a browser-based call management interface.  OnSIP describes itself as "a complete business VoIP service for 5 to 100 users."  The disadvantages of OnSIP for businesses with only a few persons appears to be cost and complexity.  Google Voice (an app pre-installed on the Nexus One) is a free service, designed for individual use, that has some capabilities similar to OnSIP.  VoxOx is another free service designed for individual use.  VoxOx offers a powerful virtual personal assistant as well as a dead-end feature that's probably even more valuable than industry-standard sorry-gotta-go scripting technologies.  BT's Ribbit provides a platform on which a wide variety of cost-effective, manageable voice telephone services can be developed.

Competition in providing manageable voice telephone services for individuals, non-employer businesses, and employer businesses with only a few employees is likely to reconnect the telephone business to the local advertising business.  AT&T introduced "Where to Buy It" telephone directories in 1928.  In 2007, U.S. Yellow Pages directories had about $14 billion in advertising revenue.  Moreover, about $71 billion of newspaper, radio, and television advertising is local advertising.[3]  Print yellow pages, newspapers, radio, and television are moving to networked digital devices.   Providing small business telephone services is likely to provide an important advantage in providing small-business advertising and local information search.  That's the historical story of the Yellow Pages.  That's a story that now seems ready to be re-enacted, but perhaps with different main characters.

*  *  *  *  *

Data: U.S small businesses and Centrex services workbook (Excel version); coded Bell Atlantic / Verizon Centrex rate elements, 1998-2009, compiled from the full rate-detail dataset

Notes:

[1] The figures for home-based businesses are my estimates.  For the source, detailed data, and estimation formulas, see the small business worksheet in the Excel version of the small business/Centrex workbook.

[2] For data details, see the Centrex worksheet.

[3]  Radio, television, and newspaper advertising, separated into local and national, is available in the full Coen Advertising dataset.  Those figures show local radio advertising, local television (cable and broadcast) advertising, and local newspaper advertising to be 53% of total advertising.  The Coen over-all local/national advertising figures for 2007 show local advertising to be 34% of total advertising.  However, the Coen over-all local/national ad figures include all direct mail advertising and almost all miscellaneous advertising as national advertising.

Offer Comments

The traditional mass distribution of yellow pages print directories isn't effective where I live.  In my 148-unit apartment complex, stacks of new yellowbooks were recently placed outside entryways during the day on Thursday.   About 7pm that evening, 146 yellowbooks remained.  The next evening, 139 remained.    The following Tuesday morning, 134 books remained.   Later that day, apartment staff recycled those remaining books.   Hence apartment residents picked up only about 10% of the yellow pages directories placed in entryways.

I also looked at yellow pages directory distribution through two CVS stores in my neighborhood.  On December 16, 2009, the day my apartment complex received the new 2010 yellowbooks, the first CVS had five 2009-2010 (old) yellowbooks, and the second had two.  The top books in both stacks were quite dusty.  I'd guess that no one had picked up a yellowbook at these stores for at least several months.  By January 6, 2010, the first CVS had removed the distribution rack and the yellowbooks, while the second CVS still had the distribution rack.  It contained one 2009-2010 yellowbook.  Hence someone picked up from CVS an old yellowbook.  Maybe someone is using it to make a spending decision, or to clean a bike, or just maybe a homeless person took it for toilet paper.

Whether someone bothers to pick up a yellow pages directory is a crude measure of the performance of yellow pages ads.  But given that only 10% of the yellow pages directories were picked up from my apartment complex, the performance of ads in those directories probably isn't very good.  The performance of ads in yellow pages directories available for pickup in CVS probably isn't much better.

New media is offering powerful new means for making advertising relevant, useful, and accountable.  Changes in advertising are disrupting major businesses and challenging well-established organizational practices.   That about 418 pounds of yellow pages directories were dropped at my apartment complex and then carted away is just a small act in the astonishing spectacle of media change.

Data: yellow pages print directory retrieval in my apartment complex (Excel version)

1 Comment

The once lucrative yellow pages print directory business is shrinking quickly in Northern Virginia.   Some simple measurements and calculations indicate that the 2010 yellowbook yellow pages listing  and advertising volume has shrunk more than 33% over the past year.

The shrinkage this year seems to have been designed to be less obvious than the shrinkage last year.  The number of yellow pages this year (1062) is only 9% less than the number of yellow pages last year (1162).  This year's yellowbook lacks the coupon section that took up 28 pages in last year's directory and also does not have any thick paper advertising pages (last year had two).   Making up for that reduction in thickness was an 80% increase in white pages from 260 pages in the 2009 book to 468 pages in the 2010 book.  The net effect of these changes was to make the 2010 yellowbook the same thickness as the 2009 yellowbook.

Changes in page size and print size allowed the yellowbook to retain its thickness.  The page area in the 2010 yellowbook is 15% smaller than in the 2009 yellowbook.   Block advertisements, e.g. a two column, quarter page advertisement, were reduced in size proportionally.  However, the print size for a plain listing (business name, address, telephone number) was increased an estimated 36% in the yellow pages and 103% in the white pages.  Ignoring block advertisements and computing equivalent pages as the number of listing that would fit on a 2009 page, then the number of equivalent yellow pages fell 33% from the 2009 book to the 2010 book.  If block advertisement space had the same proportion to plain listing space in 2010 as it did in 2009, then a 33% decline also estimates the reduction in yellow page space including block advertisements.  Comparing the two books indicates that the number and size of block advertisements (relative to page size) has decreased significantly.  Hence over-all listing and advertising volume in the 2010 yellow pages has decreased more than 33% relative to the 2009 volume.

The extent of obfuscation traditionally associated with advertising is second only to that in filings before government regulatory agencies.   Nonetheless, trends in the advertising industry point to the growing business success of advertising that emphasizes accountability.

Data: measurements and calculations comparing the 2010 yellowbook to the 2009 yellowbook (Excel version).

1 Comment

The Great Depression that began late in 1929 and extended through the early 1930s had a major effect on U.S. newspaper advertising expenditure.   U.S national advertisers spent less on newspaper advertising in 1938 than in 1919.   In 1938, spending of national advertisers in newspapers was only 57% of the corresponding advertising spending in 1929.

Greater concentration of national newspaper advertising expenditure occurred in conjunction with the collapse in advertising  expenditure.  In 1924, the top-100 national newspaper advertisers accounted for 22% of total national newspaper advertising spending.   In 1938, the top-100 national newspaper advertisers accounted for 55% of total national newspaper advertising.  This increase in the share of the top advertisers is not merely the product of mergers among top advertisers.  In 1938, the top-50 national newspapers advertisers accounted for 44% of national newspaper advertising, far greater than the 22% for the top-100 in 1924.

The current contraction in advertising spending in newspapers and magazines is more rapid than that in the Great Depression of the early 1930s.   The concentration in advertising spending that occurred in the Great Depression, if repeated in the current contraction, may help traditional newspapers.   Established personal relations and customized services are likely to be more important for securing the business of the largest advertising accounts.  Hence increased concentration in advertising spending probably favors traditional media over new media.

*  *  *  *  *

Data (for U.S.): national newspaper advertising expenditure by top-100 firms in 1924 and 1938 (Excel version); aggregate national and local newspaper advertising spending, 1915-1938 (Excel version); magazine advertising expenditure, apparently available only for 1913 to 1929 (Excel version).

1 Comment

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

Here's a Hollywood-based analysis of sign spinning:

Some of these spinners are quite good and can be fun to watch. But again, I just wonder how useful it all is. Personally, I think they should set up the sign on a stand and [hire] performers to do stuff in front of the sign.  That or women in bikinis.

That's dead-end thinking of traditional media.

Another whole world of value lives within persons.  Forget about supply and demand.  Have more fun with what you're doing.  Create more enjoyment for others.  The business will come.  And if it doesn't, you haven't wasted any time.

*  *  *  *  *

Thanks for contributions to the above video:

Sign Spinners by Douglas Galbi is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License.

Offer Comments

U.S. print newspaper and magazine advertising revenue is falling sharply.  In the first quarter of 2009, newspaper print advertising revenue declined 30% compared to the first quarter of 2008 ( $5.92 billion in 2009 Q1; $8.42 billion in 2008 Q1).  Magazine rate-card-reported advertising revenue, which does not take into account discounting likely to be more significant in more difficult economic times, declined 20% ($4.18 billion in 2009 Q1;  $5.25 billion in 2008 Q1).  These declines are associated with the current, major macro-economic contraction.  However, these sharp declines in newspaper and magazine advertising are not just a macro-economic effect.  They also indicate deep structural changes in the communications industry.

The current newspaper and magazine advertising revenue contraction is much more rapid and significant than that associated with the Great Depression.  From 1929 to 1931, newspaper and periodical advertising revenue fell only 23% and 25%, respectively.  The revenue decline subsequently accelerated and amounted to 31% and 42% for newspapers and periodicals, respectively, from 1931 to 1933.  Because periodicals have a higher share of subscriber revenue than newspapers, the more rapid decline in periodical advertising was offset to an extent by periodicals' greater revenue diversification.  The opposite comparative business model effect operates with the current more rapid decline in newspaper advertising revenue and newspapers' greater dependence on advertising.

Print media almost surely will remain a significant feature of the public landscape.   Now, however, every person can operate an astonishing, virtual multi-media printing press that the Internet provides globally.  How persons use print media undoubtedly will change enormously, and so too will the traditional business of newspapers and magazines.

Note:  Here's U.S. Census data concerning the revenue structure of U.S. newspapers and periodicals from 1880 to 2007.   Available also as an Excel file.

Update: Now includes data through 2008 from the 2008 Services Annual Survey, Tables 3.1.1 and 3.1.2.

Offer Comments

Looking at the top U.S. magazine advertisers from 1913 to 1929 shows a lot of familiar brands.  Proctor & Gamble, the leading magazine advertiser in 1913, was also the largest U.S. advertising spender from 1963 to 1986 and 1991 to 1996.[1]  Quaker Oats, Colgate, Kodak, Kellogg's Corn Flakes, Goodyear Tire, B.F. Goodrich, and General Electric, among other familiar names, were all leading magazine advertisers in 1913.  Having a brand name become well known is an enduring asset.  Astonishing new communications technologies developed since 1913 haven't obliterated brand names established with primitive print technology.

The distributional shape of print advertising spending among companies ranked by print advertising spending has changed little with the development of radio, television, and other new advertising media.  In the 1920s, the largest 75 U.S. national magazine advertisers, placed in descending order of advertising spending, are well described by a power law distribution.  The approximating linear slope of the distribution in log(ad spending) vs log(rank) space is -0.47.    The top 75 U.S. print advertisers in 1999 had a corresponding approximating slope of -0.54.   The scale of advertising spending has grown enormously along with the economy at a whole.[2]  But the shape of the leading advertiser distribution, like the share of total advertising in GDP and average advertising to sales ratios across companies, changed little from the 1920s to the end of the century.

This regularity is consistent with an analogous regularity in the company-size distribution and with company size determining advertising expenditure.  Economists have long recognized that power laws describe well company size distributions, as well as many distributions of symbolic goods, including given names.   If advertising spending is is a multiplicative share of company size (total sales), then the advertising spending distribution will have a distribution like that of company size.

*  *  *  *  *

Notes:

[1] Based on Ad Age's list of leading national advertisers, 1954-1997, summarized in Robert Coen, "Marketing elite," p. 128 in Advertising Age Special Issue, The Advertising Century (1999).  The leading advertiser from 1955 to 1962 was General Motors, and from 1987 to 1989, Phillip Morris.

[2]  Total U.S. advertising spending in 1999 was roughly 200 times greater than in 1913.  See the Coen Ad Expenditure Dataset.  The largest print advertising spender in 1999, Macy's Department Stores, spent roughly 100 times more than the largest print advertising spender in 1913.   I estimate the latter assuming that the Crowell list covers 40% of magazine advertising for the leading advertiser (Proctor & Gamble) and that magazine advertising account for half of total print advertising.

Data:  Here's the top 75 U.S. national magazine advertisers from 1913 to 1929 as an Excel workbook, with advertising expenditure calculated from Crowell's magazine set and reported in Crowell's National Markets & National Advertisers.  Here's data from Ad Age on the Top 200 U.S. advertisers in 1999 as an Excel workbook.  The top 75 print advertisers within this latter set probably is a good approximation for the top 75 print advertisers over-all.    In any case, those data are the basis for the 1999 estimate reported above.

Offer Comments