The effects of the abundance of supply and choices in the new internet marketing (:The Long Tail) By Dr Costas Kyritsis http://preveza.teiep.gr/kyritsis Department of Finance Technology and Education Institute of Epirus, Greece SDU International week Odense April 2008 The supply distribution The Pareto 80/20 rule The “long tail” effect Traditional retail: The retailer makes the majority of the money by selling more from a minority of “hits” New online retailers: The retailer makes more money by selling less numbers per title from a majority of “non-hits” The major part of area under the distribution curve……. …shifts under the “tail” instead of under the “head”. 2002:The decline of the “hitism” The search key-words distribution in Google is an empirical “Long tail” on the total “area” of searches Outline of the presentation What is the power distributions of the supply What is the Pareto Rule What is the long tail effect Why it is emerging What are its implications, for the modern world What are the relevant new online marketing techniques What is its time dynamics Examples Trends in online commerce More sales, more algorithm-fuelled recommendations. Growth of sales by positive feedback loop. Unlimited online selection is revealing new truths about what consumers want and how they want to get it in service after service. More facts about online commerce Many of our assumptions about popular taste are actually artefacts of poor supply-and-demand matching, a market response to inefficient distribution. This was the world of scarcity. Now, with online distribution and retail, we are entering a world of abundance. And the differences are profound. New facts about Books and Videos The potential book market may be twice as big as it appears to be, if only we can get over the economics of scarcity. Netflix has made a good business out of what’s unprofitable fare in movie theatres and video rental shops because it can aggregate dispersed audiences. Further facts about online commerce In a Long Tail economy, it’s more expensive to evaluate than to release. The slogan is: ”Just do it!” All this good news for consumers doesn’t have to hurt the industry. When you lower prices, people tend to buy more. So free has a cost: the psychological value of convenience. This is the “not worth it” moment where the wallet opens. Its effect is mainly on the “hits” Emergence evolution: Great Long Tail businesses can guide consumers further a field, by following the contours of their likes and dislikes, easing their exploration of the unknown. The power law distributions 1) Physics 2) The web 3) Macroeconomics 4) Ecology 5) Astrophysics etc When the size of the entities has a lower bound the distribution is the Pareto distribution rather than a power distribution In macroeconomics The income of the population, The size of the enterprises, The consumption size of energy of the enterprises, follow a Power distribution Also : The size of villages, towns and cities follows a Power distribution The famous Pareto rules “More than 80 % of the population possesses less than 20% of the total wealth” “More than 80% of the total power in the planet is consumed by less than 20% of the countries” “More than 80% of the problems in a company are created by less than 20% of the employees” “More than 80% of the result is attained by less than 20% of a total effort, the next 20% of the result requires more than 80% of a total effort” How a Power Distribution is Derived from Randomness and Growth It can be proved in mathematics that that all the next lead to a power distribution of the size of exponentially growing entities: Same initial size, different growth rate or age Same initial size and age, different growth rate Uniformly randomly different division of initial size (e,g. by a Poisson distribution) , same or different growth rate or age Linear scale graph y = C x-a Log-Log graph log(y) = log(C) - a log(x) The laws of the web: The size in pages and number of links of the sites in the Web follows a power distribution Source: The Laws of the Web Bernardo A. Huberman The MIT press 2001 Web logs distribution ranked by number of inbound links is an empirical long tail 433 weblogs arranged in rank order by number of inbound links. The data is drawn from N.Z Bear's 2002 work on the blogosphere ecosystem. The current version of this project can now be found at http://www.myelin.co.nz/ecosystem/ Mailing lists is an empirical power distribution Figure #2: All mailing lists in the Yahoo Groups Television category, ranked by number of subscribers (Data from September 2002.) The Technorati Popularity chart is an empirical “long tail” A formula for modeling Long tails Where x F(x) = the share of total volume covered by objects up to rank N50 = the number of objects that cover half of the whole volume α = the factor that defines the form of the function β = total volume Application of the model Box office sales in the U.S The 2004 book sales power distribution Table 1: Book sales in the U.S. in 2004 x = rank Cumulative volume (copies/year) Cumulative share (based on real data) Cumulative share [based on the model: F(x)] 10 17396510 2.6% 2.7% 32 31194809 4.7% 4.6% 96 53447300 8.0% 7.8% 420 100379331 15.1% 15.1% 1187 152238166 22.9% 23.4% 24234 432238757 65.0% 65.1% 91242 581332371 87.4% 87.0% 294180 650880870 97.8% 103.7% 1242185 665227287 100.0% 118.7% The 21st century and the power distribution The older smooth shape of the power distribution The anticipated 21st century new shape of the power distribution The new online retailers and physical retailers as they take deferent parts of the supply Online music popularity (Rhapsody 2005):The threshold of availability of traditional retailers and online retailers The abundance of choices in the online retailers compared to the offline retailers The abundance of new documentaries is only online Software supply Hollywood movies: The truncation of the distribution The truncation of the supply distribution seen in Log-Log Democratization of the production 1 Democratization of the production 2 Connecting supply and demand Producers, aggregators,and filters The “Long tail” rules Make everything available, no rejections Help people find it, with filters Lower the costs Let customers do most of the work One distribution and filter method cannot fit it all One price does not fit all One category does not fit all Share information Understand the power of freebies The participation of consumers in the production The “Long tail” is fed often by the consumers themselves, for that sake of getting public or for the love of doing it, rather than for money gain. This amplifies the collective intelligence in unprecedented degree. At the same time the collective opinion is correcting mistakes and increases the quality of the massively created products. An example is wikipedia, and the Linux operating system. Top-down and bottom-up messaging:Pre-filters and post-filters Traditional retail tries to “predict” demand as there is limited shelves space and high supply chain costs. These are the pre-filters as top-down messaging (experts minority wisdom) The new online retail market makes everything available and tries to measure the online demand and amplify it. These are the post-filters as bottomup messaging (Collective intelligence wisdom) Does the abundance of choices (tail) reduces the sales of the hits (head)? The sales of the existing hits do not decrease Nevertheless the industry is reducing the quantity of production and marketing of hits This has its good effect that the market is not so strongly manipulated by a small number of producers and advertisers (democratization) Huge advertising budgets of the hits are not necessary anymore, when online aggregators make more money from the non-hits. Dangers of the “hitism” Everyone wants to be a star If it is not a hit, it is a miss The only success is the mass success Low-selling=low quality If it were good it would be massively popular “Self-publish”=bad “Direct to Video”=bad “Independent”=The could not get a deal Does the abundance of the choices increases the demand or simply shifts it? The statistics show that consumers examine a larger number of products. This is made possible by powerful and smart online search filters. The demand is increasing in number of titles and items, although an individual consumer may or may not spend more money in total. The intellectualization of the population brought by the internet, shifts household expenses from basic products like food, and cloths to cultural like software, CD, MP3’s, DVD’s etc. The net results is a mutual benefit of suppliers and consumers as far as e-products are concerned Should prices rise or fall in the zone of abundance of choices (long tail)? The prices as well as the quality vary more in the tail than in the head. The average price is rather lower in the tail than in the head. But the affiliation commissions of the middle agent are higher in the tail than in the head. There is no real “Should”. The market’s balance and producers quality and expectations defines the prices in individual products The power of collective intelligence For the first time in history it is possible to measure the consumption patterns, inclinations,and tastes of the entire market of consumers in real timer while shopping. And just as quickly adjust the market to reflect them. We are leaving the information age and entering the contextual suggestion age Examples: Yahoo’s music ratings, Google’s Pagerank, MySpace Friends, Netflix users reviews, Amazon’s statistically based suggestions. Collective wisdom is a finer resolution compared to traditional professional individual critics Unlimited choices:A non-zero sum game In the fast changes of the distribution mechanisms of supply and demand it may happen that some parts may start loosing. This does not mean, in the macroscopic overall picture, that someone else is gaining their loss. In the overall picture the win-win phenomenon is the dominating and that which is driving the evolutional changes The variance of the quality increases in the long tail The demand for more powerful search filters is higher in the long tail, because of larger variance of the quality The time dynamics of the supply distribution:As time goes one the distribution flattens Review and Conclusions What is the power distributions of the supply What is the Pareto Rule What is the long tail effect Why it is emerging What are its implications, for the modern world What are the relevant new online marketing techniques What is its time dynamics Examples Sources 1 BestBuy.com Launchcast.com Salesforce.com Netflix.com Rhapsody.com soundscan.com dvdscan.com CRC standard probability and statistics tables and Formulae by daniel Zillinger-Stepen Kokosa Chapman&hall/CRC 2000 Sources 2 The Bass diffusion model (wikipedia) A practical model for analyzing long tails by Kalevi Kilkki http://firstmonday.org/issues/issue12_5/kilkki/index.html#author#author Repeating history, The Long Tail, and software demand by Dion Hinchcliffe @ 6:48 pm June 1st, 2006 http://blogs.zdnet.com/Hinchcliffe/?p=45 The sources of innovation by Eric Von Hippel http://web.mit.edu/evhippel/www/sources.htm Democratizing innovation by Eric Von Hippel http://web.mit.edu/evhippel/www Power Laws, Weblogs, and Inequality by Clay shirky http://www.shirky.com/ The Laws of the Web Bernardo A. Huberman The MIT press 2001