In the article “Hollywood Economics: How extreme uncertainty shapes the film industry,” author Arthur Vany explores how movies are in fact an information cascade.  He claims that movies that have lots of special effects, a big budget, and are filled with well-known big stars will attract more viewers because it is essentially an information cascade.  Movies that generate high opening box-office revenues act as a “signal” to other people that this movie is good.  Subsequently, other moviegoers choose to see these high-grossing movies.  This is essentially an information cascade. 

            Vany compares moviegoers to lemmings.  He claims that if you can generate great hype about the movie and do extremely well in its opening, then people who use box-office grosses as a means to evaluate the movie will act like “lemmings” and follow suit.  He calls the initial box-office revenues a “public signal” that leads to information cascades.  If the public signal is High, they will Accept.  If it is Low, they will Reject.  This leads to an Accept or Reject information cascade depending on whether or not the film received high revenues or low revenues. 

            Vany, however, makes a distinction between uninformative information cascades and informed information cascades.  The scenario I just described about using box-office revenues to decide whether to go to a movie or not is what Vany calls an uninformative information cascade.  It is considered uninformative because the only piece of information that moviegoers have is the choice of other moviegoers (whether a lot went to see the movie or not); they do not have any private information how the movie actually was.  However, Vany says that an uninformative information cascade in the movie industry is very unlikely since people do share their opinions about the movies and thus, people usually do have some private information.  Box-office grosses in the opening weeks may make an impact on the moviegoers’ decisions initially, but this uninformative information cascade is likely to be broken as moviegoers get more and more private information.  At that point, they make the choice whether to see the movie based on other people’s judge about the quality of the movie instead of just how much revenue the movie has brought in in comparison with other movies playing.  The private information that moviegoers receive can act to break the uninformative information cascade and now people choose to see a movie based on the private information they receive. 

            We can clearly see how this relates to what we have been studying in class about information cascades.  Moviegoers use a signal and either Accept or Reject as learned in class.  In this case, the signal is initially the opening box-office grosses and then gradually turns into private information about the quality of the movie.  If many people accept initially, this causes people to think that since so many people are seeing the movie, it must be good and leads to an acceptance cascade.  Similarly, if people do not go see the movie, people will act on the prior choices of moviegoers and conclude that the movie is bad and not go see it.  Private information then can also come into play to affect people’s decisions of Accept or Reject.  This is an obvious case of an information cascade. 

Link to Article:

The Google search engine has been a great tool for businesses to get their name out to consumers by biddign to have their name as top results from searches.

 However, a new tool by Goolge is angering businesses and have them questioning the monopoly Google is buiding for search engines.  This tool allows people to search website pages without actually going to the pages. For example, if one were to type “New York Post,” the search returns a list of links and the option to stay on the Google site to search within the New York Post site. The problem businesses are having with this new feature is that when a user types something in to the secondary box, Google returns the results with ads for competing companies on the top and sides of the page. Therefore, to avoid having users or potential customers go to competing sites, the company would have to pay additional money for ads on Google to ensure that users stay on their pages. These companies would also need to be the highest bidders for ad space to ensure that users see their names first in the searches instead of competing companies, therefore spending even more money on Google.  

Furthermore, since users are able to bypass many of the pages on the company’s site, the ad’s on the site are essentially useless. Companies can directly place ads on Google if they feel most users are no longer visiting actual websites, but searching on Google. Sites like Barnes and Nobles that suggest products to users based on past purchases will lose the opportunity to advertise these products to consumers if they bypass the company’s pages.

 Google argues that this new feature will benefit companies because now users will be better able to navigate to their destinations. Moreover, companies without a search option on their pages or with an inadeqaute one will greatly benefit from having the Google search engine apply to only their pages.  

Some argue that big companies do not have to fear since most users will remain loyal and not be swayed with simple ads to the competition.  Therefore, they would only lose a very small percentage of users.             

Many predict that users will find the new feature helpful and a better way to search webpages than actaully going to the site.   

References 

Tedeschi, Bob. “A New Tool From Google Alarms Sites.” New York Times. 24 Mar. 2008.             <http://www.nytimes.com/2008/03/24/business/media/24ecom.html?pagewanted=1&_r=3&ei=5087&em&en=7af4252e4367999c&ex=1206504000 >.

We have talked about a few ways to increase Google rankings of an existing page for common queries in class. A “Googlebomb” refers to a prank people attempt to cause someone else’s site to rank high for an obscure or meaningless query. An example of such prank is a search for “miserable failure” on Google brought up the official George W. Bush biography from the US White House web site a few years ago. A Google bomb is created by creating many sites that link to that page using consistent anchor text. Google originally would not alter the search result because it wished to preserve the integrity of its search engine, but in January 2007, Google changed the indexing structure; it tries to only display the comments and articles relating to the topics to minimize the effect of the prank. As of now, a search for “miserable failure” no longer return the biography of George W. Bush as the first hit.

Even though Google bomb was just a prank, in a way, it provided a method to create misleading information. Google’s methodology for ranking web pages are often good at estimating what you are looking for provided the keyword, but sometimes the return results might not be what you are really looking for. Even though it might have already tried to minimize the effect such prank, but it is still possible to spread the misleading information by Google bombing. If some misleading information relating to a keyword is ranked highest, it is more likely people click on the first link returned from the search and start reading about it, then the person might start believing in the misleading information. It is also possible to Google bomb other pages that may also provide the misleading information to push down the more relevant information to the second page so that people are less likely to see, then the likelihood of people seeing and believing the misleading information increases a lot.

http://searchengineland.com/070125-230048.php

As more social networking tools, like Facebook, begin to take over the internet, we are going to see a change in the way advertising works. Facebook has already taken a stab at this and has broken through the boundaries of conventional advertising of click through rates that we have talked about in class. Instead, they are concentrating on applications and sponsorships in which advertisers can engage the use of their potential customers. The applications that the advertisers invest in are geared towards the individual who will use them and incorporates some strategy that will catch the person’s attention and enable them to interact with the advertisement.

Michael Kingdon wrote this article (the first link) on engagement advertising and what we can expect in years to come. According to Kingdon, we can expect to see Facebook joining Google as one of the main connections into the digital world. Social networking sites will become more prominent as people strive to keep in touch with one another in an easier fashion. As for advertising, Kingdon says, “Although search will still be a powerful tool, the big money in digital marketing will be in getting people to engage with a brand.” Unlike today where people compete to buy an ad slot that has the better frequency of display or a slot with high internet traffic flow, companies will be trying to create unique ways of engaging the consumer in their advertisement. There will be no “standard ad” that will be shown to everyone on the side of Google or on a banner across the top of your email page. Ads will be targeted directly to the consumer who is viewing them.

In class we have talked about finding the best matching for advertisements based on calculating the value that the advertiser would receive based on the prices for those slots. Pretty soon, this method will become obsolete as advertisements become more advanced and high tech. It will no longer solely depend on how many impressions (the amount of times that one page is shown to one person) that the advertisement receives, but also how well that advertisement can engage a person’s interest. While currently Facebook’s applications are free for the creators, Facebook has created another advertising opportunity for companies, in which they become a “sponsor”. Here they are offered advertising on Facebook’s group pages, there own page where they can control unlimited pictures and information, and stories that pop up on user’s home pages. These have become exceedingly effective forms of advertising, yet come at quite a cost. To become a Facebook sponsor, you must pay $150,000 every 3 months. Big companies and entertainment like Victoria’s Secret and the Dave Mathews Band have jumped on this opportunity, but it will take some time for the smaller ones to be able to compete. While these smaller companies are contemplating which mode of advertisement to go through, they could create a free application. Now the only challenge they will face is advertising this form of advertisement (the application). Since there are so many applications on Facebook, it takes a very engaging and innovative application to make a profit. In the coming years, we will have to rethink the terms of which we price our advertisements as the advertising industry becomes more advanced.

http://gigaom.com/2007/11/07/future-of-online-advertising/

http://valleywag.com/tech/online-advertising/facebooks-secret-rate-card-284029.php/

Greg Mankiw, an economist at Harvard has posted to his blog, an application of Bayes rule to the 2008 Presidential Election. He finds election probabilities from the website Intrade and defines the necessary variables as follows:

“Let A be the event that a candidate wins the general election, and B be the event that a candidate wins his or her party’s nomination. Tradesports gives us the betting market’s view of P(A) and P(B). It is a safe assumption that P(B / A) = 1, that is, a candidate can win only if nominated. We can then use Bayes theorem to compute P(A / B), the probability that the candidate will win the general election conditional on being nominated.”

Mankiw calculated the conditional probabilities at the time of his post, in November of 2006, so I have updated the calculations to show the current probabilities.

According to Intrade, the current probabilities of the three main candidates to win their respective party nominations (as of 3/29/08) are as follows:

Obama: 80.1%

Clinton: 18.5%

McCain: 95.4% (His nomination has yet to be formally announced)

The current probabilities of Presidential election (as of 3/29/08) are as follows:

Obama: 47.9%

Clinton: 11.9%

McCain: 39.7%

Given these numbers we can now determine the conditional probabilities P(A/B) by taking P(B/A)*P(A) and dividing it by P(B)

These above numbers give the following conditional probabilities (the probability that the candidate will win the general election conditional on being nominated):

Obama: 59.8%

Clinton: 64.3%

McCain: 41.6%

According to the market, it seems that even though Obama is more likely to end up President than Clinton, Clinton is more likely to win the general election if she is nominated. So although it may seem as if Clinton is a long shot to win the Presidential election right now, if she can win the Democratic nomination, her odds greatly improve. Either way, it is clear from this analysis that with a little over seven months to go until the election, the markets are very much undecided as to our next president.

New data from ComScore Inc. shows that Google’s click-through rate growth has almost slowed to a stop. The rate increase for January was zero when compared to January 2007 and only a 3 percent year over year increase in February. These rate increases are abysmal when compared to prior rate increases only a few months ago that were showing double digit percentage growth. In web advertising money is made on the number of times consumers click on advertisements. Different advertisements at different places on the web page have different click-through rates because of location and prominence on the page. Using our models from class, smaller click-through rates will result in decreased values from the advertisers and resulting smaller profits for Google. Google needs to maintain increasing click-through rates to keep on pace with its forecast revenue and profit growth. Hence, these stalling click-through rates spell trouble for Google.

Google has responded to this data by stating that these decreased click-through rates are because they are working on creating more value from each click. For example, Google has decreased the space around a word that would create a click, thus making sure more of the clicks are actually intentional. Google hoping that advertisers will be willing to pay more for advertisements with smaller click-through rates because of this increased value. However, financial analysts are unsure how long it will take for Google’s per click revenue to adjust to their new value added advertisements. An advertiser will be unlikely to pay more for an advertisement unless the new value is apparent. Google will have to prove this to their clients.

However, Google’s new value scheme may not be the whole story in their decreased click-through rates. If consumers are spending less, it is reasonable to assume they will be clicking less on advertisements. An advertisement for a vacation will have more responses when they economy is at a point where consumers have excess cash to spend and are not worried about a recession. A recession will also mean that advertisers may cut their budgets. This was the view of one Lehman Brothers analyst who cut his 2008 profit targets for Google citing decreased click-through rates as well an impending recession.

http://www.msnbc.msn.com/id/23831823/

As we discussed in class, there are many things today that are examples of information cascades.  One of the most prominent issues in our image self conscious society is about weight and diet.  This blog post from a columnist  from the New York Times talks about the obsession with low-fat diets and weight-watching. The columnist is reflecting on a book that has just disproved the common belief that low-fat diets prolong your life.  The article references an economist at Brown who states that most choices are binary.  He says that people either choose to do something or not to do it. This economist, Dr. Welch, explains that logically people reach the truth of low-fat diets by acting on their beliefs. However, because of this binary “accept-reject” state of mind, he further explains, people have no option to show their belief but only to illogically accept or reject the current belief of the correct amount of fat in a diet.

            There are many other aspects about diets that have cascading effects.  I remember a long time ago when scientists argued that chocolate was actually good for you.  People didn’t care about the real information, they could either accepted this information and eat all the chocolate they wanted or reject and not change.  People wouldn’t care about the fact that this new finding only refers to dark chocolate and that it lowers blood pressure and has healthy antioxidants.  They wouldn’t realize that the sugar in such chocolate is still not good for you and there are other ways to have the same helping effects. 

This is the same as the “accept and reject” aspects of information cascades that we have studied in class.  This binary response is the cause of such illogical blunders that occur in cascades.  This kind of decision-making causes people to ignore the information and to not choose for themselves.

            Welch finally also comments on the power of  “entrepreneurs”.  He says that these a re special people who usually have enough knowledge to follow their own beliefs and make their own decisions.  Welch says they are valuable because they are able to overturn certain cascades in society.  I, however, can also see them as a potential threat to creating even more cascades.  Suppose an entrepreneur makes a mistake or a crowd misunderstands his decision, it is like to cause even more cascades.  

 http://tierneylab.blogs.nytimes.com/2007/10/09/how-the-low-fat-low-fact-cascade-just-keeps-rolling-along/ 

It took me a while to find a topic related to information cascades that wasn’t already in the blog, but I eventually came across a 16 page research paper by John H. Miller and Scott E. Page of the Sante Fe Institute about modeling standing ovations. It turns out that an information cascade model similar to the one that we have learned in class can be applied to standing ovations. Note, however, that their information cascade model does differ in some respects from ours, and that it is in general more complicated than ours.

The standing ovation problem is the following: At the end of a presentation, lecture, or other event, there is applause from the audience, and some audience members may or may not choose to stand. How can we determine if the audience will begin to give a standing ovation, or if they will simply remain seated and the applause will calm down. The application of information cascades to this problem should be fairly obvious: If audience member(s) around you choose to stand, your choice of whether or not to stand will probably be influenced by their decision(s) to stand. You probably know from experience that if everyone around you is standing and applauding, you feel pressure to stand as well, even if you didn’t enjoy the lecture or performance.

One version of the standing ovation information cascade model is as follows: Audience members choose to Stand or Sit (referred to as Good and Bad in our class discussions). An audience member can look left and right (to see the people next to him or her) and can see the people in the row directly in front of him or her (a simplified model, but there have to be some simplifications to reduce complexity). An audience member will change his decision from Sit to Stand if there are at least two more people that he can see who are standing than people he can see who are sitting (and vice versa from Stand to Sit). Consider the following example:

(three rows of three audience members, and their actions after the lecture/performance)

Stand Sit Stand

Sit Sit Sit

Sit Sit Sit

In this example, the person in the middle of the front row (he can only see the people to his left and right) chooses to Stand because the two people he can see are standing as well. Then, the people on the outsides of the second row stand up because they see three people in the front row standing, while they only see themselves and the person in the middle of the middle row sitting. Then, the person in the middle of the middle row stands, and so on, until the entire audience is standing. Here are the steps in the cascade:

Stand Stand Stand - Stand Stand Stand - Stand Stand Stand

—-Sit Sit Sit ———> Stand Sit Stand —> Stand Stand Stand

—-Sit Sit Sit ———— Sit Sit Sit —————- Sit Sit Sit

-

— Stand Stand Stand — Stand Stand Stand

-> Stand Stand Stand -> Stand Stand Stand

—–Stand Sit Stand —- Stand Stand Stand

As you can see, even though only two people are initially standing, the entire audience is standing at the end. One important difference to note between this information cascade and the one we discussed in class is that decisions are not made in sequence (anyone can change their decision to stand or sit at any time).

The paper goes on to discuss other ways of modeling standing ovations, including conformity, growth/coordination, and diffusion (see link below for more details). It concludes by pointing out that there are so many dynamics in a real auditorium and a real audience (people leaving the auditorium, people going with their friends, people sitting on different levels (e.g. balconies), where the entrances are located, the fact that people more interested in the talk tend to sit in the front, etc…) that predicting if and how a standing ovation will occur is nearly impossible in many circumstances. However, the paper was still an interesting look at another application of information cascades.

Reference: http://zia.hss.cmu.edu/miller/papers/ovation.pdf

http://www.nytimes.com/2007/10/09/science/09tier.html

http://tierneylab.blogs.nytimes.com/2007/10/09/how-the-low-fat-low-fact-cascade-just-keeps-rolling-along/

The article that I’ve chosen for this post (first link) is about how the medical world was duped into the cascading idea that a low-fat diet would lower the risk of heart disease. John Tierney, a science columnist for The New York Times, cites Gary Taubes’ book, Good Calories, Bad Calories to explain the steps that led to the commonly accepted idea, which has never been proven, and yet was endorsed by numerous health organizations, officials and doctors all the way up to the Surgeon General in 1988. It started with the confident voice of Ancel Keys, a prominent diet researcher of his time, who became convince in the 1950s that an increase in the number of cases of hearth disease in America was correlated with an increase in the amount of fat in Americans’ diets. He supported his theory with a weak study of a few countries’ populations (weak because, as Tierney points out, if he had used more countries – for which data was available – he would not have found such a correlation). Then, in 1960, Keys and an ally were on an American Heart Association committee that issued a report asserting that a low-fat diet would lower the risk of heart disease. This report made the theory popular wisdom, and in the 1970s, a senate committee, led by Senator George McGovern, issued another report warning Americans to lower their fat intake. This senate report was only supported with input from one nutritionist, Mark Hegsted, who, in turn, was hired by an assistant agriculture secretary to draw up a set of national dietary guidelines for the Department of Agriculture. In multiple instances, doubters emerged, but there were too many supporters of the theory for them to be able to convince anyone of their case. Ultimately, the idea became a commonly held belief, and according to Senator McGovern, “92% of the world’s leading doctors” endorsed it; however, as Dr. Edward H. Ahrens Jr. responded, many of those doctors were relying on second hand knowledge.

This phenomenon is a perfect example of information cascade, which we discussed in class. The author of the article, John Tierney, subsequently made a blog post (second link), which essentially explains what an information cascade is, and how it is that dietary research so often gains enormous support and tons of followers, only to later be contradicted and/or proven ineffective. To support his analysis that large groups often reach a ‘consensus’ without everyone fully understanding the issue, Tierney cites Brown University economist Dr. Ivo Welch. Welch explains that it all comes down to the fact that many choices are binary. In the case of diets, someone can endorse it or not, but they cannot partially endorse it. Therefore, after a few people have chosen to endorse a diet, it is very easy for many more that follow to believe that the diet is a good idea because they assume the endorsers cannot all be wrong. Welch explains that sometimes a cascade can change direction, for instance when a few people introduce a new idea that is opposite of the commonly held belief, and subsequently more and more people switch their beliefs, he also notes that this can occur for either the right or wrong choices. This explanation by Tierney and Welch ties in precisely to what we have learned in class and in the text about information cascades.

As more people use search engines, Search Engine Optimization (SEO) has become a profitable way for website owners looking to boost their ranking in organic search results. Some companies have made SEO into a business by learning the intricate details of a particular search engine’s algorithm and tailoring a client’s webpage for prime placement. Clearly, the focus of most SEO companies is the 500 pound gorilla in the search engine market: Google. With Google’s market share doubling the combined shares of all its competitors, website owners prefer to ranking highly on Google over any other search engine. SEO has split into two main categories: white hat SEO, and black hat SEO (sometimes known as ‘Google bombing’). White hat SEO is optimization designed to be a lasting rank improvement by not breaking any search engine rules or terms of service. Conversely, black hat SEO involves deceptive techniques that violate these rules. Google bombers attempt to keep their ranking improvement as long as possible, before the search engine takes notice and bans the offending site.

A recent blog post by SearchEngineWatch.com chronicles the fine line Google is walking by selectively permitting or banning Google bombs. In most cases, Google will ban deceptive practices. For example, BMW Germany attempted to improve its search ranking by using a secret page loaded with key words that only Google crawlers could see. When Google discovered this blatant violation of their rules, they lowered the PageRank of bmw.de to 0, effectively putting the page last in any search. At the same time, a Google search for the term ‘Jew’ used to return an anti-semitic news site as the top result. Google found itself in a public relations nightmare by allowing the link to stay, claiming the site did not use deceptive practices to increase rank. When pro-Jewish groups found out, they organized an extensive Google bombing campaign that employed massive linking to the Wikipedia entry on “Jew”, pushing it to the top of the results. Was such a ranking due to “natural” patterns and did it fall within Google’s terms of service? Unlikely. However, Google instantly found itself with an easy exit to the controversy. By allowing the Google bomb to stay, Google no longer had to face the backlash of an offensive first search result. In fact, Google placed a disclaimer on the top of the results page indicating that the anti-semitic website did not reflect the views of the company.

So, how should Google deal with Google bombs? Should it adopt a blanket policy by banning all black hat SEOs or should it selectively permit certain bombs, possibly those that could benefit its image? By what criteria can Google make such judgments? These questions cannot be answered by a complex algorithm, thus revealing a very human side to the world’s largest search engine.