http://www.shirky.com/writings/powerlaw_weblog.html

Classroom discussion has brought up the linking between weblogs (one of which I am linking myself in this post) as an example of a network governed by a power law distribution. The above article describes how this power law distribution arises, attributing the development of a seemingly more connected “community” within the larger network to inherent inequalities in a large, diverse, and free network. By not accounting for this structure by referring to individual behavior, it becomes easier to understand potential counter-intuitive examples of power law behavior in web logs.

As an online social system becomes larger, there are more options for people to express their preferences. As the number of options increases, the power law curve becomes even more extreme. This is perhaps the opposite of what one may think would happen in a situation where more options are present - that when people are presented with more opportunities for searching and linking the curve will generally flatten out as the most important aspect governing linking, a person’s time, is diminished among these options. However, the opposite occurs.

Though it is an easy model to consider, linking between blogs is not governed by uniform linking probably to every site, which would result in a relatively flat curve. People’s choices do greatly affect one another, and in many ways, a simple model of a blogging network could resemble a type of information cascade. One user would choose to link blogs with no view of other users’ signals. A second user would see this first user’s signals in addition to his own. Users farther down the line receive their own signals regarding different blogs, but will be much more likely to link to previously linked blogs based on the signals he gathered from other users. Of course, the probably of a blog being linked is also due to its visibility among users, and thus a simple cascade model does not reveal a very clear picture of blog linking because it is a function of a user’s signals regarding the blog in addition to that blog’s visibility among users. However, when considering that the linking of a blog both increases its positive signal among users in addition to its visibility, its is easier to see how even small differences among blogs (quality, preferences, solidarity) can greatly change their popularity once they begin to be linked.

Finally, this article attempts to address the potential inequality in the weblog world. This blogger clearly believes that the system is relatively fair due to the fact that there is true freedom in the weblog world, users are relatively equal, popularity does not result from a cliquish preference but a reinforcing distributed approval, and the lack of no real “A-list.” Though it may appear there is a true A-list due to that fact that 2/3 of users may fall below the mean, there is no true cutoff. The power law distribution is indeed continuous, and thus any set designations are arbitrary markers. This distribution, however, is not static. When there are only a few blogs receiving most of the new traffic while most blogs are getting below average traffic, the distribution must become more extreme. The rich do get richer. In some cases, those top web loggers may join mainstream media in that there is no way for the blogger to answer all posts, and thus becomes a mechanism for material distribution. The bottom majority will become conversational, where having your 3 closest friend read a post is more an achievable goal than having 3 random users read your blog. This is an aspect of weblogs that sites such as LiveJournal have embraced. Especially as time goes on, a growing number of blogs will becoming relegated to this sphere of the weblog world.

http://www.cnn.com/2007/TECH/04/09/googlingyourdate.ap.ap/index.html

The article “googling your date” is about how social network sites are affecting the lives of people who are trying to discover information about a date via the internet. The article explains how the rules of dating have changed for better or worse. With websites like www.facebook.com and www.myspace.com, potential daters are looking up personal information to get a better handle on their dates or their potential dates. With just an email address, you can glean all sorts of personal information about the person before you even meet them. Information attained could be in the form of pictures, hobbies or news about the person. While it is positive to see the same hobbies listed in a potential date, not all information is valuable information. Katie Laird from the article says “Don’t Google what you can’t handle.” You might find a date with a potential fetish that might turn you off, or vile pictures that are inappropriate The question at hand is social networking helping or hurting dating?

I believe social networking helps dating because there is no actual harm in reading a potential date’s online profile. Presumably, the profile you are reading is what the actual potential date wants people to read about them. Googling helps dating because it can help weed out the people you KNOW you won’t have a connection with and you can avoid wasting your time and money. Even if the profile is exaggerated or even a lie , the dater will eventually find out the true nature of the person with the face to face interaction. An interesting point in the article pointed out that some people do not want to admit to googling their date beforehand for fear of looking like a “stalker”. While this might make things awkward, I believe there are more advantages than disadvantages when it comes to “googling when you date.” In today’s environment, if a person has personal information posted on these social network websites, they should expect that a potential date would read the profile in advance. In fact, I would want someone to google me in advance so that I am not wasting my time on someone who may have no interest in me.

In an extremely interesting article in the Washington Post, an experiment was set up to see if commuters in a D.C. subway station would stop and listen to a musical performer. However, this was not any musician; the beggar leaning against the wall with his violin case open in front of him was violinist Joshua Bell, an internationally acclaimed virtuoso who had just recently filled Boston’s Symphony Hall. The situation was set up to see if commuters would take time out of their busy schedules to listen to one of the best violinists in the world, playing a $3.5 million dollar violin. A hidden camera was set up to record the reaction the commuters would have to a musician that was clearly superior to the normal quality of street performer.

There were concerns by the responsible parties that this event might cause something of a commotion in the Metro station. If only a few people had recognized the world-renowned star, they would stop and begin staring, inevitably causing a cascade that would create a larger and larger crowd.

“In preparing for this event, editors at The Post Magazine discussed how to deal with likely outcomes. The most widely held assumption was that there could well be a problem with crowd control: In a demographic as sophisticated as Washington, the thinking went, several people would surely recognize Bell. Nervous ‘what-if’ scenarios abounded. As people gathered, what if others stopped just to see what the attraction was? Word would spread through the crowd. Cameras would flash. More people flock to the scene; rush-hour pedestrian traffic backs up; tempers flare; the National Guard is called; tear gas, rubber bullets, etc.”

In fact, the opposite ended up occurring: out of the more than 1,000 people that passed during the 45 minute performance, only 1 person recognized Bell. Unfortunately, the commuter arrived at the very end, however her presence in front of Bell attracted more people to come over and watch the final minutes. While the reason that more people did not stop and watch the musician is not clear, it might have something to do with the fact that there was not enough of a signal to the other commuters that something extraordinary was going on before their very eyes. Each individual, even if he had a signal that this performer was a unique talent, did not have the necessary signals around him from the other commuters to corroborate his feeling, thus preventing a cascading effect.

Videos from the performance are posted, along with the full article at the Washington Post’s website.

Two major topics were discussed under the “The World-Wide Web and Information Access” of this course: keyword-based advertising and modelling the effectiveness of network communication through the concept of hubs and authorities. The former discussed the sale of online advertisements, i.e. links on a Google search page, using the familiar second price auction system taught earlier in the semester. By modelling the willingness of buyers to purchase a prominent location for their online ad and using a second price auction system, market clearing prices can be determined as well as profits sellers earn. The latter demonstrated a model for analyzing the effectiveness of web connectivity by assigning hub and authority scores to webpages. One conclusion that came out of this discussion, as shown in homework 4, is that mutually reinforcing pages will score higher in the long run compared to pages that just have high in-degree.

What is neglected in this overview are the methods used to create levels of prominence and to enhance web connectivity. Creating levels of prominence on a search page is fairly simple: higher paid ads are displayed higher up in the list, with perhaps premier ads placed in a colored box at the top of the page. For other layouts the design may not be so straightforward. An example of a more complicated scheme can be seen with the website for The Cornell Daily Sun, www.cornellsun.com. Ads are not sold and displayed on this webpage, however as a for-profit organization, readership is essential to the organization and the effect of content layout on the website can be modelled in terms of the search engine example. Content can be laid out with varying degrees of prominence, and as a search engine would want to order its links to maximize profits, so this organization would want to lay out its content to maximize readership, and with it profits. One obvious method to achieve this effect would be to place the most popular or important stories of each department at the top of their section and pair it with a photo to clearly delineate that this story is the most eye-grabbing one for the day. This design would attract the greatest level of readership. More subtle methods would be to order the departments on the website by their level of importance, in this case: News, Sports, etc.

Enhancing web connectivity on the website also promotes readership. Mutual reinforcement of newspaper content on the web can be achieved by presenting content through multiple media. Recent developments at the Sun, namely podcasts, video clips and online news blogs, demonstrate this effect. They are used to achieve a greater level of readership by tying in viewers more closely towards articles. Podcasts (audio clips of stories and other varying content) are accessed on the website by the icon link that looks like a sun blaring out audio. Viewers tend to be more attracted towards visually appealing images than text and are easily inclined to access the link upon entering the homepage. Those who may dislike reading newspaper articles on a website may be allured by the idea of listening to an audio clip of the story, and of these some may decide to go back and read the article online. Regular subscribers to the website may also be interested in having articles read to them and so in addition to reading articles online would choose to listen to some of the podcasts. Thus, the versions of the same content mutually reinforce each other and enhance readership. Video clips and online news blogs provide an extra dimension to article content and reinforce them in a similar fashion.

To summarize, the design of webpages is essential towards maximizing profits, and the manner in which design plays a role can be analyzed through the models for ad auctioning and web connectivity.

Herd mentality and information cascades seem to be the buzz words nowadays: every major online social forum, from YouTube to Facebook to Digg, has been analyzed as the result of the many following the opinions of the first few. Once enough people decide that something is or isn’t popular, the rest will follow in their footsteps, oftentimes disregarding their own private information or preferences. When enough people start to follow the actions of the few, disregarding their own private information, their generally coherent and synchronized reactions and behaviors lead to the idea of herd mentality. Apart from social forums, the information cascade model (which tends to lead to the formation of the human herd) can be applied to anything involving sequential decisions where new decision makers see the decisions (but not the information) of people before them, and a more or less binary decision (such as buy/don’t buy, or digg/don’t digg).

Stephen J. Dubner, in his March 21st, 2007 blogpost titled Herd Mentality? The Freakonomics of Boarding a Bus, discribes an example of herd mentality in the simple process of boarding a bus during his morning commute in Manhattan. Like most New Yorkers, Dubner takes public transportation, specifically a crosstown bus, to get his daughter to her nursery school and himself to work and inevitably suffers the overcrowding during the early morning rush hour. The bus stop closest to their house happens to be the closest one to the local subway stop, meaning it suffers from particularly heavy overcrowding because of the spillover from people taking the subway and coming above ground to get across Central Park. The obvious problem is that the more people there are trying to get onto a bus, the less of a chance Dubner has for getting onto the first bus that comes by, and an even lesser chance of getting a seat for his daughter. As a solution, Dubner and his daughter have started walking to a bus stop further west (in the opposite direction of their final destination) because it is much less crowded and greatly increases their chances of getting onto the first bus that goes by and actually getting seats. While the new bus stop requires some walking to get to, Dubner finds the benefits much outweight the cost of walking and cannot understand why none of the other bus riders bother to walk to the new bus stop and try to outsmart the rest of the herd. He proposes several possible reasons for this behavior, but expounds that the strongest reason is that the riders at the overcrowded bus stop are part of the herd, and find comfort in being a part of the whole, so despite the benefits to the individuals of breaking away from the pack and walking a few extra steps to get to the next bus stop and actually get a seat, everyone tags along “because if everyone else is doing it, it must be the thing to do.”

While certainly the herd mentality aspect of the story is quite apparent, the information cascade is not as obvious. First of all, the idea of the information cascade depends on everyone knowing about the second bus stop and therefore having to make a choice between standing at the overcrowded one or walking to the farther one. Essentially, the information cascade occurs with the first few people. The first person to get to the bus stops see no one else and decides to stay there since they’re sure to get onto the bus. The next person sees the first person at the bus stop and decides that this must be a good place to wait for the bus. The next few people who sequentially get to the bus stop see the small group waiting, which often in the city means that the bus or train is sure to come soon, and decide to wait at the first bus stop because they see the decisions of the others and value them above their own logical idea that the other bus stop might be less crowded. This turns into an information cascade, as the more people that are waiting, the more will join them, thinking that there must be some reason the people already there chose this bus stop over the farther one (aside from the cost of the walk in between).

Dubner asks for other suggestions of herd mentality and the inherent information cascade. Quite a lot of responses were posted, and interestingly enough, the vast majority samples had nothing to do with social networks or stock markets, but with everyday life. For example: someone discusses the herd mentality of choosing a line at the supermarket, where people will get onto longer check-out lines despite short ones close by, because they see the decisions of those before them and possibly infer that the existence of this longer line must mean that the shorter one is about to close or is otherwise worse. Other examples include people waiting in long lines at any car toll, or at border crossings between the US and Canada: there are always one or two car lanes which are far longer than the rest, and until a cascade of cars start switching to the shorter lanes, people stay in these longer lines, despite an obviously better options. I thought of an example specific to my own experience. I am taking the Wines elective this semester, which meets every Wednesday in Statler Auditorium. There are around 700 students in the class, all with cafeteria trays and wine glass cases. I sit on the first level, which holds approximately 500 students. When class lets out, there is a mad rush for the main doors at the back of the auditorium on the first level because these doors lead to the main stairway. Every class period, the majority of the 500 students run for these two small doors, causing a 10 minute holdup while they file out in pairs. A few students have started taking the exits in the front of the class, which lead to a less central location to the building, but are just as close to some exit as the main doors. However, despite the fact that we’re several months into the class, the number of students taking the exits at the front of the room has not significantly decreased, leading to the same question brought up by Dubner. Just like the information cascade inherent in the bus stop situation, what occurs in our class is the first few people make the decision to go to the back doors because they are close to the central stairway, the next few people see these people choosing the back and follow them, because they see the first few people’s decisions and infer that the back way must be better and faster, and this develops into a full blown stampede to the back doors, despite the availability of other, less crowded exists.

I think examples of information cascades and herd mentality are abundant in everyday situations, and aren’t limited to the formal networks discussed in class. While the examples in class are easier to form proofs around (since they have been studied), it is interesting to try to find unconventional examples and see how the models apply there. It seems that the occurence of information cascades can be somewhat generalized to situations involving a decision that in part relies on chance and can result in a potential loss (such as the loss of time, money, social connectedness). The decisions have some cost to them (in the implied loss) and it is psychologically easier for people to follow the decisions of others than to make their own, because if you follow someone else’s decisions and they turn out to be wrong, there is less personal blame. Much like every other analysis of information cascades and herd mentality, Dubner’s story points to the individual losses that result from the many following the decisions of the few. What these analyses’ don’t consider, is the personal benefit of not being responsible for the losses caused by following the herd. Additionally, they don’t consider that perhaps people have some foresight and realize that if everyone starts going to the second stop to get onto the bus earlier, then the same situation would repeat itself. What’s the equilibrium solution to these real life information cascades then?

In network cascade theory, we see that people are like sheep, ie. they follow the current popular opinion. Whether it be what they eat, what they wear, where they live, what car they drive etc, most people will tend to follow the crowd. In thinking about this, there are two different categories of network cascades that we could think about. The ones where the act of making one choice would have a long term effect, and the one where there would only be a short term effect.

 For example, choosing a restaurant to eat. Perhaps you were told that restaurant B was better than restaurant A, but everyone you saw went into restaurant A. Which one would you choose? Network cascade theory says that we would choose A. But if we are to think about the decision carefully, this is a choice with a short term effect. Perhaps the restaurant you eat at was really good. The next day you’ll have memories of the good food, and perhaps you may go there in the future. Or if the restaurant was bad, you will also just have a memory (but it will be bad) and you probably would try something else. But either way, (unless the food made you sick) the effect of your choice does not have any major ill effect on you should you choose wrong. Thus you may not think twice about following the crowd.

 However, in a decision with a long term effect, such as buying a car, following the crowd may not be such a great thing to do. Suppose you felt that car A was better, but all your friends bought car B. Perhaps car B is the ‘in’ car and car A is just simply “uncool”. But this choice is not a choice that we can easily make. We cannot simply just buy car B because “everyone else is doing it”. We have to think about the costs of buying one car over the other, is one car more fuel efficient, is one brand better than an another, which car is cheaper etc. This means that its not necessary that the network cascade theory would work so well in such situations.

 What prompted this thought was that recently a group of friends started playing World of Warcraft (WoW). It seems Blizzard was quite smart about this, but they provide a free 10 day trial subscription to the game. Thus once one friend started, it was simply quite cheap for another friend to join in, and pretty soon there was a network cascade. This choice made can be categorized as a short term effect, because the trial only lasted for 10 days and the trial was free. There was little cost to the people playing (apart from time) and they were able to follow the crowd and join in.

However after the trial ends, the decision whether to pay for the subscription switches to a long term effect decision. Whilst the subscription fee is not high, the time (possibly) spent playing is rather expensive and thus people would have to think seriously before following their friends in paying for the subscription. Of course, the addictiveness of WoW is probably what tips people over the line to subscribe to the game after the trial, but before subscribing, one would seriously consider the time spent and the money spent if one were to subscribe.

http://www.jstor.org/view/00346527/di990706/99p03304/0?frame=noframe&userID=80fdba30@cornell.edu/01cce4406400501bc33

0d&dpi=3&config=jstor

The Economics of Rumors

In this work, the authors explore a family of models of information transmission processes in regard to an investment opportunity in which expected returns are known to few people. Each individual has a cost of undertaking the project and only receives information about which other agents invested. From this information, he would like to infer the expected returns of the project and decide whether or not to invest. Specifically, the basic model has two assumptions: (i) each person chooses to invest based only on what the previous investor decided and (ii) only the first investor knows the opportunity exists, and has a private “signal” of what the expected returns are. Through various relaxations on these assumptions, the authors obtain models in which the optimal investment decisions exhibit interesting cascade effects.

Assume that the first investor, learns of the opportunity and its expected returns. If he decides to invest, then each other investor gets an opportunity to invest sequentially and only observes the decision of the previous investor (he does not have his own private “signal” of returns). Now, each investor knows the prior probability of another investor’s having high costs and the prior of an investment yielding high returns. He knows the action that each investor will take conditioned on his cost and knowlege of returns. So, if he observes the previous investor’s investing, he makes use of a bayesian posterior calculation of returns and uses this to make an investment decision based on his costs. Certain settings of the priors yield different gaurantees on the optimal investment decisions for each investor. For this model, we assume that they are set such that the second investor invests if the first investor invested (it can be shown that such parameter values exist). Now, the third investor knows how this decision process works. Therefore, his observing that the investor2 invested is equivalent to knowing that investor1 invested and therefore knowing the state of the world. It follows that everyone invests if and only if the first person invests. This model is the extreme form of our information cascade.

Next, we consider the effects of relaxing the assumption that not all the agents know of the opportunity, while still assuming that only the first investor knows of the returns. Now, since everyone knows of the opportunity, all low cost investors will always invest (low returns > low cost). However, all high cost investors invest only when high returns are sufficiently likely (determined by parameter values). Under the assumption that a high-cost investor would not invest solely on his private information, the authors show that there exists a high-cost investor that will not invest even if the previous investor invests. The reason is that if it the investor knew that the previous investor had low costs it would not give any information about the returns. It was shown that the infered probability of high returns decreases with time due to the increasing expected number of low-cost investors and therefore decreasing informativeness of the previous investment. Under this model, the market is robust to a runaway information cascade in the sense that a few early investment decisions are not enough to sway the entire population.

The previous two models fail to capture the fact that speed of transmission should depend on the importance of the information - it should spread slowly if it affects few people. Here the authors relax the assumption that the rumor is heard sequentially and consider a model where the probability of an investor learning of an oportunity in a given time interval is proportional to the number of people who have invested in the past. The analysis of this model is much more complex than the previous two, so only the main results are discussed here. Under this assumption, the rumor propagates faster if the investment yields high returns. Furthermore, the ratio of the probability that a rumor is heard at time s given low returns state to that in high returns state increases with time and is unbounded. So, as time goes on, if an investor had just heard the rumor for the first time, it is more likely to be a low returns state. In this sense the rumor gives us very precise information. This is result is in sharp contrast to both the results of the first model, in which the informativeness of the rumor remained constant throughout time, and the second model, in which the informativeness of the rumor decreased with time.

The three models considered here provide different interperetations of a rumor for each individual as time progresses. The first model exhibits dynamics similar to the model we discussed in class in which observing two people’s descisions could affect everyone else. While for the second model, the optimal strategy of certain individuals (high-cost investors) was to rely on the previous investment less and less as time went on. Finally, in our third model, the optimal investment choice as time went on was to rely entirely on having heard the rumor at that time (and realizing that it was a low returns state), and was not infulenced by private “signals” (high/low cost investor).

Yesterday, AOL announced that it will begin selling some of it’s own search advertisements in order to increase its revenue. Back in 2005 Google and AOL (Time Warner) struck a deal in which Google bought a 5 percent stake in AOL, for $1 billion, and in return Google was allowed to sell all ads that appeared next to searches made on AOL sites. This service, called AOL Search Marketplace, is currently not allowing AOL to tap into it’s full advertising market share and in response to AOL’s decision last year to offer most of it’s services for free, AOL has been looking for a way to increase advertising revenue. Beginning Monday, April 9th, AOL will begin selling a portion of the ads displayed next to searches made by AOL users, fed through Google’s advertising system. This will allow marketers who already have ads that appear on other AOL pages, to better target AOL users.

Google is willing to be more accomodating to AOL, one of it’s largest partners, to continue edging out Microsoft who fought in 2005 to be AOL’s advertising partnet. Google, who currently shares AOL search ad revenue with AOL has devised a new formula making sure that AOL now retains a larger percentage of the sale. Not only will AOL benefit from this new deal, but Google may even see an increase in it’s own revenue. If the new advertising system proves to be more appealing for many marketers, companies will bid even higher to have their ads appear on AOL search pages.

By selling it’s own ads, through Google’s advertisement auction that we discussed in class, AOL will be able to increase its own revenue by increasing both the click through rates  (CTR) of the advertisement slots on its pages as well as the values each marketer places on the ad slot. If the ads that appear next to searches on AOL pages are more relevant to AOL users, the ads will receive more hits, increasing their slot value. Other online subscription services should look to AOL as an example for how to offer public services for less, by honing in on the relevance of advertising and increasing ad slot value.

 http://www.nytimes.com/2007/04/07/technology/07aol.html

Article: “Value creation in fine arts: A system dynamics model of inverse demand and information cascades” by Philip Crossland and Faye L. Smith in Strategic Management Journal, May 2002

This article looks at information cascades in the fine arts market, specifically focusing on a company that produces porcelain sculptures of flowers and birds: Edward Marshall Boehm, Inc. Although fine art products are tangible goods, they differ from other tangible products on the market in that they also have qualities common to intangible goods, such as the expressive value of performing arts. Because these products are subject to changing social trends and creative talent, the demand of fine art products is difficult to predict.

Before purchasing artwork, it is necessary to determine its value in artistic and monetary terms. The first group to recognize the value of a work is “cognoscenti”, a group including people like museum curators. Following the cognoscenti group are two groups of customers: investors and collectors. Investors purchase artwork in order to see its monetary value appreciate, while collectors privately purchase artwork for their own pleasure. Because investors and collectors generally know less about the quality of artwork than cognoscenti, they will follow the actions of the cognoscenti group; however, these two groups will be “motivated to purchase by different information cues” from the cognoscenti.

In the “information cascade… from the cognoscenti to the investors and collectors”, both investors and collectors will follow the purchasing actions of the cognoscenti. Given their different purchasing motives, however, a “greater lag time” can be expected between the decisions of the cognoscenti and collectors than between those of the cognoscenti and investors. This article explains how there are similarities and differences between information cascade in “cultural industries” and manufacturing industries. While a cultural industry is similar to a manufacturing industry in that companies can also have a competitive advantage in the market, a company in the cultural industry achieves this by having an “inimitable” balance between artistic talent and technological process.

In class with the example of video players, we learned that it is not the technological superiority of a product that matters, but rather whether that product has been accepted by the majority of the public. This article suggests that a similar effect seems to be occuring in the fine arts market as well.

http://www.nytimes.com/2006/09/23/technology/23click.html?pagewanted=1&ei=5070&en=d53f58f7e87516f9&ex=1176177600

note: I believe that you may need a NYtimes online account to view this article, since it dates back to Sept. 2006.

Although advertising is a key component in any business’ strategy to attract new customers, there are some forms of advertisements that may be financially detrimental to a company. Obviously, it takes money to advertise, but usually the influx of customers and their money will create a balance, or preferably a surplus. But as the article above shows, pay-per-click advertising may cost companies money. The advertisement links are being clicked, and the companies are being charged, but sales are not generated. The reason for this catastrophe: click fraud.

Click fraud is essentially the act of clicking a paid link without being a prospective customer. These clicks can be generated by an automated computer program, or by people who actually spend time to click the same links over and over. Click Forensics, a consulting firm, estimates that 14% of total clicks are fraudulent. Though at first it may not seem like a significant loss, one must be aware of the figures of online paid search advertisements – about $7 billion in the fiscal year 2006. This means that some companies can lose up to tens of thousands of dollars in a matter of months due to click fraud.

However, pay-per-click advertising has not always been so troublesome. Mr. Burton, a business owner said that he used to make money with pay-per-click advertising, when it wasn’t such an expensive method. This is understandable, because even if there were some click fraud in the past, the sheer cost of pay-per-click advertising was lower and thus the impact of a little fraud was negligible. But nowadays click fraud is becoming rampant. All competitors on the same list of paid links want to be at the top, and so they resort to spam clicking links to push companies past their budget. This motive is a sad one indeed, but in such a cutthroat world of advertising of today, it is not a surprise.

What can be done about this? Unfortunately, not much. The internet is so vast that it is not easy to track every possible click, and it is not easy to come up with a solution to doing so. Big companies with a lot of money can afford to pay for some measure of fraud such as programs or even a professional pay-per-click management company. But for many smaller companies there is only so much that they can do.