This is the summary of a behavioral targeting article by Saranga Komanduri, et al. You can get the pdf of the behavioral targeting article here: AdChoices? Compliance with Online Behavioral Advertising Notice and Choice Requirements

Online behavioral advertising is the practice of tracking consumers activities online to target advertising. Part of the responsibilities of companies engaging in online behavioral advertising is to listen to the privacy concerns of their consumers. Thus, in 1999, the Network Advertising Initiative (NAI) was created. Now, 66 companies are members of the NAI and they offer services for consumers to opt-out of online behavioral advertising. Another set of organizations formed the Digital Advertising Alliance (DAA) and they too have set principles of self regulation for online behavioral advertising.

The FTC determines what the next step is, and this study helps by investigating how effective opt-out and notices are. This paper also studies if the regulations implemented by DAA and NAI are being followed by several websites.

Online Behavioral Advertising

With online behavioral advertising, users are tracked to give them advertisements that are relevenant to their interests. Cookies are used to do the tracking. Despite its effectiveness, privacy concerns are being raised because data is being collected in secret, and this can create profiles of users tagged with sensitive personal information.

A huge majority of Americans who participated in various services found behavioral advertising tobe invasive. They did not want ads that reflect their interests, and wanted websites to erase their personal information immediately. Thus, the opt-out mechanism was invented. Furthermore, self-regulatory methods were invented in the form of notices and website privacy policies, these policies are so hard to read that users can’t read them every time they visit a website. However, a study shows 50 percent of users believe that if a website has a privacy policy, that website won’t share their personal data.

DAA and NAI Principles

DAA has seven principles. Education Principle (to main the website of DAA as an educational website), Transparency Principle (inspect ads and websites as they share certain info), Consumer Control Principle (provide opt out mechanism), Security Data Principle (data security requirements), Material Changes Principle (require companies consent before making changes), Sensitive Data Principle (steps for handling sensitive data), and Accountability Principle (develop compliance programs).

These are the same principles to NAI, along with additional principles that are irrelevant to the study.

Methodology

66 NAI members were checked during February to March 2011 for the following requirements. Privacy notice requirements at the front page of their websites. Their privacy policy. Opt-out cookies from the opt-out mechanisms by NAI and DAA. Compliance with the notice requirements enhanced from the principles of DAA, through inspection of the ads posted in the websites. These enhancements are only for behavioral advertisements, so ads that excluded an ad network were not inspected. The estimate is that 80 percent of the ads are behavioral, and rely on third party cookies.

Results

Non-contextual ads were looked for in 100 websites, and 400 pages of these websites. 164 of these pages are monitored by the NAI members. The enhanced notice requirement requires the notice to be in the same page that contains the behavioral ads, but upon investigating these 164 pages, only 35 percent put enhanced notices, which is a huge compliance gap. The study further sees most of these enhanced notices are from the advertisers, the companies purchasing the ads, and not the members of NAI.

For Privacy Notice requirement, only Audience Science stated in its privacy policy that it abides by the principles of DAA. It is therefore the only NAA member which fully complies with the requirements for privacy notice. Aside from this fact, 83 percent of the members would be compliant. For choice requirement, the opt out mechanisms of DAA and NAA were checked in this study. The two mechanisms have shown inconsistency between them, such as setting different cookies and getting different content from the same advertiser. Other results show that these mechanisms do not work in Safari. These are some of the results obtained from this study.

Implications for Public Policy

Compliance to DAA Principles is slow and infrequent in some aspects, opt-out mechanisms contain errors. Furthermore, the definition of online behavioral advertising by the DAA and NAI may not be enough to ease concerns regarding privacy. Furthermore, two members of NAI, Undertone and Valueclick, imposes their demands and limitations to a user that visits their site, through messages that direct them to the privacy policy, and this may surprise the user as to the limitations of his or her rights upon reading. Finally, some members of NAI have their own opt out features, usually going beyond what is set by NAI, which is a good thing for privacy.

It is important to know how people interact with a SERP, or a search engine results page. One way to know this is to make an eye tracking study and understand the variables which affect user’s behavior with these circumstances. This is the summary of a behavioral targeting article by George Buscher, Susan Dumais, and Edward Cutrell. You can get the pdf of the behavioral targeting article here: The Good, the Bad and the Random: an Eye-Tracking Study of Ad Quality in Web Search.

Web Search Behavior

Factors affecting web search behavior include search results quality, presentation, search task type, and individual differences among users. Users that search the Web do so for three major goals, resource or transactional, navigational or informational. Users that search for navigational purposes are faster and more successful. A lot of researchers study the variations of search behavior among different users. The conclusions are obvious; experts in searching have more techniques to effectively find a solution for their search purposes.

SERP Eye Tracking

Results have shown that the way users examine a SERP is a function of the relevance of the search results and their position. In fact, users highly favor those results that are at the most effective positions in a SERP. There are two kinds of searchers, the economic and the exhaustive. Exhaustive searchers carefully look at all entries to find which one fits their query best, and economic searchers usually just click the most noticeable relevant entry. Aside from the organic search results, SERPs also have other components, such as rich snippets, related queries, spell suggestions, and sponsored links. Not too many researches focus on these other parts, except for sponsored links. They make up 10 to 23 percent of the links located in a SERP. In one study, 51 percent of users only notice the organic search results, and don’t look at the sponsored ads.

What influences a user’s behavior when he searches? These are, search result relevance, the search task, and individual characteristics of users. This study is a comprehensive and detailed look at other components of a SERP aside from the organic search results, investigation of ad quality, and determining user attention and user behavior with varying ad positions.

Methods

Eye-tracking is used for this study. The data obtained from this approach will help in the investigation of how search processes and strategies go about. Participants of this study were given 32 search tasks, with 16 of these tasks having a navigational search purpose, and all are commercial tasks to make the presence of ads in the SERPs relevant. An example task description is, Find the special offers page for Southwest Airlines, and they were given an initial task query, which is “southwest special offers.” This gives the researchers the ability to compare the initial task queries of the participants. The positions of the solutions for each task also vary; some within the top 3 results in the organic search, and others within 4-6 results and greater than 6. After the initial query, the participants were then given the liberty to do anything and click anywhere in the SERP.

Each SERP in this study had 3 ads on the top, 5 ads on the right. Ad quality varies between bad and good. Good ads come from Yahoo, Bing or Google, and bad ads came from those commercial search engines as well, but they were generated from terms that are a subset of the initial task query were used. For example, a good quality ad for ibuprofen side effects is that same phrase, while a bad one would be free sound effects. This study uses their own search interface which is modeled from some commercial search engine.

The tasks are grouped together into 4 blocks, and there are three kinds of blocks, the Good, the Bad, and Random. Good blocks mostly contain good ads, bad blocks mostly follow bad ads, and random blocks contain half good and half bad ads which are randomly assigned to each of the eight consecutive trials in that block. Participants are grouped according to conditions. There are three conditions, GB, BG and RR.

Procedure, Apparatus, Participants and Measures

Participants were introduced to the study, then the eye tracker was calibrated. Participants then underwent a practice task before proceeding with the next 32 tasks. After doing the tasks, they answered a questionnaire asking them to describe their experience during the Web search, among others. It took one hour per participant to perform the entire procedure. The apparatus used was an LCD monitor, Internet Explorer 7 browser, and Tobii x50 eyetracker using the software Tobii Studio. 38 participants performed the experiment, 17 are male and 21 are female, and their age ranges from 26 to 60 years.

Results

For an investigation on the general gaze distribution on a SERP, the results obtained in this study are in line with the results of previous researches. With regards to the top results, there are more clicks than attention in that area. The top and right ads receive more attention than clicks. The top ads get the same amount of attention as the organic results, but much fewer clicks.

Users spend more time when presented with informational tasks than with navigational tasks. When doing informational tasks, users spend most of their time in the upper search box and organic search results. The extra time that they had were strikingly not spent on the top ads. There was no even distribution of attention, but most of that was given to the top two organic search results. Furthermore, users spend twice as much time on the search box for informational tasks than navigational tasks, implifying they research more for these tasks.

For good quality ads, users spend twice as much time on the top ads. Less attention was given to the organic search results for good quality ads. Indeed, these ads had a direct effect on the performance and attention of the users towards the SERP. However, right rail ads seem to be ignored for the most part. Furthermore, the order in which users see pages with good ads or bad ads have a strong effect on the behavior of users. For displaying good and bad ads in random order, the ads tend to be ignored even more, despite their quality. For consistently showing good ads, the ads get more attention and clicks.

Many young adults love hanging out in Facebook, Myspace, or other social networking sites. These users share personally identifiable information to a personal profile, which is like their own web page, usually posting information related to their demographics, contacts, and name. This article is the summary of a study written by Mariea Grubbs Hoy and George Milne  about the gender differences regarding online privacy among 18 to 24 year olds, and the use of personal information beyond social connection purposes. You can get the pdf of the behavioral targeting article here: Gender Differences in Privacy-Related Measures For Young Adult Facebook Users.

Advertisers have used the content found in social networks to target users with relevant ads. For example, advertisers can send out bridal shop ads to women whose profile relationship status is set to Engaged. Consumer searches, visited Web pages, and viewed content are also used for targeting, in a strategy which is collectively known as behavioral advertising or behavioral marketing.

Users who join Social Networking Sites don’t expect to see ads relevant to their interests. They only joined because they want to connect with their friends or acquaintances, and so to most of these users behavioral advertising is just an invasion of privacy. What advertisers can do is really understand how social networking site users behave and what they believe in, specially for the women, who use these sites more than men, and they are found to be more concerned about privacy.

Social Networking and Privacy

Social Networking Sites used to be about groups and forming online communities that have common interests. But it has evolved to focus more on the way people communicate with others, and a way to connect with “friends” and know their social spheres. As it turns out, online interactions in these sites encourage sharing of personal information to others. The more personal information you share about yourself, the more findable and knowable you are. Many young adults know that they leave personal information as they interact with these sites, but most of them don’t really care about having a limit to the amount of information they are willing to share about themselves. In fact, studies show that most college students share high levels of personal information in their social networking site profiles.

Facebook

Facebook is the most popular social networking site. Most of its users share a huge amount of personal information, due to several reasons, including peer pressure and ignorance to the implications of disclosing personal information. This implies that social networking sites are data rich environments, and motivates Facebook to use behavioral targeting effectively for their advertising endeavors. If you have an account, you must be familiar with applications asking you to allow access to personal information before proceeding. Controversially, some of these apps gather information which they don’t really need.

Research Questions

The study will primarily examine gender differences related to privacy concerns, beliefs and awareness about the use of Social Networking Site profile information for behavioral advertising. These are the three main questions. First, what is the difference between men and women in terms of privacy beliefs and concerns with respect to social networking sites? Second, what is their difference in terms of beliefs and awareness about the use of profile information in their sites beyond their original purpose? Finally, what is their difference in terms of privacy-related behaviors?

18 to 24 year old adults from the United States were recruited through Facebook and surveyed. In summary, 29 students joined a Facebook group as part of their research class, and they invited all of their friends to join as well. The final profile for analysis includes 589 respondents, 72.7 percent female and 27.3 percent male.

Privacy Concerns and Beliefs

A five-point scale was used to assess the beliefs of the respondents, 1 being very unconcerned, to 5 being very concerned. Results showed that women were more concerned than men about the privacy of their personal information. In terms of the ability of social networking sites to protect their privacy is concerned, both genders neither agree nor disagree, and both agree that sites should alert them when companies use behavioral targeting.

Beliefs and Awareness of Profile Usage Beyond Original Purpose

Both genders are a bit aware of advertisers’ practice on social networking sites, although female are more likely to go against this practice. When men have greater privacy concern, they tend to agree more on the fact that advertisers use their site information for targeting ads, that they should know how their personal information is used, and that they should be alerted if someone’s tracking them. Women have the same results, but agreed less on social networking sites doing well in protecting their privacy for increasing privacy concerns.

Privacy Related Behaviors

More women read the privacy policy of a social networking site before joining, and a lot more men haven’t read the privacy policy at all. Women have done all of these more than men: untagged pictures, being careful about accepting friends, monitor profile, careful about posting pictures, controlling privacy settings and using the privacy feature. More women ask others to delete some personal information posted about them that they didn’t want to appear. For privacy protection, men’s strategies include lying, technology, managing image, being careful. Women’s strategies are lying, review, being careful, and post hoc control.

Implications for Behavioral Advertising

For the sample of young adults surveyed in the study, there is a general lack of concern. That is mostly because of the benefits from following a social networking format, e.g. it’s fun to browse other people’s profiles form information. However, they should know and be aware of how their personal info is used by advertisers. So this study suggests that when advertisers are using the information of a user, it should alert the user. There is already an initiative for this, acted on by the group Future of Privacy Forum, by coming up with a standardized logo for each time users are alerted about the use of their personal information. This logo will be paired with the sentence, “Why did I get this ad?” which links to a page explaining behavioral targeting. Furthermore, if awareness as to how the data in social networking ads are used for advertising is enhanced, that will benefit both the advertiser and the social networking site.

To be successful with advertising in the past, you needed to mass advertise. Even with that strategy, you know only a few people you actually reached with your ads are interested in your products and services. Now, with online advertising, both the advertiser and customer will benefit from targeted ads where advertisers can expect to get a higher response rate and consumers will be kept from seeing irrelevant advertisements. This is the summary of an article by Aleecia McDonald and Lorrie Faith Cranor. You can get the pdf of the behavioral targeting article here: Beliefs and Behaviors: Internet Users’ Understanding of Behavioral Advertising.

One form of targeted advertising is behavioral advertising. It involves collecting data from online consumers regarding their browsing behaviors, and creating user profiles out of these data to give out the most relevant ads to each individual consumer. There are many ways to implement behavioral advertising. One of them is through the use of cookies. Advertising networks set and read cookies to create user profiles based on a user’s possible location determined by IP address, sex, age range, ads clicked, sites visited, etc.

Google and Yahoo have far-reaching prowess in terms of mass media. Each month, Yahoo gets around 500 million unique visitors, while Google gets 1.5 billion. Google beacons reach 92 of the top 100 website in the world, and 88 percent of some 400,000 websites. Collecting vast amounts of data is a huge task for these web giants, and the consequences for privacy harm is huge as well. In fact, legislatures in the United States have called for new Internet privacy regulations. Furthermore, Industries aim to be self-regulating in terms of privacy, know what consumer’s reactions are to privacy loss and online practices.

Research Methods

A laboratory study and an online survey was conducted in this study. The first part of this study saw 14 subjects asked to participate in a research about internet advertising for a university, which was done in Pittsburgh, and started from September 28 to October 1 2009. The second study is participated by 314 subjects from a website called Mechanical Turk. They were asked to participate in a study which lasted for 20 to 30 minutes, and was held at the end of the April 2010. 64 questions were asked which were categorized into questions regarding online purchase, paying for privacy, cookies and the computing environment, how cookies work, screen shot ads, opt-out pages, behavioral advertising through email, and questions on demographics.

Perception About Cookies

This study focuses on first party and third party cookies, and disregarding other less popular cookies (flash cookies, browser fingerprinting, session cookies). First and third-party cookies are well known, and users are more likely to employ cookie management if they are concerned with their privacy preferences over behavioral advertising. During the survey, however, about 33 percent of participants said they are unsure what cookies are. More than a third are partially correct about the nature of cookies.

In response to the questions regarding how cookies can help them, the respondents had varied answers. Some relate cookies to saving passwords, storing website preferences, and three responded that cookies involve personalized advertisement. 91 percent of the participants know that cookies are stored in their personal computers. Most participants also know that cookies can be used for saving passwords and website and advertisement personalization. 75 percent of respondents know that cookies record the websites they visited, and 50 percent believe cookies can be used, along with other information, to know their names.

There are false beliefs as well among the respondents. 50 percent falsely believe that without cookies, information about their locations will be kept private, and 50 percent falsely believe cookies record their home address, name and where their computers are bought. Furthermore 44 percent falsely believe that cookies are the reason why the forward and backward browser arrows work. 30 percent, and 59 percent are uncertain, falsely believe that the law states cookies are not allowed to contain credit card information.

Managing Cookies

In questions regarding how often the respondents clear cookies, 9 percent said they never do it, 9 percent said they do it once a year or less, 16 percent said a few times annually, 10 percent says monthly, 17 percent says a couple of times per month, 16 percent says weekly, 12 percent daily and 8 percent said every time they end a browsing session. However, the respondents are not so sure why they clear cookies. They are not sure when they should retain or delete cookies, how they work, what their benefits are. The confusion among consumer’s regarding cookies is indeed high. Some respondents don’t know how to delete cookies even if they want to, some delete cookies for reasons based on false issues. In turn, the advertising community is put at a disadvantage.

Cookies and Browser History

More than half of respondents do not understand that browser history is independently stored from cookies. Some respondents also falsely believe that not logging in to a website will keep them from being tracked. It is possible that the user interfaces of browsers contribute to the confusion. Clearing cookies and clearing browsing history all belong to one page in a browser. So as they clear both, they create a false connection between the two. In fact, 35 percent said they are sure cookies and browsing history are the same.

Tailored Content and Privacy Concerns

A substantial group of people have a good understanding about contextual search advertisements. All participants know that Google has ads, that these ads are located in the right side of the search page, and that there are also sponsored links. The respondents also know that advertisers pay Google to display ads, but many are not sure about the payment method.

A lot of participants know about behavioral advertising, but only a minority believe that email advertising exists today. That implies that very few are informed regarding the business model of Gmail and how it works. In general, the participants prioritize online privacy, and that tracking to them is invasive. Only 10 percent said they do not care if advertisers collect personal data, but only 15 percent said they will stop browsing websites that have behavioral advertising. Furthermore, people are less than willing to pay for privacy preservation if they value privacy so much, a result that seems to contradict intuition.

Conclusion

From the results of the study, it can be concluded that many people view behavioral advertising as a threat to their privacy. However, the study also shows that these people have very limited knowledge regarding how behavioral advertising works; that their knowledge is insufficient so as to make decisions on taking care of their online privacy. Furthermore, there is a small group of users who are quite positive about targeted advertising.

In this digital age, search engines play the most important role. Every day, search engines process over 4 billion queries from users looking for a plethora of information. It is just natural then for advertisers to take advantage of the popularity of search engines such as Google and Yahoo, and now, search advertising is a multi billion dollar industry. This is the summary of a research by Alexandre de Corniere. You can get the pdf of the behavioral targeting article here: Targeting with Consumer Search: an Economic Analysis of Keyword Advertising.

One of the reasons why search advertising is so popular is because it is the cheapest way to advertise right now. It only costs 8.5 dollars, compared to using mail advertising which costs 70 dollars, or the yellow pages for 20 dollars, among other forms of advertising. Pay per click basis, and the fact that it is intent-related in nature makes search advertising cheap.

Intent-related advertising

Intent-related advertising is different from content-related advertising. Intent-related means looking for ways to know what consumers really want to purchase. Search advertising is a kind of intent-related advertising because it innately targets people who are interested in the products they are looking for through a keyword search. For example, if you want to advertise a hotel near the Eiffel tower, using local newspapers is not effective because the locality is less likely to check in your hotel. Furthermore, TV advertising is quite expensive. Search advertising is the best option, as it targets those consumers who are looking for hotels near the Eiffel Tower exclusively, and your search advertisement is most likely to get a positive response.

Per Click Pricing

Per click pricing is a great way for advertisers to conserve resources on ads that were not used by the most relevant consumers. That is, you only pay for the ads that have actually been seen by consumers, and which they have been looking for in the first place.

The research paper introduces a kind of targeted advertising which has the features of per click pricing and intent-related advertising. The model in this study involves firms that are differentiated horizontally, and consumers that have no idea about the products and prices of the firms. Firms choose keywords they want to target. Consumers type keywords and look at the links that appear and respond to them. The search engine acts as the third party between the involved parties.

Efficient Advertising

Sections of the study investigate a stiuation wherein the search engine cannot serve as an intermediary for helping firms target consumers. However, the researchers point out that when a firm can indeed targeted consumers, that leads to consumers being more happy, because they are getting relevant ads and have better options. This study also predicts that because of the ad relevance, users will tend to visit only one firm, reducing their search cost.

Model

In this study, the consumers basically don’t know anything about the firms, so they have to search before purchasing. Consumers can only interact with firms through a search engine, an intermediary wherein firms show their target keywords, and consumers show the keywords that interest them. A consumer enters a keyword, and all the firms that can relate show their ads to the consumer. The consumer then knows more about the firm and the user obtains a search cost. If the consumer clicks on an ad, the firm pays the search engine a certain amount.

The main strategy of a firm is to choose the right prices and the right set of keywords, one that results to maximized profit. The strategy of consumers is to use a stopping rule, which is the best response to the strategies created by the firms.

Optimal Advertising and Pricing Strategy

This study has a surprising prediction; that a consumer buys a product from the firm it first visits. This is only a consequence of the model used in the study, and it just implies the targeting behavior of firms as they reach out to the consumers who have a high probability of purchasing their products. Some assumptions in the study include the following: Before buying, a consumer will search more than once if the firm doesn’t target specific key words. There are three other assumptions which are for the technical aspect of the model.

The proposition of this study is that the overall welfare is increased if firms target their customers, as there is a reduction of search costs and poor matching between consumer and product. In terms of the equilibrium price, this targeting technology has both an internalization and outside option effect. Internalization effect means that firms are only interested in targeting customers who will buy from them. The outside option effect means that if a consumer sees a targeted ad, he understands that continuing to search will give him a better option, and he gets closer to what he truly desires to purchase.

Strategic Search Engine

Search engines are concerned about how ads are displayed. In fact, Google, for example, uses a quality score which factors in on ranking keyword bidders. The same is true with broad match technology, which matches consumers with keywords that are not really what the firms targeted but are close enough. A firm’s advertisement for “web hosting” will appear on a consumer’s screen if the consumer searched for “webhost”. This allows firms to save resources and time finding exactly what the right keywords they should use.

If firms have the ability to target consumer’s, then they will benefit from a lot of potential gains in terms of efficiency. With a small enough cost for advertising, targeting can lead to smaller prices.

With the popularity of mobile smart phones, and the emergence of mobile advertising, this paper is a summary of a behavioral targeting article by Byoungjim Kip, et al., which is all about AdNext. You can get the pdf of the behavioral targeting article here: AdNext.

Mobile Advertising

The popularity of mobile advertising keeps on growing, and by 2012, 19.1 billion dollars will have been spent for this application worldwide. In the new mobile era, research for more effective mobile advertising is crucial for both mobile users and advertisers.

The COEX Mall, the largest commercial complex in South Korea, has over 260 shops and 100,000 visitors per day. This mall, and other huge malls around the wall are perfect places for mobile advertising. Advertisers take advantage of the fact that most people visit the mall to purchase goods and/or services, and shoppers can use mobile advertising to familiarize themselves with the services and goods offered by the stores.

Customer Targeting

Customer targeting is important in mobile advertising, because you want to identify the mobile users who are most likely to purchase the products or services you advertise. The ads themselves become more effective. Furthermore, spacial and temporal relevance are important as well. If a mobile user sees a relevant ad, and is close to the store that sells the product, or if the ad comes at a time when the user will most probably purchase it soon, then the user will very likely purchase the advertised product or service. Present mobile advertising is limited in that they only targeted users based on location, which is insufficient information if the purpose is targeting the user with an ad that he or she really desires to respond to.

AdNext incorporates spatial and temporal relevance in the ads by predicting a user’s next visit place. AdNext does this by studying the sequential visit patterns of users collectively. For example, users who have been to the mall cinema might generally go to the cafe next. Of course, human behavior is unpredictable, so probabilistic reasoning plays a critical role in designing AdNext. In particular, this study uses a prediction model built from Bayesian networks.

AdNext Design

In mobile phone users, AdNext clients collect the user place visit history, which is built through Wi-fi fingerprints as the user visit one place in the commercial complex and another. The client then sends this information to the AdNext server, which will then send the relevant ad to the user after building a prediction model from the visit history. This is called online prediction mode. The server is also responsible for collecting feedback data, such as issue counts, click counts, among others. This will help train the server for offline learning mode, along with the preexisting collective data from a large number of users.

Collection of place visit histories is challenging because store-level accuracy for localization is not common, and determining the visit times for creating an accurate visit pattern is difficult. Wi-fi localization techniques are used since there are many Wi-fi access points in commercial complexes such as COEX Mall, which saves the researchers from installing additional facilities. Using an Elekspot ID, detection of place visit is done through current location detection, then location change validation.

Next Visit Place Prediction

It’s hard to predict the next place a user visits in the mall because people tend to have uncertain behavior, and users may not want to give away their place visit history. But the researchers argue that collectively, there are cause and effect relationships between several places in the mall, and sequential visit patterns are formed. An example pattern is cinema to restaurant to cafe. Probabilistic models for this study is made from Bayesian networks. The primary features of this model include age, gender, visit time, visit duration, and visit place. Furthermore, selecting the relevant ad is also important. Ads based on place category can lead to too many ads to fit in one small mobile phone screen, so an evaluation method for scoring relevant ads is proposed. An ad scores higher if the store offering the product is closest to where the user is right now, and if the user rates that place highly.

The setting of this research involves a group of users who, at any time, can request a web server to load a web page. The web server, upon loading the requested web page, will also load with it a banner ad (the process known as impression), which the user can respond to or not. This is the summary of an article by Fabrizio Caruso, Giovanni Giuffrida, and Calogero Zarba. You can get the pdf of the behavioral targeting article here: Behavioral On-Line Advertising.

Several things will occur: impressions, clicks, registrations, keywords in a search query, page views, among others. Registration refers to a users response after clicking the banner, such as purchasing the product advertised. Registrations, clicks and impressions for a given banner can lead to profit. For simplicity, profits generated from clicks are to be initially investigated. The goal is to maximize the number of clicks and the profit obtained from all the clicks. This study uses the Bayesian model, a much simpler model compared to other studies, for investigating how to maximize the number of clicks for a banner.

This study uses a cookie embedded in a user’s browser, and all the information gathered from that cookie, and use an algorithm to use that information and come up with the most suitable banner ad for that particular user.

Banner Selection Algorithm

This study uses a banner selection algorithm aimed at maximizing the probability that a user will click the banner ad presented in the web page he or she requested. To do so, the web server selects the most appropriate banner from a set of candidate banners. The web server chooses the banner with the highest score, where score is defined as a function of the profit, or cost per click of the banner, and the “rule”. The rule is a mathematical function involving the summation of probabilities, to say the least. But when the value of rule is highest, it means that banner has the highest probability of being clicked. The probabilities used in the algorithm are determined through the concept of click-through rates, which is the number of clicks on a banner divided by the number of impressions of the banner.

Avoiding User’s Boredom

This strategy is implemented so that the user will not keep on seeing the same ads over and over again. This is by the addition of a throttle function to the score, which decreases for any given banner as the number of impressions of that banner increases. Aside from avoiding user boredom, this also increases the probability estimation of a click.

Generalization of Score

While previously, only clicks are investigated for profits, the generalization includes profits gained from impressions and clicks, so that the total score now includes the addition of a variable called imp_profit. And by treating other factors, such as registrations, as click-like events, the score can be generalized further.

Ad Networks help advertisers post ads on publishers’ websites. The two latter entities rarely contact each other. Both advertising networks and advertisers want to improve their ad campaigns, and so they look for means to enhance ad features such as the formats, channels, and marketing forms. These are only short term solutions, so AdMotional offers long-term solutions to make ads highly effective. This is the summary of a behavioral targeting article by Manfred Meyer, et al. You can get the pdf of the behavioral targeting article here: AdMotional: Toward Personalized Online Ads.

In summary, AdMotional improves ads for long-term effectiveness through optimization of targeting, and personalization. Targeting means choosing ads suited for a particular group of customers, while personalization refers to customizing ads for a specific customer. The system architecture used in this study incorporates these two factors.

System Architecture

The design project can be seen as embedded from within an ad server infrastructure, which allows it to benefit from being able to deliver ads quickly. The sole purpose of this project is to aid the generation and selection process of campaign ads as the ad server requests it.

The website that a consumer loads contains URLs to scripts located in the ad server. The browser of the consumer will then request these URLs, and so the ad server responds by referring these requests to the AdMotional system. The system will also get valuable information about the consumer from the ad server, such as URL, type of browser, language, and even browsing history and parameters of top priority. These relevant information will be used by AdMotional to select the most relevant campaign ad.

After doing so, AdMotional will then customize these ads in terms of font size, background color, among others. For now, customization is capable of generating jpg and png images and html. This customized rich media ad will then be sent back to the ad server, which will then be sent to the consumer browser for display.

The campaign selection and personalization are made more accurate by gathering internal and external data. This is done to include an emotional dimension to the consumer model; or what the consumer’s situation is at present. Non-emotional factors are used to cut down the number of possible campaigns, but the final choice for most relevant campaign is largely determined through these emotional factors.

The architecture of the system is composed of five major components: communication, campaign selection, dynamic ad creation, learning and optimization, and data storage and external services.

Campaign Selection and Emotional Targeting

Campaign selection is responsible for collecting targeting information and determining which campaign ad is best for the consumer that requests it. To enhance the targeting procedure, three techniques are combined together: situational targeting (referring to location), behavioral targeting (browsing behavior) and contextual targeting (web page info). But this study also proposes emotional targeting, a kind of targeting which models the emotions and the possible feelings of the consumer, which can be obtained from a complicated emotional classification system, which is a function of external services, consumer requests, among others.

Dynamic Ad Creation

AdMotional requires advertisers to provide ad templates that can be customized in several points. This provides additional work for advertisers, but a lot of personalized designs can be made from these templates, so the hard work of manually designing ads is reduced. Furthermore, from the advertiser’s point of view, one would prefer as many control parameters as possible, but this is constrained by the limited time delivery of ads by ad servers. This is resolved through two different kinds of creation methods and the process of creating personalized ads is divided into three modules, e.g. one for producing pixel images, etc.

Learning Component

Depending on the intentions of advertisers, the learning component is divided into two types: performance campaigns and branding campaigns. The objective of performance campaigns is to improve the immediate response of the consumers, while the objective of branding campaigns is to strengthen the brand of the advertiser on a long term basis. For performance campaigns, the most effective personalized ads are clustered after their effectiveness have been measured. The less successful personalized ads are important information as well, and adjustments to what the next personalized ad design should be are suggested. Recommendations are not on individual criteria, but rather in bunches or groups of criteria because the variables involved are rarely independent variables, e.g background vs foreground color.

This is the summary of a news article in 2009 by Miguel Helft. The following is a link to the original behavioral targeting article: Google to Offer Ads Based on Interests

Interest-based Advertising

Google starts implementing interest-based advertising, which is a form of behavioral targeting that is widely accepted by it’s competitors, including Yahoo. However, it has drawn criticism from the Congress and advocates of privacy. To answer such privacy issues, Google introduced an opt out feature, and an additional step for users to be able to see their profile and edit what Google compiled as their interests. These two steps are applauded by privacy advocates, and especially so towards the former.

Most likely, other companies which are second to Google may follow what the huge company has done. But many privacy advocates also said that Google needs to make people know more about how they are being tracked.

Categorizing Consumers

Implementing behavioral targeting into Google Ads is the first byproduct of the company’s acquisition of DoubleClick. In this strategy Google uses a small piece of text, known as a cookie, found in a user’s browser to gather information about the browsing behavior of that particular user. This technique will then categorize the user based on what he or she is most interested in (e.g. cars, sports). That user will then see ads displayed in the websites he or she visits that are relevant to his or her interests. And it doesn’t matter if the website content is related to the interest or not.

Benefits of Targeted Google Ads

Users, publishers, and advertisers will all benefit from behavioral targeting. Users can get more relevant ads, publishers can get money out of their websites, and advertisers would find it easier to reach their target audience. Google plans to have 20 categories with 600 subcategories, excluding sensitive categories such as religion.

Users will not know that they are being tracked, or what is done with their data, but if they click an Ads by Google link, that will take them to a site which will explain why they are getting targeted ads. As to the reaction of publishers and advertisers, they are a bit wary about this new technology, simply because of the far-reaching power of Google. True, it might get them to have higher paid ads, but they are nervous about giving Google information about their customers. Of course, Google gives publishers the option of opting out from this behavioral targeting strategy.

Behavioral advertising is a type of marketing that targets the interests of consumers in order to effectively stimulate their minds and concentrate their interests on a product. It is widely used in the online world in order to draw customers into websites that are similar to those they demonstrate interest in by studying browsing habits and purchase habits. A great example of behavior advertising is the crawling ads along the side of your email box when you open your email. These ads use the keywords in your email to determine which advertisements will best catch your eye by looking at companies or common themes in you inbox.

For those who intend to market goods to consumers in the online world the field of behavioral advertising is one that you cannot afford to ignore. More and more companies are using these advertising techniques in order to widen their sales. Thus, if you fail to also utilize behavioral advertising tactics your competition will snag your potential consumers before you get a chance to advertise your own product. The simple truth is that behavioral advertising allows you to direct your message straight to your target audience instead of wasting time and money trying to find consumers who may be interested.

As mentioned, behavioral advertising is an online tactic that is being used more and more in the real world as well to snag consumers by showing them products and messages that they want to hear. For example, type in great toothache relief into Google and you will not only see peoples’ blogs and opinions, but in the sponsored advertising links you will also see many products that allegedly provide just that. This is the basic point of behavior advertising, to identify users’ needs before they have a chance to fulfill them elsewhere.

When it comes to actually using behavioral advertising you have two choices in the web world: either hire consultants to gather material and search to help you find your target audience and their browsing habits, or simply pay for services that deliver your content to an audience as the need arises. There are many online advertising venues that will deliver this information to you, with Google Advertising being one of the most recognizable. This is much easier and less time consuming than trying to compile enough information on your own to deliver the same results.

The outlook of behavioral advertising as a marketing field is only growing as the web continues to grow in popularity. The internet is quickly becoming the most relied on tool for news, information, and shopping in modern day society which pretty much ensures that behavioral advertising will be around fo years to come. Thus, if you want to fully exploit the internet and its many business opportunities it will pay to begin to employ and utilize behavioral advertising techniques at your company or in your own personal endeavors.