Business Analytics: Facebook, Caesars Entertainment, Online Gambling, and Amazon

From New Media Business Blog

Jump to: navigation, search
Business Analytics

According to the American multinational developer of analytics software, SAS Institute, business analytics is a "Multidimensional field that uses mathematics, statistics, predictive modeling and machine-learning techniques to find meaningful patterns and knowledge in recorded data.”[1]

Contents

Business Analytics

Why are they Important?

Business Analytics are important because it can help companies uncover correlations and patterns that can help them make better business decisions.

It can also help answer the following types of questions regarding business outcomes:

  • What happened?
  • How or why did it happen?
  • What is happening now?
  • What is likely to happen next?

What can Analytics do?

Business analytics can help with a variety of business activities that can help a company discover the big-picture and what really motivates customer behavior.


This video shows how the staff at Walmart use business analytics in their day-to-day operations:


Walmart Global Customer Insights and Analytics

Three Types of Analytics

There are three types of analytics to be aware of:

  • Descriptive Analytics: Descriptive statistics have been around the longest. They are a summary of collected data points. These are the models that will help you understand what happened and why. For example, everything from how many clicks a page receives to how many units are produced vs. how many are sold.
  • Predictive Analytics: Predictive Analytics involves running hundreds or thousands of models quickly with past data and predictive algorithms to help you determine the probability of what will happen next. It is most often used to predict consumer behavior.
  • Prescriptive Analytics: Prescriptive analytics answers the question of what to do by providing information on optimal decisions based on the predicted future scenarios.

Who Works with Analytics?

There are various types of staff members that are involved in Business Analytics. A few of them include:

  • Business Managers: They identify the issue, makes decision based on analysis, and monitors return based on the business decision.
  • Business Analysts: They conduct data exploration/visualization and work to identify key variables that influence outcomes.
  • IT/Data Management Teams: are usually in charge of facilitating data preparation, model deployment, and data monitoring.
  • Data Scientists/Miners: are responsible for performing complex exploratory analysis, descriptive segmentation, and predictive modeling.


Companies and Industries

Facebook

Facebook Logo

Facebook has the ability to collect all kinds of information about its users as it is shown in its privacy policy. The policy states that:

"We collect the content and other information you provide when you use our Services, including when you sign up for an account, create or share, and message or communicate with others. This can include information in or about the content you provide, such as the location of a photo or the date a file was created. We also collect information about how you use our Services, such as the types of content you view or engage with or the frequency and duration of your activities.”[1]

As a result, when Facebook users accept the terms and conditions of the privacy policy they are essentially allowing the company to collect information about many aspects of their lives.

How Private is the Information?

Some of the information people give to Facebook when they fill out their profile is public, such as their age range, language and country. Facebook use a part of their profile, called Public Profile, to help connect people with friends and family. The Public Profile includes their name, gender, username and user ID (account number), profile picture, cover photo and networks.

If other people share information about someone, they can choose to make it public without that person's consent. Also when people comment on other people’s public posts, the comment is public as well.

Facebook Pages and public groups are public spaces. Anyone who can see the page or group can see the post or comment. Generally, when people post or comment on a Page or to a public group, a story can be published in News Feed as well as other places on or off Facebook.[2]

By default, Facebook tends to make everything users put on its network fairly public. However, people have the option of changing their privacy settings and decide what information they want to make public, share with friends, or keep for themselves.

This video shows how simple it is to get people's information from a simple Like on Facebook:


How Private is your Personal Information?

Facebook's Cookies

Facebook tracks data in different ways for members who have signed in and are using their accounts, for members who are logged-off, and for non-members. The tracking process begins when people initially visit www.facebook.com. If they choose to sign up for a new account, Facebook inserts two different types of tracking cookies in your browser, a "session cookie" and a "browser cookie." If they choose not to become a member, and move on, they only get the browser cookie.[1]

Session Cookie

Facebook's Cookies
  • Works when someone is logged-on Facebook and surfing the web
  • Tracks and logs user's name, e-mail address, friends and all data associated with their profile on Facebook
  • Collects IP address, screen resolution, operating system and browser version
  • Compiles a running log of all webpage visits for 90 days, continually deleting entries for the oldest day and adding the newest to this log

Browser Cookie

  • Works when someone is logged-out Facebook and surfing the web or is a non-member
  • Tracks and logs a unique alphanumeric identifier (a number unique to the cookie), but no personal information
  • Collects IP address, screen resolution, Operating system and browser version
  • Compiles a running log of all webpage visits for 90 days, continually deleting entries for the oldest day and adding the newest to this log


Facebook Tracking

Facebook uses the tracked data to strengthen security, increase ad revenue and enhance user experience.[2]

Olé - Argentine sports news site using Facebook plug-ins

Security

Facebook argues that the cookies are primarily use as security tools. The company says that without cookies every visit could be considered an untrusted login and people would have to go through a complicated log-in process to prove that they are the legitimate owners of their accounts.[2] Therefore, cookies help by:

  • Preventing the creation of fake accounts
  • Reducing the risk of users' accounts being taken over by other people
  • Protecting users' content against theft

Ad Revenue

Advertising revenue is Facebook's biggest source of income.[2]

  • In 2015 approximately 45% of the company’s income came from advertising revenue
  • 80% of the 45% came from mobile ad sales

User Experience

Facebook offers to put Like buttons and other social plug-ins all around the web. Companies that allow them are able to monitor and control their web traffic more efficiently and promote its products or services via the social networking service. By doing this, Facebook is:

  • Improving the usability of the like buttons
  • Revamping plug-in services of third party websites,
  • Making advertisement more relevant

However, these plug-ins allow Facebook to collect information as soon as the page is loading. There is no need to click on the like button to let Facebook store user's information. Thus, if enough sites participate, Facebook could assemble a vast amount of data about Internet users' browsing habits even if they are not logged on their site. Therefore, some people say that by putting a Like button on a site, the site is potentially selling out users' privacy even if they never pressed that button.[3]


Facial Recognition Technology

Tracking Methods

Now, Facebook has been investing in image processing and face recognition capabilities that allows the company to know how people look from the photos they have shared on their profile. [4] They have the ability to explore other Facebook profiles to find a picture of a specific person. Therefore, some critics say that eventually a photo will be enough to find someone all over the internet.

Also, personal data is exchanged when people use apps like FarmVille through Facebook. It has even been reported that signing up for these apps will give those companies who created the apps access to user's personal information.[5]

Facebook is developing more sophisticated tools to measure the effectiveness of ads. For example, the company is considering how long users look at a post instead of relying on users to click “like”, making a more complete analysis of the users’ behaviour.[6]


This video illustrates how Facebook gain access to all the information that people have on their phones by using its Messenger app:


Facebook's Messenger app

Data Science Team

Data Science Team Facebook Page

A team that has developed ways to predict the intelligence of users, their political views, and even their emotional stability. They are also able to predict weeks before it is happening, that a user will change their relationship status from 'single' to 'in a relationship' by analyzing his/her activity (messages, posts, likes, comments, shares).[1]

As an example, the Data Science Team conducted a study of dog people versus cat people and they found out that:[2]

Dog people have more friends

  • Dog people are more outgoing, measured in terms of Facebook friends
  • On average, dog people have 26 more Facebook friends than cat people
  • Cat people get invited to more events

Dating: Yes, cat people are more likely to be single

  • About 30% of cat people are single, compared to just 24% of dog people
  • Being single and a cat lover isn't related to age or gender

Books, TV, and movies

  • Cat people seem to have more indoor activities: They disproportionately like books, TV, and movies (measured in terms of Facebook Page likes)
  • Cat people are especially fond of fantasy, sci-fi, and anime
  • Dog people like love stories and things about dogs


This study can be considered fun. The problem is that the data science team conducts all sorts of studies without people knowing they are using their information:


Data Science Team Various Studies

Caesars Entertainment

Caesar's Entertainment Logo

Caesar's Entertainment (formerly Harrah's Casino) was one of the first retail casinos to really identify the value of using big data. Former CEO Gary Loveman was the "pioneer" of utilizing analytics to ensure big players were kept happy, and were returning to the casino. Most data gathered came from their "Total Rewards" loyalty program. In addition to using this program, casinos such as Caesar's entertainment also have access to cameras that tell important stakeholders which areas are hot, and which games could benefit from more tables being open.

Total Rewards Loyalty Program

Casinos rely mainly on loyalty programs to track their customers. These programs are completely voluntary, so unlike other companies in other industries, you know where your data is going once you sign up. These programs give an incentive to players, as they can accumulate points to earn rewards within the casino, and redeem them. To quote former Caesar's CEO Gary Loveman, “You play more, you get treated better. You play less, we still treat you well, just not quite as well.” Spending more money at the casino would lead to an increased level of rewards for the consumer if they possessed a loyalty card, so there is no incentive not to apply for one. Loyalty programs can span outside the range of a casino setting as well. In the case of Caesar's loyalty program, they were able to partner with VISA to create a Total Rewards credit card. This provides additional information to casinos, as Caesar's would now be able to see a portion of your spending habits, based on the amount of points that are gained. Without loyalty programs, casinos would have a tougher time tracking the habits of their clientele.


Caesar's Total Rewards Program

Information Gathered by Casinos

The main piece of information that casinos are trying to gather from their customers is the PV, or player value. The player value is calculated by different variables, such as how often you visit, how long you play, and what your denomination preferences are when you decide to play. All of this information is combined with the end goal of assessing whether or not it is worthwhile for the casinos to spend their time on you. For example, if you were a daily gambler who spent $500 every day, the casinos would be more likely to send you coupons and other exclusive offers than a player who plays once a month for $20. Casinos primarily do not focus too much on whether a player wins or loses, as there are inherent biases that favour the house in every game. These house advantages implies that there is a price paid everytime someone plays a game.

Crowd Contouring

Crowd contouring is a method in which casinos utilize heat cams to determine which areas are hot, and which areas are cold. As more people group around one table, the area corresponding to the popular game becomes red, and managers may make a decision on whether or not to open another table. This method is particularly interesting as it gives the casino an opportunity to test out certain layouts of where games should be. Over time, this information gathered can allow casinos to maximize their profits by becoming more efficient. Another type of crowd contouring is one where the heat map is overlayed onto the city that the casino is in. As the map changes colours over time, casinos are able to accurately predict when people from a certain demographic, age group, or even neighbourhood will come for a visit. From the information gathered, the casino can send out timely promotions to certain districts within the city, and have the knowledge of when exactly the most popular time is for certain groups to visit, and tailor their offerings to the market segment.

Heat map of casino floor
Heat map of U.S city

















Online Gambling Industry

Information Gathering

Online gambling sites gather information from their users from the very beginning when they register. When registering, basic information is required and this allows the sites to have information such as your name, birth date, address, email address and etc. Next, before users begin play, money must first be deposited. When money is deposited, information such as how much money users are willing to play with, credit card information and where the money is coming from is gained from this. Once everything is completed and users have won money, users would like to withdraw money and before users are able to do so, they have to give 2 pieces of information to the sites, their birth certificate or passport and a government ID.

Information Storage

Information gathered by the gambling sites are mostly stored in database warehouses created using Apache Hadoop and Apache Hive.[1] Once the data is stored in the warehouses, data analyst will mine the data and categorize them into several categories, for example, product data, spending patterns, demographic and betting history.

Uses Of Information

Customer Acquisition

When customers are visiting the website, data is collected by cookies. Data like this can be used to help identify who are the most likely to respond to a sales offer from the type of games users play and the parts of the sites they visit most often. Sales offers on the sites include free betting vouchers and also enhanced odds for the newer players to entice them to place their first bet. Customer recognition can be further improved with this information and it is beneficial to gambling sites because it prevents them from giving out incentives to the wrong people, ultimately, wasting their time and money. [2]

Up-sell and Cross sell

With all the information, gambling sites also use it to understand which games are played together or bet together based on the data. According to Finneran, understanding which games are played together will help gambling sites to identify important cross sell opportunities for higher value games.[3] For example, gambling sites will promote games such as slots or even promote their mobile applications with advertisements while users are deciding on their bets.

Next Product Offering

After gathering information, gambling sites have a good understanding that gamers are attracted to play games in a certain order. Predictive analysis models are used in this case and they help identify these patterns. Patterns from the analysis help to guide gambling sites offer the next most appropriate next product which ultimately, help maximize their profits. Examples of this would be, gambling sites providing coupons to individuals who have lost money on sports bets and offering them coupons for playing poker or blackjack.[3]

Attrition and Churn

Information is also used to help prevent the attrition and churn on the website. For online games churn of users is very easy because once people lose their money, they will stop playing altogether so in order to entice the user to come back rewards should be given to them. To help aid in identifying the churn for the site, Churn and loyalty models are used. These models are based on two definitions, product churn and also customer churn. Product churn being not losing the user as a client, but only losing revenue from that user in a particular product and customer churn being losing the customer to the competition.[3]

Lifetime Value

Lifetime value which is one of the most important information needed by gambling sites to estimate the amount of money they can earn from a user. With this in mind, they will be able to figure out which individual user should they turn more of their attention to and offer them more deals and rewards. When estimating the lifetime value, data collected helps to estimate the expected revenue from each customer over their lifetime on the site. According to Finneran, this is done on the basis of the expected products which are the games, cost marketing, cost of account management, and their expected lifetime from the day of first product acquisition to churn. [3]

Risks Models

Data collected are also used to generate risk models. With risk models, it helps to prevent credit risk and also fraud. For certain sites who do give out credit to some of their users, they access credit risk by how often the user is able to pay back their loans. Similar to the banks, gambling sites will also cut of the user’s ability to borrow more credit until they pay back their loans. For fraud, gambling sites access them by using all the information given to them by the users. The most common frauds conducted by users would be money laundering, collusion and credit card fraud, so, the accurate prediction of these frauds are crucial to the sites. An example of how a user goes about doing this is by making registering accounts with fake information and then using it for illegal activity. [3]

Player Risk Prevention

Gambling sites also help to access the player risk. Player risk is when a player is at risk of putting themselves into a detrimental financial position after losing a lot of money. Data collected is used to access the player’s pattern of play and sees how much money they have lost within a certain period of time. Once the amount of lost have surpassed a set amount then the organization steps in either sending a message asking them to stop and access how much they spent or by sending them a promotion that stops them from gambling. Most of the time, acts of kindness like this are not always altruistic as they are usually required by the government so, it differs from site to site on the methods they use.

Ethical Issues Surrounding The Industry

Private Information gathered

Information be it private or even basic information, it is all being gathered by the sites. Whether it is right or wrong for them to have all this information at their disposal is still up for debate but, one thing is for certain, they are using this information for their personal gain and like Facebook they are really close to the ethical line of whether or not it is right to benefit off other people’s information. [4]

Game Fraud

Another big issue surrounding the online gambling industry is game fraud. Game fraud is when a site designs game in a way to take advantage of the user’s urge to gamble and making it highly in favor of the site instead of making it even. By the time the user realizes the problem, it is already too late and they would have lost more than expected.[4]

Gambling Addiction

With the introduction of mobile applications for gambling websites, gambling has never been easier. With this said, the ease of online gambling has caused many individuals to become addicted. According to a study done by the state university of New York, “more than 80 percent of Americans gamble every year and between three to five percent of Americans have a gambling problem”.[5]Furthermore, according to the state university of New York, “An estimated 750000 of U.S. youth – between the ages of 14 and 21 – are problem gamblers”.[5]With this in mind, especially since youths are the most prevalent users of mobile devices, that is why the accessibility has caused them to become addicted.

Amazon

Amazon Logo

How does Amazon operate?

  • Amazon is a Marketplace Model. An online marketplace (or online e-commerce marketplace) is a type of e-commerce site where product or service information is provided by multiple third parties, whereas transactions are processed by the marketplace operator. Online marketplaces are the primary type of multichannel ecommerce and can be described as a "simple and convenient portal" to streamline the production process. In an online marketplace, consumer transactions are processed by the marketplace operator and then delivered and fulfilled by the participating retailers or wholesalers (often called drop shipping). Other capabilities might include auctioning (forward or reverse), catalogs, ordering, wanted advertisement, trading exchange functionality and capabilities like RFQ, RFI or RFP. These type of sites allow users to register and sell single items to a large number of items for a "post-selling" fee.[6]
  • Direct Amazon-to-Buyer sales approach.
  • Multi-leveled e-commerce strategy.

There is now a program that lets those "associates" build entire websites based on Amazon's platform. They can literally create mini Amazon websites if they want to, building on Amazon's huge database of products and applications for their own purposes. As long as any purchases go through Amazon, they can build a site called Amazonish.com, pull products directly from Amazon's servers, write their own guides and recommendations and earn a cut of any sales. Amazon has become a software developer's playground.[7]

Amazon & Data

Customer tracking is an Amazon's stronghold. Amazon tracks everything people do on their website.
The collected data allows Amazon to offer:

  • Personalized recommendations
  • Price optimization
  • Targeted marketing

Sections customer would be familiar which includes:

  • Wish List
  • Browsing history
  • Recommendations for you
  • Related items you have been viewed
  • The other buyers who also bought the items that you view


Personalized Recommendation

Additional 10%-30% revenue from suggestions.

  • Provides recommendations on inventory stock to sellers.
  • Provides suggestions on new product choices they should add to their inventory.
  • Gift giving recommendations can be used for the customers never visited Amazon before.

A recent development in customer tracking actually collects information on people who may have never visited Amazon.com. Amazon's gift-giving recommendations collect data on the stuff you buy for other people.[7]. For example, if someone buys a toy train set in December and ships it to their friend's son, Amazon knows that they gave gifts to a boy aged four to 10 who lives somewhere. When the next holiday comes, Amazon will give the buyer all the sorts of ideas what they should get for their friend's son.
The factors would be considered into the Algorithm:

How Big Data Is Used In Amazon Recommendation Systems To Change Our Lives
  • Purchase history
  • Browsing history
  • Impact of friends
  • Trends
  • Social media mentions
  • Purchases made by customers with similar purchase history

Real-time Product Pricing

Shopping on Amazon is more like playing the stock market than shopping at a local big box store. Prices don’t just change with the season or when something goes on sale. They fluctuate as often as every 15 minutes. Amazon has long determined prices using lots of fancy math. But an increasing number of the more than 2 million third-party merchants in the Amazon Marketplace also rely on algorithms to take stock of supply, demand, and prices offered by their competition, then adjust their own prices, in real time.[1].
In the year 2000, Amazon was caught with charging different customers different prices for the same item and claimed it was for “random price tests”. Finally, refunds were given and an apology was provided.
Now there is an implemented price protection policy. If the customers find a lower price compare to the price they paid for, they could get the refund within seven days. Be sure to keep the receipt.

Amazon and Advertisers

Amazon has its In-house platform targets ads to people who have visited and then left Amazon’s sites.
Amazon does not directly give advertisers access to its trove of information about individuals’ data, instead they would create target audiences, such as people who recently purchased digital cameras. A marketer selling digital camera accessories could then use Amazon’s technology to show the ads to the people in that category.

Anticipatory Shipping

Amazon optimizes delivery time through anticipatory shipping. The following video shows how anticipatory shipping works.


How does "Anticipatory shipping" works?

Comments on "Anticipatory Shipping"

  • Ecommerce expert and CEO of Venda, Eric Abensur, predicts "anticipatory shipping will be an undeniable, market-dominating triumph. If it isn't for one teeny issue, that is. “If it were indeed possible to accurately predict that a customer will purchase the product they’re searching for, then anticipatory shipping could increase productivity and revenue.”. “Unfortunately, the nature of a prediction is that it isn’t an exact science.”[1]
  • Leon Brits, also from the digital commerce field, said “One of the biggest issues online retailers are trying to solve is reducing their returns rates, as the postage cost is often picked up by the retailer and eats into their margins. Anticipatory shipping can only lead to an increase in returns and repackaging, and so doesn’t seem like a fantastic idea.” [1]


The images here are the flow charts for the "Anticipatory shipping".

Anticipatory Shipping Chart 1
Anticipatory Shipping Chart 2

The Future of Business Analytics

Business Analytics is leading to a more personalized consumer experience. With a vast amount of personal data now available to companies, they will be able to predict a customer's behavior with great accuracy and precision.

As stated above, a simple action such as "liking" a Facebook page transmitted a great amount of personal information about coffee drinkers. Could there be a reality in the future where companies could know a great deal about you without any action required on your part? Time will tell.

References

  1. 1.0 1.1 http://www.mycustomer.com/selling/ecommerce/amazons-anticipatory-shipping-innovative-or-insane
Personal tools