Instructor: Jing Gong

Weekly Question #9 (Due Friday, November 20)

Leave your response as a comment on this post by the beginning of class on November 20, 2015. Remember, it only needs to be three or four sentences. For these weekly questions, I’m mainly interested in your opinions, not so much particular “facts” from the class! If you sign in using your AccessNet ID and password you won’t have to fill in the name, email and captcha fields when you leave your comment.

Answer one of the two questions below (not both):

  1. Name and describe a business question that you could answer using clustering. What data would you collect to perform the analysis? Don’t use an example we’ve covered in class.
  2. Let’s say you performed a cluster analysis and came up with a series of market segments, describing the demographics of three different sets of customers for your business (big spenders, average spenders, small spenders). Your manager is reluctant to trust your analysis because she has a lot of experience and this contradicts what she believes is true about her customers. How would you convince her that your analysis is correct?

18 Responses to Weekly Question #9 (Due Friday, November 20)

  • A business question that could be answered using clustering could be to identify our customers prescription history using our database at my family’s local pharmacy. We could begin to store our customers prescriptions into our database as well as other crucial details about the customer that we would like to keep on file. We would be able to tell what brand of the drugs the customers like. This would allow us to predict the customers basic desires which would allow us to really complete the experience and make sure they come back to the pharmacy. This is how we would use clustering to benefit us.

  • A business question that can be answered using clustering is for a retail store such as Express or American Eagle and how many of their customers use their membership cards for discounts and coupons. The data needed to do for this analysis is the customer’s name, how often they’ve been to the store, membership card ID, and how often they used the card.

  • 1. Clustering could answer the business question of where to advertise a business’ products. For example, by using clustering, a company could find out what web sites their target market it most likely to use for the longest amounts of time. Information such as age would be important because advertising to teenagers would be more effective on Twitter or Instagram then say, Facebook, which now has many older users. Other information such as gender, past online searches and geographic location would be useful.

  • 1. Clustering could be used in a business setting to determine the correlation between types of clients that a business might appeal to. As apposed to looking at what products could be marketed together, a business could use clustering to better target their marketing campaigns toward individuals or groups who are similar to their current customers.

  • You could use cluster analysis to answer business questions regarding how costumers felt about a product or service. You could use data such as their ratings and other data about the customers in order to identify any problems and betteryour company

  • An academic question you could answer using clustering is: Which students in a given class are close to a 4.0 GPA? The data you can collect to perform the current analysis is every student’s current GPA and their projected grade in the given class.

  • A business question that could be answered using clustering could be how much specific medical equipment sells for based on the quantity that is sold. Given that there is generally room for negotiation, some clients will pay less than others and some will pay full price. Given multiple clusters on multiple equipment pieces sold at multiple prices, one can see which product is selling the best and most closely to its asking price.

  • 1. Name and describe a business question that you could answer using clustering. What data would you collect to perform the analysis?

    A business question I can answer using clustering is to find out whether consumers who purchase merchandise for high end prices are loyal. In order to answer this question, I would collect data from high end price merchandise in the store and calculate how many people purchased this product. In addition, I would collect information on how frequent the consumer made purchases. Furthermore, I would find out what style of clothing they were purchasing in order to keep them coming into the store.

  • A good example of how we could use clustering to answer a business question would be relating to grocery store loyalty cards. Say they want to know what type of shopper that you are. They could cluster shoppers into 4-5 broad categories based on what they buy. Some shoppers may only go in to buy soda and snacks, while other will buy more meat, and still others will focus on grains and veggies. If you’re wondering which items to give coupons / discount for which customers, you can answer using clustering in this way.

  • Clustering could be used to find what ages are buying tickets for a concert. People could be grouped by similar ages to view what concerts children are seeing vs. adults. Some artists attract an older crowd and some attract a younger crowd.

  • A business question that could be answered using clustering is what clothing items are most likely to be bought together in a retail store. By tracking and analyzing customer purchases, and classifying the data by seasonality, a business can have a better idea of how to layout their stores throughout the year in order to maximize complementary purchases by customers. By placing items that are bought frequently together near each other, customer time in the store will be reduced, satisfaction will be increased, and repeat purchases due to convenience will lead to more business for that retail store.

  • A business question that you could answer using clustering would be product value over time. the data that you would need for this would be the times at which the products were purchased and the quantity of which were purchased. Depending on how the cluster displays you can tell as to weather the product is losing its value or increasing.

  • One way a business could using clustering to help enhance their business is clustering age groups to see who uses technology, the results from this can transfer into how businesses are going to go forth in marketing their products. If a certain age group is clustered to use a certain type of technology more, the company can use that technology in particular to reach out to them.

  • NFL recruiters could use clustering to determine what players to draft. They could collect data on colleges attended, position, and player stats to help pick out top prospects for the NFL. For example, they could use clustering information to find college quarterbacks with the most potential to be successful in a professional career.

  • I would do a cluster analysis on a clothing boutique. I would segment loyalty card customers in 6 various groups according to their buying behavior. I would use variables like customer age, gender, location, favorite brands, style of clothing, etc. By doing the cluster analysis we would be able to identity similarities with respect to certain customer behaviors.

  • One question Apple Inc. could answer is what state orders the most iPhone 6s’s. They could cluster orders of the iPhone by state then determine which cluster is the largest. I would collect data from online sales and in store sales of the iPhone 6s’s and see which region produces the most sales.

  • A business question that I could answer using clustering is if i was part of Microsoft, I can find out what age of consumers buy Microsoft products the most. I could group them into clusters by age and find out what age bracket is the most.

  • A business question that we could answer after this example would be, “Would a shopping mall be more profitable if they redesign and position stores differently to attract specific market segments instead of their current layout? For example, lets say a large shopping mall wants to see the various types of people coming to the mall, shopping at certain stores, and how to attract more people in addition to the current customers. The mall owners know that there are a few different market segments, and they want to know if the should re-design and position the stores differently/alter their offered services (events and discounts) to attract more profitable market segments. Some data that we would need to answer this would be: “Do customers prefer to have “fun” and be social while they shop?”, “Do customers go in the mall looking strictly for deals/ do they shop with a specific budget in mind?” “How often do the customers visit the mall?” “What is the customers household income, and how does that affect their shopping habits/time spent in the mall?” After obtaining the results from the data, we could use cluster analysis to determine the best business decision for this scenario.

  • I would use clustering in a social network analysis. Specifically, if i wanted to recognize communities of large groups of people to market to. Clustering can be used to solve product research, and determine which groups of people would be more likely to purchase the product.

  • A business question that you can answer by using clustering is finding out why people would donate the most based off information on where they live and what type of person. For example in a non profit organization you’re consistently asking for donations from many people and there will be many people who say no more than yes. So by by understanding which type of people would donate you can increase your chances of getting more donations from those people. Of course you wouldn’t ignore the people who you think would say no however by targeting the people who are more likely to donate you can increase the funds you’re getting through donations and utilize it earlier.

  • A question that you could answer using clustering would be how could you prevent a catastrophic loss as an insurer for flood insurance? The data you would collect to perform the analysis would be the number of houses with flood insurance in a given area. To avoid a catastrophic loss, the insurance company would diversify the risks into different flood regions.

  • A business question that you could use clustering to solve could be which ad spots on TV reach which audiences. Telecom companies are able to collect data on which demographics watch each show, as well as collect viewership ratings. This data would allow companies to sell specify the ad space it sells, and charge a higher price for space where there is large viewership and a desirable demographic. If my manager was against using data clustering, I would walk him through the data to show him historically the accuracy of using this method.

  • There is a case I learned from my Risk and Insurance intro class. There is a conception called Group Insurance and here are many methods to control adverse selection. The first thing we concern is the reason the group exists. For example, all people from this group are from the same company, they are Employer/ Employee based group, or it’s a professional associations, alumni associations even veterans groups. Usually each group focus on one specific area to insure. Why we need to put similar people into a group, the reason is the more similar they are in the same group, the lower risk this group faces. For example, if we provides a car insure, we should put the people who just got drive license 3 years ago together or put the people have more than 2 car incidents within one year together, because the risk of those groups are higher than that with no incident for over 10 years driving time. We should look at their age, city, driving time, driving behavior, incident background and so on.

  • We can use clustering in demographic population dispersion. For example, we can use clustering to know the distance of residential housing from the center city, and we compute the population’s gradient according to the distance gradient, this way we can know how many people living far from center city or close to center city.

  • Clustering could identify a person’s preference for a certain sandwich to order at Subway. Subway can collect data on the demographics of people’s ethnicities in certain areas. This can allow Subway to figure out certain people’s upbringing and taste preference. Also, they can figure out what time during the days that people would buy a certain sandwich so they can stock up on the bread or material needed to make the sandwich. This can help Subway be more efficient and increase their sales.

  • An example of a business question that can be answered based on clustering modeling is when a specific demographic should be marketed to based on social media accounts. For example companies have started tracking and clustering certain demographics of who is on social media during certain events on tv or occurring in real life. As a result the business are able to analyze when to market their product at what optimal times on social media websites so a person is more likely to click on the advertisement. You can also see the max number and minimum number of people on social media at certain times which will help your decision as well.

  • An example of business clustering is in insurance. Here they cluster you based off of your risk in the risk pool. From there they cluster you into whether they will insure you or not and how much they will insure you for.

  • We can use clustering in an attempt to look at cars sold in a specific region. We can groups the cars acording to the location by the style that was sold. We can look at Range Rover and what style was popular in the Philadelphia Area. We may be able to look at the distance between dealerships and determine if sales would be affected by opening another location with in a certain region.

  • Clustering can be used to determine the client relationships. And you know how to combine the production together to find the specific data and information you need. Also, the clustering can find the employee’s performance, which could collect the data to find the evaluate the employee. Besides, through the clustering, the company could find the customer needs, which can make the company more efficient.

  • Data clustering could be used in a lot of ways to answer different set of business analysis problems. Professional sports organizations could use data clustering to see which games are attended the most and maybe specifically which players the fans came to see. This data could be used to sell jerseys of those specific players for a greater revenue.

Leave a Reply to Ben Bucceri Cancel reply

Your email address will not be published. Required fields are marked *