Business intelligence and analytics promise to deliver correct, nearly real-time information to decision makers, and the analytic tools help them quickly understand the information and take action. There are six analytic functionalities that BI systems deliver to achieve these ends:
- Production reports: These are predefined reports based on industryspecific requirements (see Table 12.5).
- Parameterized reports: Users enter several parameters as in a pivot table to filter data and isolate impacts of parameters. For instance, you might want to enter region and time of day to understand how sales of a product vary by region and time. If you were Starbucks, you might find that customers in the East buy most of their coffee in the morning, whereas in the Northwest customers buy coffee throughout the day. This finding might lead to different marketing and ad campaigns in each region.
- Dashboards/scorecards: These are visual tools for presenting performance data defined by users.
- Ad hoc query/search/report creation: These allow users to create their own reports based on queries and searches
- Drill down: This is the ability to move from a high-level summary to a more detailed view
- Forecasts, scenarios, models: These include the ability to perform linear forecasting, what-if scenario analysis, and analyze data using standard statistical tools.
Who Uses Business Intelligence and Business Analytics?
In previous chapters, we have described the different information constituencies in business firms—from senior managers to middle managers, analysts, and operational employees. This also holds true for BI and BA systems (see Figure 12.4). Over 80 percent of the audience for BI consists of casual users who rely largely on production reports. Senior executives tend to use BI to monitor firm activities using visual interfaces like dashboards and scorecards. Middle managers and analysts are much more likely to be immersed in the data and software, entering queries and slicing and dicing the data along different dimensions. Operational employees will, along with customers and suppliers, be looking mostly at prepackaged reports.
Production Reports
The most widely used output of a BI suite of tools are pre-packaged production reports. Table 12.5 illustrates some common predefined reports from Oracle’s BI suite of tools.
Predictive Analytics
An important capability of business intelligence analytics is the ability to model future events and behaviors, such as the probability that a customer will respond to an offer to purchase a product. Predictive analytics use statisticalanalysis, data mining techniques, historical data, and assumptions about future conditions to predict future trends and behavior patterns. Variables that can be measured to predict future behavior are identified. For example, an insurance company might use variables such as age, gender, and driving record as predictors of driving safety when issuing auto insurance policies. A collection of such predictors is combined into a predictive model for forecasting future probabilities with an acceptable level of reliability. FedEx has been using predictive analytics to develop models that predict how customers will respond to price changes and new services, which customers are most at risk of switching to competitors, and how much revenue will be generated by new storefront or drop-box locations. The accuracy rate of FedEx’s predictive analytics system ranges from 65 to 90 percent.
Predictive analytics are being incorporated into numerous business intelligence applications for sales, marketing, finance, fraud detection, and health care. One of the most well-known applications is credit scoring, which is used throughout the financial services industry. When you apply for a new credit card, scoring models process your credit history, loan application, and purchase data to determine your likelihood of making future credit payments on time. Telecommunications companies use predictive analytics to identify which customers are most profitable, which are most likely to leave, and which newservices and plans will be most likely to retain customers. Health care insurers have been analyzing data for years to identify which patients are most likely to generate high costs.
Many companies employ predictive analytics to predict response to direct marketing campaigns. By identifying customers less likely to respond, companies are able to lower their marketing and sales costs by bypassing this group and focusing their resources on customers who have been identified as more promising. For instance, the U.S. division of The Body Shop plc used predictive analytics and its database of catalog, Web, and retail store customers to identify customers who were more likely to make catalog purchases. That information helped the company build a more precise and targeted mailing list for its catalogs, improving the response rate for catalog mailings and catalog revenues.
Big Data Analytics
Many online retailers have capabilities for making personalized online product recommendations to their Web site visitors to help stimulate purchases and guide their decisions about what merchandise to stock. However, most of these product recommendations are based on the behaviors of similar groups of customers, such as those with incomes under $50,000 or whose ages are between 18–25. Now some are starting to analyze the tremendous quantities of online and in-store customer data they collect along with social media data to make these recommendations more individualized. Major online companies such as Walmart, Netflix, and eBay are analyzing big data from their customer transactions and social media streams to create real-time personalized shopping experiences. These efforts are translating into higher customer spending and customer retention rates.
EBay uses Hunch.com, which it acquired in 2011, to deliver customized recommendations to individual users based on their specific set of tastes. Hunch has built a massive database that includes data from customer purchases, social networks, and signals from around the Web. Hunch is able to analyze the data to create a “taste graph” that maps users with their predicted affinity for products, services, Web sites, and other people, and use this information to create customized recommendations. The Hunch “taste graph” includes predictions on about 500 million people, 200 million objects (such as videos, gadgets, or books), and 30 billion connections between people and objects. To generate accurate predictions in near real-time, Hunch transformed each person’s tastes into a more manageable “taste fingerprint” extracted from the larger taste graph.
Hunch.com’s prediction technology is helping eBay develop recommendations of items that might not be immediately obvious for users to purchase from its online marketplace. For example, for a coin collector purchasing on eBay, Hunch might recommend microscopes that are especially useful for coin analysis. Hunch could also become an important tool for eBay sellers if its customer profiles help them make better decisions about which items to offer, the content they use to describe their inventory, and perhaps even the advertising they use to promote their eBay listings
Many companies employ predictive analytics to predict response to direct marketing campaigns. By identifying customers less likely to respond, companies are able to lower their marketing and sales costs by bypassing this group and focusing their resources on customers who have been identified as more promising. For instance, the U.S. division of The Body Shop plc used predictive analytics and its database of catalog, Web, and retail store customers to identify customers who were more likely to make catalog purchases. That information helped the company build a more precise and targeted mailing list for its catalogs, improving the response rate for catalog mailings and catalog revenues.
Big Data Analytics
Many online retailers have capabilities for making personalized online product recommendations to their Web site visitors to help stimulate purchases and guide their decisions about what merchandise to stock. However, most of these product recommendations are based on the behaviors of similar groups of customers, such as those with incomes under $50,000 or whose ages are between 18–25. Now some are starting to analyze the tremendous quantities of online and in-store customer data they collect along with social media data to make these recommendations more individualized. Major online companies such as Walmart, Netflix, and eBay are analyzing big data from their customer transactions and social media streams to create real-time personalized shopping experiences. These efforts are translating into higher customer spending and customer retention rates.
EBay uses Hunch.com, which it acquired in 2011, to deliver customized recommendations to individual users based on their specific set of tastes. Hunch has built a massive database that includes data from customer purchases, social networks, and signals from around the Web. Hunch is able to analyze the data to create a “taste graph” that maps users with their predicted affinity for products, services, Web sites, and other people, and use this information to create customized recommendations. The Hunch “taste graph” includes predictions on about 500 million people, 200 million objects (such as videos, gadgets, or books), and 30 billion connections between people and objects. To generate accurate predictions in near real-time, Hunch transformed each person’s tastes into a more manageable “taste fingerprint” extracted from the larger taste graph.
Hunch.com’s prediction technology is helping eBay develop recommendations of items that might not be immediately obvious for users to purchase from its online marketplace. For example, for a coin collector purchasing on eBay, Hunch might recommend microscopes that are especially useful for coin analysis. Hunch could also become an important tool for eBay sellers if its customer profiles help them make better decisions about which items to offer, the content they use to describe their inventory, and perhaps even the advertising they use to promote their eBay listings
No comments:
Post a Comment