Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
Upcoming SlideShare
Apache Flink: Real-World Use Cases for Streaming Analytics
Download to read offline and view in fullscreen.


Case Studies Utilizing Real Time Data Analytics

Download to read offline

The goal of this paper is to explore executive perceptions and opinions about real time data applications and operations. We have interviewed over forty (40) Key Innovation Leaders who have been cited as the most innovative thinkers within the world of analytics and have documented distinctive case studies which have clearly optimized business intelligence.

Related Books

Free with a 30 day trial from Scribd

See all

Case Studies Utilizing Real Time Data Analytics

  1. 1. Case Studies Utilizing Real-Time Data Analytics
  2. 2. The goal of this paper is to explore executive perceptions and opinions about real time dataapplications and operations. We have interviewed over forty (40) Key Innovation Leaders whohave been cited as the most innovative thinkers within the world of analytics and havedocumented distinctive case studies which have clearly optimized business intelligence.With every passing day, the pace of business is accelerating. Paradigm shifts in the globaleconomy combined with rapid technology adoption are forcing business executives to contendwith a new global marketplace fraught with uncertainty and constant change. To be successfulunder these new market dynamics, firms must adjust operational processes, corporate strategiesand business models at lightning speed—allowing them to leverage intelligence instantly andtake immediate action. At the same time, they must make sure their decisions are informed byproper data and analysis.The interviews and research conducted focused on large and medium-sized business units in foursectors of the economy—consumer products, technology, oil and gas, and retailing.The data revealed some interesting findings:  Executives realize they need to take their businesses into real time. In fact, 30% of firms already derive considerable benefit from real-time business, and nearly two-thirds of companies yet to implement real-time business techniques plan to do so over the next five years. With at least one-fifth of companies in every sector or region using real-time business techniques, its value to businesses of all types is growing. For the majority of the businesses that are behind these leaders, the imperative is to make up lost ground.  There are some surprising leaders—and laggards. The majority of oil and gas firms have implemented a real-time business approach, particularly as part of their production processes and financial and business risk management strategies. This indicates the importance of real-time operations particularly for complex, capital-intensive firms. Consumer product and retailing firms, meanwhile, lag in terms of implementation. Retailers in particular are behind in the implementation of real-time operations for customer experience and supply-chain management, two critical areas for retailing success. Retailers also report their efforts around real-time business to be less effective than firms in other industries. 2
  3. 3. Real-Time Business in the New Global Marketplace  Early adopters have seen substantial results. The main strategic goals of executives choosing to implement real-time business techniques are to increase market share and enhance service and quality advantages. Operationally, real-time business is proving especially effective in delivering improvements in customer experience, production processes and supply chain management. But perhaps most striking is the tangible return on investment: Those able to estimate put revenues gains at over 20%, and cost reduction at nearly 20%. In fact, future gains are expected to exceed earlier ones, with revenue increases of 28%.  But challenges are significant, in a number of areas. Respondents cite a lack of technology as a main obstacle, as well as suppliers’ unfamiliarity with real- time systems and a lack of internal expertise. Indeed, the move to real-time operations can be extremely complex, and will require careful planning and strategy to ensure success.  Those planning to roll out real-time systems have different strategies than early adopters. They will put particular focus on financial and business risk, production processes and supply-chain management. Meanwhile, almost all early adopters plan to carry on investing, though they will shift their focus from sales and marketing towards product design and innovation, while continuing to focus on customer experience.The need for speedWe live in a world that increasingly communicates and operates in real time. Over the pastdecade, the spread of social media and mobile phone use, combined with the explosion ofaccessible information and increasing interconnectedness of global markets and cultures, hasdramatically sped the pace of our daily lives. The evidence is clear in the dissemination ofinternational news: For example, when US Airways Flight 1549 emergency-landed on NewYork’s Hudson River after striking a flock of geese during takeoff in January 2009, it took onlythree minutes for the first “tweet” to alert the world 3
  4. 4. The pace of business, too, has dramatically accelerated. Customers and clients expect theirqueries and problems to be addressed immediately. In an uncertain global market where there aremany more competitors and constant change, executives know that success depends on theability to take fast action and make rapid decisions. Indeed, in the business world today, speed isnot only expected, but required. As Rupert Murdoch noted: “The world is changing very fast.Big will not beat small anymore. It will be the fast beating the slow.”But are companies prepared to take their firms into real time?While nearly all respondents agree that doing so is critical, only one-third of businesses havealready implemented real-time business applications in some way (though 65% of those whohave not yet implemented real-time systems expect to do so within five years). And theexperience is far from uniform. For example, companies that focus on business-to-businessoperations (35%), are more likely to have moved to real-time than those in the business-toconsumer space (16%). And almost half of the very large businesses studied— those with annualrevenues in excess of US$25 billion—have introduced real-time business, significantly morethan those with a turnover of less than US$1 billion (22%) or between US$1 billion and US$10billion (28%).Adopting a real-time business approach can provide companies with myriad benefits atboth the operational and management levels  At an operational level : by speeding up data capture and simplifying processes, executives can reduce inventories, minimize business risks, lower operational costs, accelerate speed to market, foster productivity and better meet customer needs.  At the management level, by accelerating decision-making and planning, executives can exploit market opportunities faster, identify competitive threats sooner, cope with market shifts more quickly and transform stagnating businesses.  With the internet transforming business, mobility moving to center stage and on-demand computing becoming the norm, global firms are on the threshold of an era that will radically transform operations. Companies need to consider how to adopt real-time into the fabric of their firms, or risk falling behind their competitors. 4
  5. 5. What does it mean to be a “real-time business”?For purposes of our research, “real-time business” refers to processes that allow companies toconduct a range of business activities instantaneously. The activities that make up realtimebusiness can, therefore, incorporate all aspects of business, including gathering and acting onbusiness intelligence, developing promotional and marketing tactics, controlling and adjustingproduction processes, managing inventory, identifying and managing business risks, closing andfulfilling sales and meeting customer needs.Real-time operations depend on in-memory analytics, which takes a different approach thantraditional business intelligence systems. Rather than storing information on various externaldisks and caching bits of data in a computer’s random access memory (RAM), real-timeintelligence puts the data directly into RAM. Advancements in computing power and storagehave made it possible to store vastly more amounts of data in RAM than ever before, and allowsfor extremely fast query responsesWhich companies are ahead of the curve in adopting real-time operations? The answers aresurprising. According to respondents in the four industries surveyed, oil and gas firms (54%)take the lead, particularly in production processes and assessing financial and business risk.In sharp contrast, only 20% of consumer products and 29% of retailers have moved toward real-time operations. 5
  6. 6. In the oil and gas industry, real-time data is critical to monitor drilling sites. Shell, for example,maintains six real-time operation sites around the world, in Houston; New Orleans, La.; Miri,Malaysia; Muscat, Oman; Port Harcourt, Nigeria; and Aberdeen, UK. These centers providearound-the-clock monitoring of the firm’s assets and operations.Considering some of the strides of the world’s most formidable retailers and consumer productscompanies in championing real-time operations (Wal-Mart and P&G, to name but a few)—andthe clear opportunities to improve critical areas of the business such as supply chain managementand inventory control, it seems particularly surprising that these sectors would rate so low on oursurvey.Sunil Verma, Chief Information Officer and Senior Vice President of US-based clothing retailerThe Children’s Place, is not surprised. “In terms of technology, the retail industry is a follower,”he says. “It’s an old industry which is changing quickly because of the internet. We quicklyunderstand that there is a competitive advantage in doing real-time analytics and the idea that weneed to infuse every customer touch point with some element of personalization. The minute youstart to internalize that, the need for real-time computing becomes more urgent.”According to Ranjay Gulati, the Jaime and Josefina Chua Tiampo Professor of BusinessAdministration at Harvard Business School, the reasons may center on a lack of understanding ofits benefits. “The retail sector is pretty fast-moving— looks at how quickly fads and fashionschange. Yet retailers don’t have that kind of cycle speed on being able to look at the data and runanalytics on it. Even worse, the people who have the data and the people who need the data arenot sitting in the same place—in fact; the people who need the information might not even knowit exists. So there is a huge disconnect.” 6
  7. 7. Still, the fact that so many respondents say their firms are planning to move in this directionpoints to an understanding that this disconnect exists—an encouraging step. “I don’t think we’reat the tipping point yet,” says Professor Gulati. “It’s still early stages. But we are approaching apoint where business competition is becoming so strong that it will eventually force companiesto find a way to get it done.”Drivers of adoptionFor “early implementers”—firms that have already adopted real-time business methods, use ispervasive across the enterprise: Upwards of 95% of these respondents have some form of real-time business applications in place across all departments. And the commitment to real-timebusiness is substantial. For each of the functions identified (consumer experience; supply-chainmanagement; production; sales and marketing; financial and business risk; and product designand innovation), over 50% of early implementers report a high or very high degree of adoption.Customer experience and supply chain management top the ranking (each at 64%), followedclosely by production processes (63%), and sales and marketing (62%).Naturally, obtaining a strategic benefit is a key driver for implementing real-time operations.Among early adopters, growth in market share (77%) is the most frequently cited driver foradoption. This is followed by building service and/or quality advantages (71%), expandingaddressable markets (68%) and building cost advantages (67%). 7
  8. 8. Meanwhile, “planners”—firms that are in the process of moving to real-time operations orintend to do so over the next five years—comprise some 46% of survey respondents. For theserespondents, building service and/or quality advantage is most frequently identified (87%) asan important or very important driver. This is followed by ambitions to grow market share(83%) and building cost advantages (82%). 8
  9. 9. Adoption strategies of planners are different from those of early implementers. By function,production process and financial and business risk lead initial investment among planners, withroughly 77% of those surveyed reporting fairly or very significant plans in these aspects of theiroperations. The next areas for investment are supply chain management, where 75% of plannershave fairly or very significant plans, and product design and innovation (74%).Among oil and gas businesses that plan to make their first efforts to move to real time, mostconsider building service/quality advantage and cost advantages as key drivers with 95%indicating this is quite or very important. Among new adopters in the retail sector, the greatestperceived benefit is building cost advantages (85%). In contrast, a higher proportion of newadopters in the consumer products sector (87%) cite increasing market share as the greatestanticipated benefits.Benefits and opportunitiesClearly, the purpose of real-time operations is to help companies make better decisions andspeed time to market. As such, it is of little surprise that early adopters cite increased revenue(31%) and cost reduction (35%) as the most important benefits. In this area, real-time operationsseem to offer a tangible benefit: On average, companies that have implemented such systems areseeing revenue gains of 21% and cost reductions of 19%. In fact, among early adopters, 77%report revenue gains. 9
  10. 10. For some companies, the benefits are even more substantial. Oil and gas firms report revenuegains of 36%, which are significantly greater than those in the consumer products sector(revenue gains of 14%). Business-to-business operations (27%), meanwhile, also reportsignificantly higher revenue gains than business-to-customer units (13%). The higher figuresmay be attributed to the fact that firms in these categories are further along the implementationcurve, suggesting a connection between real-time operations and revenue.At GE Energy, work is under way to embed real-time sensing software and controls intoequipment to run its plants. “There is a huge wave of activity that is taking place now arounddata collection,” says Dr. Peter Evans, head of strategy for GE Energy. “We are putting in placethe frameworks for better decision-making about how to run these plants at higher performance.”For example, GE is now sharing with its clients the real-time data it collects from its fleet of over1,000 gas turbines so customers can benchmark performance. “We collect and analyze the data,and turn n it into better solutions,” he says.For UK-based movie streaming service LOVEFiLM, which has more than 1.6 million membersand manages four million rentals monthly across Europe, the conversion rate of trial to paidsubscriptions is critical. “A 1% or 2% change in conversion has a huge effect on the bottomline,” says Mike Blakemore, the firm’s Chief Technical Office. “So we manage that constantly.”Real-time systems are key to that effort: In addition to crucial housekeeping, such as trackingcredit cards and payments in real time, LOVEFiLM can tailor its offerings to members based ontheir past preferences.Furthermore, because LOVEFiLM can track instantly which films are most popular among itscustomer segments, its content team knows which films it should promote to drive sales. “Thesooner we have the customer data,” says Mr. Blakemore, “the quicker we can adapt.” It’s astrategy that has paid off. In February 2008, Amazon became the firm’s largest shareholder; thisJanuary, the online retail giant announced that it would take full control of the LOVEFiLM.The benefits of real-time operations are not limited to companies whose business models aremainly digitally. Among airline executives, for example, the expression “wheels up” is a keyphrase: Once a plane takes off, the value of an empty seat is zero—it can never again generaterevenue. To keep financial performance aloft at Continental Airlines (now a unit of UnitedContinental Holdings), real-time analytics track empty seats prior to takeoff to put as manytravelers on board as possible. At any time, Continental executives can see an accurate accountof revenue-generating passengers on its flights.But real-time data helps with more than in-flight operations. It is now critical in helping theairline manage seat availability for its 2,000 daily flights, each of which can have as many as 30different seating classifications. In the past, says Anne Marie Reynolds, director of the airline’sdata warehouse, it was difficult to manage the complexity of the various pricing options—updates were done nightly in a batch process. “The science behind it hasn’t changed all thatdramatically but now we are able to respond within a few minutes rather than waiting until thenext morning.” That alone, she says, has meant millions in increased revenues. 10
  11. 11. Study respondents also expect to see substantial benefits from real-time operations in the future.For the sample as a whole, the expected gains in revenue average 28%; the expected costreductions average 20%; and the gains in productivity 24%.At The Children’s Place, says Mr. Verma, the company is undergoing a significanttransformation. “We have a new mandate to elevate what we do both in terms of product and interms of how we service the customer,” he says. Real-time operations are critical to thatmandate, particularly in terms of personalization. “Having access to immediate customer data isa huge opportunity,” he says. For example, in the past it took 30 days to know if a customer hadmade a purchase online or in a store. “Because a customer’s status couldn’t be updatedimmediately,” he explains, “they may have received suboptimal offers or incentives.”Despite being most pessimistic about future revenue gains and cost reductions, respondents inthe consumer products sector anticipate the biggest increase in these measures (when comparingbetween what has been realized in the past and what is expected for the future). This is likelybecause these firms have yet to adopt these systems in a fundamental way. In contrast,respondents in the oil and gas sector expect just the opposite—future revenue gains and costreductions are expected to be smaller than what was previously realized, though they will remainhigher than in other sectors. This further underscores the maturity of adoption in this industry.Assessing EffectivenessWhile early implementers agree that real-time business is effective, there is some variation in thedegree of effectiveness among functions. For example, early adopters rank its effectiveness morehighly in consumer experience (69%) and production processes (65%) than in financial andbusiness risk (58%) and sales and marketing (55%). 11
  12. 12. Real-time operations are seen as more effective by firms in the oil and gas sector than in theother industries surveyed, perhaps as a consequence of the high level of adoption within thesector. Apart from sales and marketing, a higher proportion of oil and gas respondents see real-time business to be fairly or highly effective than the sample as a whole.This gap is particularly marked for product design and innovation, and financial and businessrisks, underscoring the power of real-time information in complex, capital-intensiveorganizations.The weak ranking of sales and marketing likely reflects the commodity nature of much of theindustry’s output. Meanwhile, early implementers from the retail sector are more likely to see theinitiative as ineffective or of limited effectiveness than their colleagues in other industries.Future InvestmentsThe vast majority of early adopters plan further investment across their business functions. Forthem, consumer experience, supply-chain management, and product design and innovation ranktop as the business functions where further investment in real-time business will be directed,with over 55% reporting fairly or very significant plans in these aspects of their operationsFewer early implementers in oil and gas plan for further investment in sales and marketingfunctions; instead, they plan to extend real-time approaches in production processes, supply-chain management, and product design and innovation. Early adopters in retail, meanwhile, havemore modest plans for investment than other sectors. Nevertheless, for all business functionsexcept production processes, more than 40% of these firms plan for intensive expansion of theircapabilities.As noted earlier, those planning to roll out real-time systems have different strategies than earlyadopters. Rather than focusing on customer experience, they will focus on sales, marketing andsupply-chain management. These firms may be missing an important aspect of real-timebusiness, as early adopters rate consumer experience as the most effective area for real-timeoperations. Meanwhile, almost all early adopters plan to continue investing and expanding theirreal-time efforts, though they will move their spending from sales and marketing to productdesign and innovation, while continuing to focus on customers. 12
  13. 13. Barriers to operating in real timeDespite the understood value of real-time systems, significant challenges remain with regard toadoption. At the top of this list, respondents note the lack of availability or adoption oftechnology (59%), suppliers’ unfamiliarity with real-time systems (58%) and a general lack ofinternal knowledge or awareness of (58%) real-time operations.Professor Gulati is not surprised by these findings, and points to data silos as one culprit. Similarto the aforementioned barriers in retail, when information is not transparent or easily accessible,“you have a situation where the person who owns the data, even if they are able to analyze it andfigure it out, is not in a position to impact change.” The resulting opacity around what types ofdata exist and how they might best improve the business continues to hold companies back.“You need to either move the decisions to the people who have the information, or move theinformation to the people who are making the decisions,” he says.At The Children’s Place, the challenges have been both cultural and technical. When Mr. Vermaarrived at the company a year ago, there were at least six separate merchandising systems, eachwith its own set of data. Moving to a single system requires full-on collaboration andcommunication—setting standards for data, and creating processes for generating reports andsharing information. At the same time, asks Mr. Verma, “how do you get people to believe thatthe new way is better?” Like any organization undergoing transformation, “there’s cautionamong associates who have been here a while who have seen things work a certain way. Youhave to demonstrate there are alter natives that have better outcomes. It’s all about educating.”To that end, Mr. Verma has established a cross-functional IT innovation group at the company.“If you give your team the opportunity to work in different environments and understand howsystems within individual business areas work together to connect business processes, theybecome more open to change.”By sector, consumer products firms see a lack of available technology as a particular barrier, andgenerally judge the obstacles in the way of implementing real-time business to be greater thanother sectors. Surprising given their lead in using realtime business, oil and gas companies seesubstantial barriers around communicating the benefits of real-time operations. From a culturalperspective, “it takes time to roll these things out so that it becomes embedded in day-to-day 13
  14. 14. operations,” says Dr. Evans. “Given the high stakes that are involved in sustaining large energysystems, both trust and capability need to be built for wide-spread deployment into pipelines,networks and electrical distribution systems to be achieved.”At Continental Airlines, Ms. Reynolds says the greatest challenge for her group will beintegrating the airline’s real-time revenue management systems with its new parent company,United Airlines. Thankfully, “United reservations will be done on the Continental system oncethey’re fully integrated, so all we have to do is figure out how to back populate the Unitedhistory and incorporate the new United forecasting system into the revenue managementcalculations. It’s a lot of work—there are cases where United’s data doesn’t have all theattributes that Continental has, but we expect to be fully integrated by the middle of next year.”For the small group of respondents who do not plan to implement real-time operations, morethan half cite a lack of availability of technology, a lack of internal expertise and an inability tocommunicate the benefits as quite or very important. It’s possible that for these firms, real-timedata is not necessary for operations. However, it is more likely that these respondents have notyet fully considered how such systems could improve their business.For firms that are at the beginning of their efforts to move to real time, Professor Gulati saysthere are three steps executives must consider. The first is to make sure the information that isgathered across the firm is accurate. Professor Gulati refers to this as the informationarchitecture. This may seem obvious, but executives constantly underestimate the complexity ofcreating a single version of the truth.The second step is to understand the processes by which information is shared across the firm,and how departments need to use information to perform critical tasks. “This is theorganizational architecture,” Professor Gulati says. “Without clarity around who needs access towhat information, you’re working in a vacuum. No technology ever solved a business problemby itself.”The third step is to take a hard look at how decisions are actually made inside companies—whichcan be different from understanding the organizational processes. Professor Gulati calls this the 14
  15. 15. decision architecture. “Think in terms of the reporting, and who has what mandates.” This willrequire the IT team to work closely with executives across functions to ensure that the systemsare designed to be truly useful. “What happens in a lot of the initiatives that are CIO-driven isthat they work on the information, build a data warehouse and then add analytics. And then whatdo you find? Nobody uses it.Conclusion: Looking aheadAs our respondents confirm, use of real-time systems will only grow more pervasive in thecoming years. According to research firm Gartner, 30% of business analytics tools will use in-memory functions to add scale and computational speed by 2014. And 30% of businessintelligence applications will use predictive forecasting.As noted earlier in this report, executives expect to reap substantial revenues as a result of real-time adoption. Mr. Verma of The Children’s Place says that real-time systems will beinstrumental in helping the company move to enable mobile commerce—a move that he believeswill transform the retail industry. “In the near future customers aren’t going to come into a storeto find their size. It will be more a scenario where on a whim, a customer decides he or she needssomething. They will use their mobile device to figure out what they are going to buy and wherethey are going to buy it. Then purchase it wirelessly, and simply pick it up in the store.”At LOVEFiLM, plans are under way to use real-time data to improve the streaming quality offilms.“We have capacity to scale based on demand, and the only constraints are around thephysical infrastructure associated with people’s network connections in the home,” says Mr.Blakemore. “What we can do with real-time data is to look at the network down to customers’internet service providers, and see which are the best local points of delivery.” In other words,“If customers are having buffering issues or difficulty with playbacks because the network isoverloaded, we can switch to a different route to deliver content.” That, says Mr. Blakemore, willdramatically improve service—and help it gain market share.At GE Energy, one of the longer-term objectives for real-time data is “to allow consumers tohave more information and control of their energy use,” says Dr. Evans. While some of thatwork is already happening with GE’s business-to business customers, Dr. Evans envisions a daywhen a smart electricity grid provides transparent information to consumers so they can bettercontrol the electricity consumed by the various appliances in their home. 15
  16. 16. Real-Time Data AnalyticsIdentified and Developed Case Studies 16
  17. 17. Case Study 1: KTH – Royal Institute of TechnologyOverviewAnalyzing large volumes of streaming data in real time is leading to smarter, more efficient andenvironmentally friendly traffic in urban areasBusiness Need:Researchers at KTH, Sweden’s leading technical university, gather real-time traffic data from avariety of sources such as GPS from large numbers of vehicles, radar sensors on motorways,congestion charging, weather, etc. The integration and analysis of the data in order to bettermanage traffic is a difficult task.Solution:Collected data is now flowing into a unique software tool that analyzes large volumes ofstreaming, real-time data, both structured and unstructured. The data is then used to helpintelligently identify current conditions, and estimate how long it would take to travel from pointto point in the city, offer advice on various travel alternatives, such as routes, and eventually helpimprove traffic in a metropolitan area.Benefits:Uses diverse data, including GPS locations, weather conditions, speeds and flows from sensorson motorways, incidents and roadworks -Enters data into the InfoSphere Streams software,which can handle all types of data, both structured and unstructured -Handles, in real time, thelarge traffic and traffic-related data streams to enable researchers to quickly analyze currenttraffic conditions and develop historical databases for monitoring and more efficientmanagement of the systemReal Business Results  Uses diverse data, including GPS locations, weather conditions, speeds and flows from sensors on motorways, incidents and road works  Enters data into the InfoSphere Streams software, which can handle all types of data, both structured and unstructured.  Handles, in real time, the large traffic and traffic-related data streams to enable researchers to quickly analyze current traffic conditions and develop historical databases for monitoring and more efficient management of the system 17
  18. 18. Case Study 2: Pepperl+FuchsOverview:After evaluating different alternatives, Pepperl+Fuchs decided to implement the log-basedtechnology IBM® InfoSphere™ Change Data Capture (previously DataMirror) for datareplication in the heterogeneous database environment. This means that data changes withinheterogeneous data storage systems can be monitored in real time and only the changed datatransferred.Business need:Following the introduction of a central Enterprise Resource Planning (ERP) system, data indifferent worldwide locations had to be integrated and synchronized so that production processesand reporting on site can continue to be optimally supported.Solution:Pepperl+Fuchs implemented IBM InfoSphere Change Date Capture (previously DataMirror) forreal-time data synchronization between different database platforms. The solution providesminimal load on the central ERP system, high performance and reliable data integrity.Benefits:Error tolerance and stability reduce administrator costs. Efficient implementation by a smallteam within just a few days helped ensure low implementation costs and transparent trainingefforts. High data transfer speed via existing landlines helps ensure low communication costs.Operational data stores at the production locations in the United States and Singapore receivecurrent information at all times. The load of the central ERP system is significantly reduced;expensive batch runs are not necessary; and operational systems are not exposed to risk. 18
  19. 19. Case Study 3: DucatiOverviewAround the world motorcycle manufacturers are rolling out new models to delight bikers anddefine success on and off the track. Ducati is responsible for the construction, importation anddistribution of motorcycles.Business need:Ducati had been employing expensive manual methods to track its parts via a system of paper-based spreadsheets. Human error occurred from time to time and delayed production.Solution:A service oriented architecture-based RFID solution alerts management if an incorrect part isbeing incorporated into the manufacturing of a motorcycle so that the company can rectify thesituation before the vehicle is manufactured, sold and driven.Benefits:Projected 50% savings on revision campaigns and maintenance and monitoring of productionline ~ 30% decrease in time for scheduling and verification of products ~20% savings inreclaimed scrap and missing partsTracking parts automaticallyLeveraging a strong preexisting business relationship, Ducati decided to design and implementthe RFID tag solution, which would enable real-time, automatic identification of each vehicleand vehicle component along the entire production chain.“This solution is the backbone of our development,” says the coordinator of informatics systems.“We export the data coming from the RFID devices into a Java™ program in a web architecture.We have a message system which alerts management to issues with components.”Reducing errors and improving production efficiencyThe manufacturer deployed a total of three IBM® System x® 336 servers running the SUSELinux® V10 operating system to support IBM WebSphere® software and an IBM DB2® 9 dataserver. Employees use wireless devices to scan RFID-tagged components and VINs into thesystem. Approximately 100 users leverage the DB2 9 data server, 60 of whom can access itconcurrently. A custom-developed Java Platform, Enterprise Edition (Java EE) application—powered by IBM WebSphere Application Server Network Deployment V6 software—willenable employees to track inventory throughout the production life cycle.The manufacturer anticipates that by replacing its legacy component-tracking system with anRFID solution, it will reduce errors and improve production efficiency. For the company, fewer 19
  20. 20. errors translates into a higher-quality finished product, which will help improve customersatisfaction and could potentially expand the company’s already large share of one Europeanmotorcycle market. The new solution will also help improve the company’s ability to providesuppliers with accurate inventory data. Using real-time data collected from the WebSphere-basedinventory-tracking application, employees can now tell suppliers if a certain product is in stock,and quickly order new materials when inventory levels are low.Up to 50 percent cost savingsIf a motorcycle is destined to be delivered to France and the factory installs a speedometer thatdisplays miles instead of kilometers, the solution alerts management in real time. “That way wereduce the amount of errors directly in the assembly line,” says the coordinator of informaticssystems. “Another important innovation is replenishment of the components from the line to thewarehouse.” We also have fewer employees verifying parts manually, which saves costs.Altogether, the cost of monitoring, maintenance, replenishment is 50 percent less.“With the RFID solution we have seen a 30 percent decrease in the time for scheduling andverification of production, so products reach the market 30 percent faster. In addition, we’veseen a 20 percent decrease in the cost of missing parts and scrap. We’ve had improvements inother indirect costs because of the efficiency of the production line. All of these have shownimprovements in quality and service.”Preventing revision campaignsAs yet, Ducati hasn’t had a revision campaign in which it has had to test its new technology―asign that prevention is happening in the factory. “By avoiding the cost of one revision campaignwe have paid for this entire RFID implementation,” says the coordinator of informatics systems.“So we are going to continue to invest in this technology. In the future, we would like to exportthis SOA technology, this business project management, to our outside suppliers, so that oursuppliers can have this information at the same time we do. We also want to know what oursuppliers are doing at a particular moment for us, and therefore we will be able to consider ifsomething will be just in time or if it’s delayed. We’re also eager to change over to the smartfactory with smart supervision. And we will be doing this all in real time using Web 2.0technology and the latest innovations in SOA and business project management, thanks to IBM’shelp.”The improvements to the assembly line and inventory system have brought about efficienciesthat were unthinkable just three to four years ago, and the coordinator of informatics systemshopes there will be more benefits to enjoy three or four years from nowSmarter car manufacturingDucati was struggling with a parts tracking system that was based on paper and prone to error.When a wrong part found its way onto a motorcycle on the assembly line, it was expensive tocorrect the mistake. Now the company is working smarter and benefitting from savings in timeand money with the automatic monitoring and alerting performed by its RFID parts trackingsolution. 20
  21. 21. Case Study 4: Heathrow AirportOverview:In recent years, Heathrow faced a challenge not only keeping up with existing traffic but bracingfor major increases in traffic. Due to political and geographic issues, the airport had to livewithin its existing footprint. In short, it needed to be a model of efficiency.By harnessing the analytic capabilities of a BPM system, Heathrow managers have been able totune ground crews to incoming traffic by linking to the Euro Air Space system to give crews afull 30 minutes warning before a plane arrives. A business activity monitoring (BAM) capabilityis also used to manage crews, gate assignments, docking and undocking, refueling, cleaning, andmore.In addition to providing a visual and intuitive indicator of where planes are in the process,numerous scenarios can be invoked for different levels of traffic and delays. For instance, ifsignificant delays are expected for departing passengers, security systems are adjusted to keeppeople in retail areas rather than gate areas for a longer period. This has resulted in happierpassengers and security workers and also had the serendipitous result of substantially raisingretail revenues.“Today’s BPM handles expected exceptions but with newer tools you can start looking forunexpected exceptions,” says Jim Sinur of Gartner Research. For instance, he indicated, you canstart to look for signs of inflation or deflation. You can look at complex event patterns andcertain recurring events and combine that with modeling.“That is the top level—where someone managing the process gets notification either fromcomplex events or modeling tools that there is a good chance something unusual is happening,”says Sinur.Seek, model, adaptSinur says Gartner talks about using BPM in conjunction with a “Seek, Model, and Adapt”approach. To date, BPM has tended to focus on the adapting part of the equation and may havemissed the “seeking” part, often because information exists across multiple processes.He says the most successful users of BPM real time analytics should consider taking differentapproaches, depending on whether they are a planning-oriented or reactive company anddepending on the specific analytic capabilities their system provides. Plan-oriented companiesoften prepare for different scenarios and compose alternative responses for each—so when BPMprovides an early warning, they can leverage that information for quick action. Reactivecompanies, on the other hand, may be able to leverage elements of real time analysis to “learn bydoing”—inventing new ways to operate quickly with the help of BPM analytics. “Some BPMsystems can provide modeling tools and automated process discovery to look for patterns,” saysSinur. “This stuff is hot and will be for the next decade—companies are just starting to discoverthe potential this represents,” he adds. 21
  22. 22. Case Study 5: KEMCOOverview:Korea Energy Management Corporation (KEMCO) is an organization that contributes thenational economy and develops and supplies future energy. KEMCO has created and providedthe data lookup screens by OLAP dedicated tool and key business system, every time line-of-business users request. Meanwhile, as green growth has become the keyword to stand for thesetimes and the related laws have been embodied, the organization felt the need of having the newclass of business efficiency. KEMCO selected the Voluntary Agreement (VA)-related businesssystem started from 1998 as the first target for innovation and then they got the development ofBI system going in the end of 2009. It has driven the BI project to maximize the ease-of-usethrough the Microsofts solution instead of the existing OLAP tools. Through this, having noticedthe business efficiency of the user environment where they can draw the statistics they want onExcel without moving to lookup screen each time, KEMCO has a plan to extend the user-oriented BI into their other system areas.SituationThe statistical and analytical business is of the highest importance to Korea EntergyManagement Corporation (KEMCO) leading the charge as it tries to implement the green growthand low carbon economy. Since it is the organization managing the demands for energy, thestatistics and analysis is the key to its all businesses.KEMCO has made a number of statistical and analytical data viewed in a data lookup page bykey business system. Prior to creating reports, they had to log in the data lookup site of therelated business system and extract the data. And if there is any hard-to-acquire information fromthe view, they had requested it to Information Statistics Division and KEMCO via e-mails orphone calls. After collecting data, they arranged and analyzed it in Excel to create reportsordinarily.The need of new class of efficiency has not been on the rise in such a business process until thegreen emerged as the keyword passing through this time. In a major move towards green growtheverywhere including economy and society, the government declared the low carbon greengrowth as a national vision in 2008.To support this institutionally, the preparation for the Framework Act on Low Carbon, GreenGrowth (hereinafter, Green Act) is under discussion commonly. Naturally, as the KEMCOsstature and role has grown, for statistics and analysis business as well, they need to prepare thefuture changes with the stream of times.VA (Voluntary Agreement)-related business was picked up as the first target for innovation. VAis the non-compulsory energy saving measurement where a corporation sets itself the objectiveof energy saving or greenhouse gas emission reduction for five years and makes an agreementwith government, and then the government helps the company fulfill this agreement to make theobjectives achieved with the collaborative effort of two parties. 22
  23. 23. KEMCO has collected and analyzed a number of information for evaluating and monitoring VArelated corporations fulfillment. As the demand for this statistical and analytical information hasincreased in an era of Green Act, the VA related business is chosen first as a target forinnovation.Yeong-sang Yoon, Leader of Information Statistics Division, KEMCO, stated, "Analytic worksrelated to VA has increased. It is because of increased corporations participating to VA, andGreen Act is the key background as well. KEMCO manages the industrial development segment,one of targets to be managed by Ministry of Environment, with regard to the enforcement ofGreen Act."Actually, in addition to its existing job of establishing the energy saving goal and thenmonitoring and evaluating its enforcement, its statistical and analytical activity will be extendedto the measurement, reporting and verification of greenhouse gas emission for the emissiontrading scheme between corporations regarding the industrial development initiatives.SolutionKEMCO selected the new approach of BI (Business Intelligence) for innovating statistical andanalytical business related to VA system. Making a decision that it was no more efficient todevelop the screen each time needed, KEMCO viewed BI the best way, since the kind and thenumber of materials required to analyze increased.KEMCO grappled with the problem of finding technical methodology for implementing BI.It thought adding BI to VA would only have a marginal effect in the light of its own experience,unless it could overcome the limitations of OLAP dedicated tool. Yoon-geun Han, Director ofRealWeb which was responsible for this BI project, explained, "KEMCO has ever introduced thePowerPlay of Cognos and OLAP tools developed by a domestic development firm. As it difficultto use these tools without expertise, its usage was not too high. Thus these proprietary tools werenot appropriate for satisfying the needs of users to extract statistics they want if needed."The system developed during August to December 2008 was configured to make the dataexported from VA systems operated on SQL Server 2005 and distributed and shared through theMicrosoft Office SharePoint Server 2007s Excel Services. The existing lookup screens wereremoved except the page used to access by its staff outside the company. Besides, About 10 datamarts were created to reside on SQL Server 2008.Finishing BI development for VA system successfully, KEMCO is considering extending BI intoits other systems step by step. Mr. Yoon stated, "We have noticed that BI was very useful fromthe case of VA system. Now we are in discussions with Microsoft about applying BI to ClimateChange related business and have a plan of approaching to all sorts of statistical and analyticalbusiness which we do from a enterprise-wide BI perspective in the mid- and long-term." 23
  24. 24. BenefitsComplete the data collection that used to take 3 days in real timeStaffs dealing with data analysis related to VA are feeling that all the before and after of BIimplementation are certainly different. This is because they spend less time to collect materials,thus they have to only focus on writing reports.Mr. Yoon said, "If we failed to get information we want in the lookup screen, we have to requestthe materials to the Information Statistics Division and it took at least 2-3 days to receive it. Nowwe dont have to spend our time to wait materials. We can bring the data on data marts in systeminto Excel."Respond immediately to the request of a variety of statistical reportsAnalysts dont have to drill down to 10~20 lookup pages or request the materials they need to theinformation statistics division, which increase their responsiveness to businesses.Mr. Yoon said, "Unlike general enterprises, a public institution has to consider the statistics inbuilding business systems. It has to submit the data to superior authorities if they request. Afterthe SQL Server based BI system was implemented, we dont have to look up the law data case bycase, freeing us to focus on our own tasks including writing graphs, which makes ourresponsiveness to statistical and analytical business increased."Reduce cost about 60% compared to that of OLAP dedicated toolKEMCO broke stereotypes that BI is the expensive tool from this project. It was not a challengefollowed by big investment once the organization actually tried to do it. They found that theirwill and practice were important.Mr. Yoon explained, "While Introducing OLAP tool will need at least 100 million won to beinvested, we spent less than 30 million won in this BI system implementation."KEMCO expects that these cost reduction effects will really come into its own in the case ofapplying BI into the system that many people use in the future. Unlike the VA system that 50analysts use, in case of the system that has to support more users, cost benefits of Microsofts BIsolution is expected to be much higher compared to that of the traditional OLAP tool.Ensure a smooth transitionOne of results of moving from lookup screen to BI is the smooth transition of VA relatedbusiness. Mr. Yoon said, "A person in charge is changed regularly in our organization, so therewas a difficulty in transition for the successor in analytical business. Analysts often need theirpersonalized environment due to the nature of statistical and analytical tasks. Thus in case ofdelivering information based on lookup screen in the past, the new analyst often requested a newscreen to be developed. Now using Excel ensures a smooth transition 24
  25. 25. Case Study 6: Faro HospitalOverview:Faro Hospital provides medical services for citizens in the Algarve region of Portugal. Theorganisation wanted to improve decision making and increase efficiency. It deployed a businessintelligence solution based on Microsoft SQL Server 2008 R2 data management software.Funding and resourcing strategies are now based on accurate, real-time data, which has helpedthe hospital reduce costs and improve patient care.Business NeedsFounded in 1979, Faro Hospital provides medical services for more than 250,000 people in theAlgarve. The hospital complex covers 46,500 square meters and has 530 beds. In December1999, it became a state-managed organisation, with specific government targets. In addition toproviding high-quality medical care to patients, the hospital managers needed to streamlineprocesses and minimize costs.Until recently, each department used its own applications to manage patient records andemployee schedules. Francisco Serra, Head of IT Strategy for Faro Hospital, says: “It wasslow—often impossible—to share information across departments quickly.” For example, anytime nurses needed outpatient treatment information they had to request a manual transfer of therecords from the outpatient clinic. This slowed patient treatments, and prevented the hospitalfrom providing the quality of care to which it aspired.”This lack of integration also delayed decision making. “It took a long time to consolidatesufficient information to achieve an overview of the entire hospital,” says Serra. “Managers hadno day-to-day view of the organization’s activities, so it was difficult for them to make crucialdecisions quickly. For example, if a problem occurred, such as a need to arrange cover for asurgeon or specialist, it was impossible to make a fast decision about how to resolve it.”Collecting the data was also a lot of work for employees. “Staff were spending too much timecollecting, organising, and re-entering information, instead of doing their jobs. For example, if anemployee needed information from another department, such as medical records, it would haveto be requested and manually delivered. We needed to increase employee efficiency and startgenerating a better return on investment from our workforce,” says Serra.SolutionIn January 2009, Faro Hospital engaged with Microsoft Services to create a hospital managementsolution based on Microsoft SQL Server 2008 data management software, Microsoft OfficeSharePoint Server 2007, and Microsoft Office PerformancePoint Server 2007 businessintelligence software. The solution consolidates data, such as patient records, appointments, andstaff schedules, into a centralized repository. 25
  26. 26. Office SharePoint Server 2007 acts as the hospital intranet where employees can update andview patient data. If a patient is transferred to a different department or specialist, thisinformation is always captured in Office SharePoint Server 2007, ensuring that treatment historycan be accessed by doctors and nurses in different departments.BenefitsSince deploying SQL Server 2008, Faro Hospital can make decisions based on real-time data.This helps the hospital to manage resources effectively, minimize costs and reduce employeeworkloads. As a result, patient and employee satisfaction has increased.  Managers make more informed decisions. Faro Hospital creates reports with up-to- date information about every part of the hospital, and can make intelligent decisions based on that data.  Costs are reduced. “With SQL Server 2008 R2, we can identify areas where we are failing to get a return on investment,” says Serra. “We can then reallocate these resources to make sure everything is as cost-effective as possible.”  Quality of care has improved. Patient appointments are easier to manage than before. Serra says: “We can optimize the number of patient appointments in a day, because we can forecast them more accurately. By looking at historical data we can get a good idea of how many employees are needed to meet patient demands, and plan for cover if any personnel are absent on leave or as a result of sickness.”  Patient care has improved. Because different departments can collaborate more efficiently, patients get a better service. Serra says: “All departments have access to accurate data, and can share information, such as medical records, with ease. As a result, patients receive a consistently good service across all departments, getting care very quickly.”  Employee satisfaction has increased. “Employees have noticed a distinct improvement in services, and are much happier as a result. There’s less pressure on them because planning has improved, and they have more time available to do their jobs,” says Serra.  Improved shipping efficiency—and thus profitability—by eliminating silos across business lines and enabling users to model and analyze shipments to determine where they can consolidate or add multiple stops to improve service or reduce costs  Exceeded expectations for productivity improvements and cost savings at the eight sites deployed to date, already saving more than US$2 million  Provided more predictable processes in outbound shipping, enabling planners to more quickly and easily manage exceptions, which now happens much less frequently  Freed planners from day-to-day outbound shipping monitoring, instead enabling them to work on continuous improvement by reaching out to customers to see how they can better meet their needs 26
  27. 27.  Benefited from integrations between Oracle E-Business Suite, as well as third-party applications, which is helping to improve efficiency and reduce administrative effort 27
  28. 28. Predictive Data AnalyticsIdentified and Developed Case Studies 28
  29. 29. Case Study 1: Ingersoll RandOverview:Ingersoll Rand is a global, diversified industrial firm providing products, services, and solutionsthat enhance the quality of air and comfort in homes, buildings, and transport and that protectfood and perishables, secure homes and commercial properties, and increase industrialproductivity and efficiency.Given the breadth of its product offerings and its global reach, the company has a complexnetwork of manufacturing and distribution operations. To increase efficiency while ensuring thatit continues to deliver the superior service its customers expect, Ingersoll Rand implementedOracle Transportation Management to coordinate its outbound shipping across various lines ofbusiness and facilities.Even while the application was implemented currently at eight sites of the company’s more than70 locations, it already exceeded expectations for improved productivity and cost savings. WithOracle Transportation Management, users can view shipping plans across lines of business andfacilities, enabling them to adjust shipments to increase efficiency, such as making multiple stopswithin one shipment. The application has already contributed to more than 2 million dollars insavings and increased on-time delivery rates.Challenges  Provide superior customer service by ensuring quality, cost-effective, on-time delivery of products across multiple business lines, ranging from Trane heating and cooling systems to Schlage locks  Manage high shipping volumes, amounting to several thousand lines per day, from several facilities, including discrete manufacturing centers and warehouses  Meet tight delivery windows for industrial products, such as compressed air systems, which must arrive when an installation crew is on site to deploy it to avoid undue customer costs and project delaysSolutions  Deployed Oracle Transportation Management to replace legacy transportation management systems and provide a streamlined and integrated outbound shipping process―leveraging truckload, less-than-truck-load, and parcel modes―across four diverse business sectors  Rolled out the application as part of an Oracle E-Business Suite implementation at eight sites with plans to roll out to all Ingersoll Rand manufacturing and distribution locations  Increased the percentage of on-time deliveries―with on-time deliveries now averaging above 95% 29
  30. 30. Case Study 2: EMC Insurance CompaniesOverview:EMC Insurance Companies struggled with pinpointing the right amount of money to hold inreserve against potential case payouts; holding back either too much or too little could bedisadvantageous to the firm’s bottom line. After a year in which the company experienced a run-up in reserves, EMC took steps to improve financial reserve management. The company selectedPolyVista, advanced data analytics software built on Microsoft SQL Server Analysis Services, touncover anomalies, correlations, relationships, and patterns hidden within the firm’s warehouseof claim data. After deploying the solution, EMC was able to improve financial reservemanagement, identify claims requiring special attention, improve data quality, support executivedecision making with improved analysis, and better manage expenses.SituationEMC Insurance Companies sell workers’ compensation, property, and casualty insurancethrough independent insurance agents in 16 branch offices across the United States. Founded in1911, EMC is one of the largest insurance companies in its home state of Iowa. With assets ofapproximately U.S.$3 billion, the insurance firm has the bulk of its volume in commercial linesof business, with the remaining 15 percent in personal lines.The Challenge of Financial Reserve ManagementEMC manages its loss reserves—money set aside for paying future insurance claims—by settingup a reserve for every case that the company handles. “When a claim walks in the door, we try todetermine what is the most likely outcome for that claim,” explains Rich Schulz, Senior VicePresident of Claims for EMC Insurance Companies. Strategic management of financial reservescan be a key contributor to the success of any insurance firm, and they are important indicatorsthat are looked at by both insurance regulators and investors.“It’s a real balancing act; you don’t want to underreserve or overreserve accounts. Consistency iscritical,” says Schulz. “In 2004, we saw that our loss reserves were starting to creep up and yet,with all the tools we had, we were unable to pinpoint exactly what was happening. We realizedwe had to get a better handle on the way we were setting our reserves, and I was brought in toEMC to improve our management of our financial reserves.”Schulz wanted to be able to predict the company’s claim outcomes on an ongoing basis and,hopefully, identify and closely monitor those cases likely to have negative outcomes such asincreased costs, additional services, or a lengthy resolution. “We had all this claim data and notan effective way to look at it,” says Schulz. “We wanted a tool that would allow us to look attrends that the company was experiencing and be able to figure out what was driving them.” Helooked at a number of predictive modeling tools available on the market, but found that initialinvestment costs were high and the cost of maintaining those tools was far beyond what thecompany was willing to pay. 30
  31. 31. The Need to Identify TrendsEMC stored a great deal of data in its internal claim processing system, but managers needed theability to sift through that data more efficiently. For example, executives at EMC wanted to trackand understand company trends related to claim frequency, which is the ratio of claim volume topremium volume, and severity, which is the ratio of claim cost to premium volume. Not only didthey need to track these measures in order to manage their business, but executives also wereexpected to provide explanations of sudden trend changes to industry analysts.“Right before a first quarter earnings call, I was asked about a change in the severity andfrequency of workers’ compensation claims,” says Schulz. “At the time, I had to call individualbranches to inquire as to what was going on.” EMC had no good system of reporting that couldbe quickly used to discover the reasons behind unexpected changes.Schulz also needed to be able to track expenses both across the company and by individualbranch, and he wanted the flexibility to analyze expenses in consideration with other factors.“Currently, most claim people are under a lot of stress about expenses,” Schulz says. “The abilityto understand where expenses were running high is important. All of our adjustment expensesare open to analysis. [For example,] do we have one or two business units that are overusingcertain types of vendors to investigate claims?”SolutionIn need of an affordable and effective solution, EMC Insurance Companies heard aboutPolyVista business intelligence software, which provides two important capabilities: A businessuser can use it to drill up and down on data, and it can automatically identify anomalies,correlations, relationships, and patterns that are hidden in that data—all without requiring customcode or predefining the anticipated results.EMC was interested in determining whether the PolyVista software, which uses Microsoft SQLServer Analysis Services as its core engine, could help the firm with its data analysis needs.Analysis Services has become the dominant multidimensional engine in the market. It provides arobust open development environment that allows us to build and enhance our uniquesophisticated analytic product at an affordable price point.EMC decided to move forward with implementation of the business intelligence solution.Because the company had not previously used SQL Server 2005 Analysis Services, key ITpersonnel at EMC completed training.After EMC completed the training, PolyVista sent a senior consultant to EMC to train key ITstaff members and business users on the business intelligence software. During the fall of 2007,the complete implementation of the business intelligence solution at EMC was finished in lessthan eight weeks.As of February 2010, there are 24 PolyVista users at EMC—8 powers users and 16 casual users.EMC has relied on the power users to complete complex analysis and to assist other users withthe software. 31
  32. 32. BenefitsFrom the beginning, EMC Insurance Companies anticipated many uses for the businessintelligence solution. The initial goal was to improve reserve account management; however, theinsurance firm is also building predictive modeling into the claim system and has improved thequality of data. Finally, EMC is improving decision making and controlling expenses.Effective Management of Financial ReservesEMC relies on PolyVista as its primary tool for managing case reserves. Since theimplementation of the business intelligence solution, EMC has managed its financial reservesmore accurately and the company has avoided the volatility that can occur when reserves are notconsistently and accurately established.The software has helped the firm ensure that all branches are in compliance with the company’sreserve guidelines.”Predictive Modeling System for Improved Claim OutcomesEMC is using PolyVista to add predictive capabilities to the automated claim system. Now, thecompany can identify casualty and workers’ compensation claims that are likely to have anegative outcome and assign these claims to the company’s most experienced people to helpmitigate potential risks. If a claim is not properly identified or assigned, it can take longer toresolve and result in additional costs and services. By recognizing the complexity and riskfactors early on, an experienced claim manager can expedite the claim with better results for boththe claimant and EMC.Better Data QualityEnhanced reliability and accuracy of data with the business intelligence solution was animportant realization for EMC.In the highly regulated insurance industry, data quality is an important aspect of compliance. Forexample, Schulz says, “Under the federal government’s Medicare Secondary Payer laws, wehave to have the age of the claimant or insured in our system. We now can run reports showingthe portion of claims that don’t have an age or have an age that doesn’t match other data ordoesn’t make sense.” These reports allow EMC to identify and fix errors before submitting datato the Medicare program.Improved Executive Analysis to Support Decision MakingDuring periods of rapid change, EMC relies on its business intelligence solution to helpexecutives turn data into knowledge and make more informed decisions. “When you haveweather events or the economy is changing in a specific area, you need to understand how it isaffecting the business,” says Schulz.Effective Control of ExpensesBy using the business intelligence solution, EMC is able to track expenses and control costs thatare tied to specific buys, and match data by business unit. 32
  33. 33. Case Study 3: National Science FoundationOverview:The NSF is helping to predict earthquakes across the world in real time. The system has beendeveloped by a consortium of universities and government agencies, with funding from NSF(National Science Foundation), to provide an infrastructure of networked tools for research inocean science – constructing an internet-based system to collect and share data.Real-time Seismic Monitoring ProgramThis is a large system with 16,000 land and sea-based sensors, each sending several channels ofseismic event data at a rate of 40 data points per second. The application executes in an AmazonEC2 Cloud on a cluster of SQLstream servers, connected by an AMQP message bus.SQLstream’s AMQP adapter (built with RabbitMQ) enables the streaming SQL application toview the AMQP bus as a domain of input and output data streams. The initial prototype wasupgraded to a full-scale system running on a cluster of servers without any changes to thestreaming SQL application.SQLstream’s contribution to the project, an application that processes seismographic data in real-time, demonstrates: 1. An operational deployment of streaming SQL for scientific calculations in real-time 2. Real-time distributed processing in an Amazon EC2 cloud using SQLstream and AMQP 3. Rapid development and rollout of real-time data applications using standards-based streaming SQL. 33
  34. 34. Seismic Monitoring:The sensor network contains about 16,000 seismic sensors, organized into grids, covering largeparts of the North American continent and the adjacent oceans (see illustration below). Eachsensor measures the motion of the ground under it in three dimensions, and transmits its data asseveral digitized channels, typically sampled 40 times a second.While the rate of each signal channel is modest (since seismic waves are low-frequency),this adds up to a large amount of data to process in real time. Moreover, the rules for detecting aseismic event are heuristics that apply to a time interval of several minutes: so the application hasto calculate some quantities from the raw data and to store these calculated values over a timewindow.But to detect an earthquake reliably, it’s better to monitor all the sensors at once, looking for adisturbance in the signals that first appear in one place, and then appear in nearby places: adisturbance signal that propagates and changes shape in a way consistent with the physics of aseismic wave.Real-time sensor network management in an EC2 CloudMonitoring tens of thousands of signal channels arriving at 40 sample points per second is acomplicated problem, but it can be made simpler by breaking it into stages. We’re interested inearthquakes, which are infrequent, so SQLstream first reduces the amount of data by scanningeach channel for patterns that suggest the beginning, the peak or the end of a quake: in otherwords reduce the dense signal to a sequence of interesting events. Then we can look for eventsdetected on other channels that could be due to the same quake propagating in physical space andtime.In the first phase, we built a real-time seismic event detector in streaming SQL.We translated a scientific algorithm into streaming SQL, and connected to the scientific sensordata infrastructure – using our AMQP adapter. This involved less than 100 lines of streamingSQL.In the second phase, the prototype was scaled up to a full-size system, dealing with 16,000sensor channels, running on multiple SQLstream server nodes created automatically in anAmazon Elastic Cloud. This expansion required no changes to the streaming SQL applicationdeveloped for the initial prototype – simply running the same streaming SQL application insidean elastic container/manager. 34
  35. 35. Real-time Event Detection with Streaming SQL:The sensor data processing pipeline has five functional stages 1. Reading Messages – over the AMQP adapter. A configuration parameter specifies which “topics” SQLstream subscribes to, that is, which set of sensor channels it receives. 2. Unpack Data – a user defined transform (UDX) unpacks data channel messages into individual data points. 3. Signal Processing – extract higher order information from the raw data to identify seismic events given background noise and non-seismic ‘bumps’. An example function would be to calculate the ratios of multiple rolling averages of the signal value (x) over different time windows. 4. Event Detection – The next stage applies heuristic rules to the processed data streams – the output is much sparser: a stream of significant events, each indicating a possible start/peak/stop of a seismic wave on a particular channel. If events of the correct type occur within a correct interval of each other (as shown in the Signal Plot Diagram), they are accepted as significant. 5. Writing Messages – output (publish) significant events over the AMQP adapter. 35
  36. 36. Case Study 4: Laboratorios IndasOverview:Laboratorios Indas is the largest manufacturer of sanitary and hygiene products in Spain.Headquartered in Madrid, the company distributes its products through retail channels such aspharmacies, hypermarkets and supermarkets, as well as hospitals throughout the country.Business need:The company’s existing method of collating sales data was neither reliable nor user-friendly, andthe inability of sales teams to share client information prevented analysis of customer behavior,which limited potential sales revenue.Solution:Laboratorios Indas has implemented a business intelligence solution that provides preciseintelligence on where and how sales efforts should be focused in order to maximize profitability.The company now integrates and analyzes large quantities of sales data captured and uploadedvia sales reps’ PDAs, together with demographic data to build sophisticated statistical modelsthat provide deep insights into likely sales patterns, at the individual product level, forpharmacies in different locations.Results:Increased average sales per pharmacy by 37 percent over the prior year. Reduced the averagetime it takes sales reps to compile sales data from store visits by 92 percent, from five minutes to25 seconds.Benefits:Began generating commercial benefits just four months after implementation by providinginsights into the likely product demand patterns for individual pharmacies. This enabled thecompany to target key pharmacies more systematically with visits, emails and mailings to helpdrive sales.InstrumentedCaptures all sales data and order history, plus client and customer demographics, in a central datawarehouse for data mining.InterconnectedThis technology integrates sales information from PDAs in the field with population size,socioeconomic levels and number of pharmacies in the region, providing real-time predictivemarketing support. 36
  37. 37. IntelligentIncreases sales per pharmacy by 37 percent by using statistical analysis to leverage real-timeorder history and local demand patterns. It also reduces effort for sales teams while optimizingsales visits, strengthening customer loyalty and expanding the business. 37
  • ssuser7833951

    Jun. 10, 2015
  • YuHan6

    Mar. 22, 2015
  • kishorejets

    Jun. 23, 2013

The goal of this paper is to explore executive perceptions and opinions about real time data applications and operations. We have interviewed over forty (40) Key Innovation Leaders who have been cited as the most innovative thinkers within the world of analytics and have documented distinctive case studies which have clearly optimized business intelligence.


Total views


On Slideshare


From embeds


Number of embeds