Case Studies Utilizing Real-Time Data Analytics

2

Project Overview:

The goal of this report 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.

With every passing day, the pace of business is accelerating. Paradigm shifts in the global

economy combined with rapid technology adoption are forcing business executives to contend

with a new global marketplace fraught with uncertainty and constant change. To be successful

under these new market dynamics, firms must adjust operational processes, corporate strategies

and business models at lightning speed—allowing them to leverage intelligence instantly and

take immediate action. At the same time, they must make sure their decisions are informed by

proper data and analysis.

The interviews and research conducted focused on large and medium-sized business units in four

sectors 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.

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 speed

We live in a world that increasingly communicates and operates in real time. Over the past

decade, the spread of social media and mobile phone use, combined with the explosion of

accessible information and increasing interconnectedness of global markets and cultures, has

dramatically sped the pace of our daily lives. The evidence is clear in the dissemination of

international news: For example, when US Airways Flight 1549 emergency-landed on New

York’s Hudson River after striking a flock of geese during takeoff in January 2009, it took only

three minutes for the first “tweet” to alert the world

4

The pace of business, too, has dramatically accelerated. Customers and clients expect their

queries and problems to be addressed immediately. In an uncertain global market where there are

many more competitors and constant change, executives know that success depends on the

ability to take fast action and make rapid decisions. Indeed, in the business world today, speed is

not 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 have

already implemented real-time business applications in some way (though 65% of those who

have not yet implemented real-time systems expect to do so within five years). And the

experience is far from uniform. For example, companies that focus on business-to-business

operations (35%), are more likely to have moved to real-time than those in the business-to

consumer space (16%). And almost half of the very large businesses studied— those with annual

revenues in excess of US$25 billion—have introduced real-time business, significantly more

than those with a turnover of less than US$1 billion (22%) or between US$1 billion and US$10

billion (28%).

Adopting a real-time business approach can provide companies with myriad benefits at

both 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.

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 to

conduct a range of business activities instantaneously. The activities that make up realtime

business can, therefore, incorporate all aspects of business, including gathering and acting on

business intelligence, developing promotional and marketing tactics, controlling and adjusting

production processes, managing inventory, identifying and managing business risks, closing and

fulfilling sales and meeting customer needs.

Real-time operations depend on in-memory analytics, which takes a different approach than

traditional business intelligence systems. Rather than storing information on various external

disks and caching bits of data in a computer’s random access memory (RAM), real-time

intelligence puts the data directly into RAM. Advancements in computing power and storage

have made it possible to store vastly more amounts of data in RAM than ever before, and allows

for extremely fast query responses

Which companies are ahead of the curve in adopting real-time operations? The answers are

surprising. 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.

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 provide

around-the-clock monitoring of the firm’s assets and operations.

Considering some of the strides of the world’s most formidable retailers and consumer products

companies in championing real-time operations (Wal-Mart and P&G, to name but a few)—and

the clear opportunities to improve critical areas of the business such as supply chain management

and inventory control, it seems particularly surprising that these sectors would rate so low on our

survey.

Sunil Verma, Chief Information Officer and Senior Vice President of US-based clothing retailer

The 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 quickly

understand that there is a competitive advantage in doing real-time analytics and the idea that we

need to infuse every customer touch point with some element of personalization. The minute you

start to internalize that, the need for real-time computing becomes more urgent.”

According to Ranjay Gulati, the Jaime and Josefina Chua Tiampo Professor of Business

Administration at Harvard Business School, the reasons may center on a lack of understanding of

its benefits. “The retail sector is pretty fast-moving— looks at how quickly fads and fashions

change. Yet retailers don’t have that kind of cycle speed on being able to look at the data and run

analytics on it. Even worse, the people who have the data and the people who need the data are

not sitting in the same place—in fact; the people who need the information might not even know

it exists. So there is a huge disconnect.”

7

Still, the fact that so many respondents say their firms are planning to move in this direction

points to an understanding that this disconnect exists—an encouraging step. “I don’t think we’re

at the tipping point yet,” says Professor Gulati. “It’s still early stages. But we are approaching a

point where business competition is becoming so strong that it will eventually force companies

to find a way to get it done.”

Drivers of adoption

For “early implementers”—firms that have already adopted real-time business methods, use is

pervasive 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-time

business is substantial. For each of the functions identified (consumer experience; supply-chain

management; production; sales and marketing; financial and business risk; and product design

and 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%), followed

closely 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 for

adoption. This is followed by building service and/or quality advantages (71%), expanding

addressable markets (68%) and building cost advantages (67%).

8

Meanwhile, “planners”—firms that are in the process of moving to real-time operations or

intend to do so over the next five years—comprise some 46% of survey respondents. For these

respondents, building service and/or quality advantage is most frequently identified (87%) as

an important or very important driver. This is followed by ambitions to grow market share

(83%) and building cost advantages (82%).

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, with

roughly 77% of those surveyed reporting fairly or very significant plans in these aspects of their

operations. The next areas for investment are supply chain management, where 75% of planners

have 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, most

consider 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 greatest

perceived benefit is building cost advantages (85%). In contrast, a higher proportion of new

adopters in the consumer products sector (87%) cite increasing market share as the greatest

anticipated benefits.

Benefits and opportunities

Clearly, the purpose of real-time operations is to help companies make better decisions and

speed 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 operations

seem to offer a tangible benefit: On average, companies that have implemented such systems are

seeing revenue gains of 21% and cost reductions of 19%. In fact, among early adopters, 77%

report revenue gains.

10

For some companies, the benefits are even more substantial. Oil and gas firms report revenue

gains of 36%, which are significantly greater than those in the consumer products sector

(revenue gains of 14%). Business-to-business operations (27%), meanwhile, also report

significantly higher revenue gains than business-to-customer units (13%). The higher figures

may be attributed to the fact that firms in these categories are further along the implementation

curve, suggesting a connection between real-time operations and revenue.

At GE Energy, work is under way to embed real-time sensing software and controls into

equipment to run its plants. “There is a huge wave of activity that is taking place now around

data collection,” says Dr. Peter Evans, head of strategy for GE Energy. “We are putting in place

the 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 over

1,000 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 members

and manages four million rentals monthly across Europe, the conversion rate of trial to paid

subscriptions is critical. “A 1% or 2% change in conversion has a huge effect on the bottom

line,” 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 tracking

credit cards and payments in real time, LOVEFiLM can tailor its offerings to members based on

their past preferences.

Furthermore, because LOVEFiLM can track instantly which films are most popular among its

customer segments, its content team knows which films it should promote to drive sales. “The

sooner we have the customer data,” says Mr. Blakemore, “the quicker we can adapt.” It’s a

strategy that has paid off. In February 2008, Amazon became the firm’s largest shareholder; this

January, 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 are

mainly digitally. Among airline executives, for example, the expression “wheels up” is a key

phrase: Once a plane takes off, the value of an empty seat is zero—it can never again generate

revenue. To keep financial performance aloft at Continental Airlines (now a unit of United

Continental Holdings), real-time analytics track empty seats prior to takeoff to put as many

travelers on board as possible. At any time, Continental executives can see an accurate account

of revenue-generating passengers on its flights.

But real-time data helps with more than in-flight operations. It is now critical in helping the

airline manage seat availability for its 2,000 daily flights, each of which can have as many as 30

different seating classifications. In the past, says Anne Marie Reynolds, director of the airline’s

data 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 that

dramatically but now we are able to respond within a few minutes rather than waiting until the

next morning.” That alone, she says, has meant millions in increased revenues.

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 cost

reductions average 20%; and the gains in productivity 24%.

At The Children’s Place, says Mr. Verma, the company is undergoing a significant

transformation. “We have a new mandate to elevate what we do both in terms of product and in

terms of how we service the customer,” he says. Real-time operations are critical to that

mandate, particularly in terms of personalization. “Having access to immediate customer data is

a huge opportunity,” he says. For example, in the past it took 30 days to know if a customer had

made a purchase online or in a store. “Because a customer’s status couldn’t be updated

immediately,” he explains, “they may have received suboptimal offers or incentives.”

Despite being most pessimistic about future revenue gains and cost reductions, respondents in

the consumer products sector anticipate the biggest increase in these measures (when comparing

between what has been realized in the past and what is expected for the future). This is likely

because 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 cost

reductions are expected to be smaller than what was previously realized, though they will remain

higher than in other sectors. This further underscores the maturity of adoption in this industry.

Assessing Effectiveness

While early implementers agree that real-time business is effective, there is some variation in the

degree of effectiveness among functions. For example, early adopters rank its effectiveness more

highly in consumer experience (69%) and production processes (65%) than in financial and

business risk (58%) and sales and marketing (55%).

12

Real-time operations are seen as more effective by firms in the oil and gas sector than in the

other industries surveyed, perhaps as a consequence of the high level of adoption within the

sector. 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 business

risks, underscoring the power of real-time information in complex, capital-intensive

organizations.

The weak ranking of sales and marketing likely reflects the commodity nature of much of the

industry’s output. Meanwhile, early implementers from the retail sector are more likely to see the

initiative as ineffective or of limited effectiveness than their colleagues in other industries.

Future Investments

The vast majority of early adopters plan further investment across their business functions. For

them, consumer experience, supply-chain management, and product design and innovation rank

top 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 operations

Fewer early implementers in oil and gas plan for further investment in sales and marketing

functions; instead, they plan to extend real-time approaches in production processes, supply-

chain management, and product design and innovation. Early adopters in retail, meanwhile, have

more modest plans for investment than other sectors. Nevertheless, for all business functions

except production processes, more than 40% of these firms plan for intensive expansion of their

capabilities.

As noted earlier, those planning to roll out real-time systems have different strategies than early

adopters. Rather than focusing on customer experience, they will focus on sales, marketing and

supply-chain management. These firms may be missing an important aspect of real-time

business, as early adopters rate consumer experience as the most effective area for real-time

operations. Meanwhile, almost all early adopters plan to continue investing and expanding their

real-time efforts, though they will move their spending from sales and marketing to product

design and innovation, while continuing to focus on customers.

13

Barriers to operating in real time

Despite the understood value of real-time systems, significant challenges remain with regard to

adoption. At the top of this list, respondents note the lack of availability or adoption of

technology (59%), suppliers’ unfamiliarity with real-time systems (58%) and a general lack of

internal knowledge or awareness of (58%) real-time operations.

Professor Gulati is not surprised by these findings, and points to data silos as one culprit. Similar

to 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 and

figure it out, is not in a position to impact change.” The resulting opacity around what types of

data 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 the

information 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. Verma

arrived at the company a year ago, there were at least six separate merchandising systems, each

with its own set of data. Moving to a single system requires full-on collaboration and

communication—setting standards for data, and creating processes for generating reports and

sharing information. At the same time, asks Mr. Verma, “how do you get people to believe that

the new way is better?” Like any organization undergoing transformation, “there’s caution

among associates who have been here a while who have seen things work a certain way. You

have 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 how

systems within individual business areas work together to connect business processes, they

become more open to change.”

By sector, consumer products firms see a lack of available technology as a particular barrier, and

generally judge the obstacles in the way of implementing real-time business to be greater than

other sectors. Surprising given their lead in using realtime business, oil and gas companies see

substantial barriers around communicating the benefits of real-time operations. From a cultural

perspective, “it takes time to roll these things out so that it becomes embedded in day-to-day

14

operations,” says Dr. Evans. “Given the high stakes that are involved in sustaining large energy

systems, 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 be

integrating 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 once

they’re fully integrated, so all we have to do is figure out how to back populate the United

history and incorporate the new United forecasting system into the revenue management

calculations. It’s a lot of work—there are cases where United’s data doesn’t have all the

attributes 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, more

than half cite a lack of availability of technology, a lack of internal expertise and an inability to

communicate the benefits as quite or very important. It’s possible that for these firms, real-time

data is not necessary for operations. However, it is more likely that these respondents have not

yet 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 says

there are three steps executives must consider. The first is to make sure the information that is

gathered across the firm is accurate. Professor Gulati refers to this as the information

architecture. This may seem obvious, but executives constantly underestimate the complexity of

creating 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 the

organizational architecture,” Professor Gulati says. “Without clarity around who needs access to

what information, you’re working in a vacuum. No technology ever solved a business problem

by itself.”

The third step is to take a hard look at how decisions are actually made inside companies—which

can be different from understanding the organizational processes. Professor Gulati calls this the

15

decision architecture. “Think in terms of the reporting, and who has what mandates.” This will

require the IT team to work closely with executives across functions to ensure that the systems

are designed to be truly useful. “What happens in a lot of the initiatives that are CIO-driven is

that they work on the information, build a data warehouse and then add analytics. And then what

do you find? Nobody uses it.

Conclusion: Looking ahead

As our respondents confirm, use of real-time systems will only grow more pervasive in the

coming 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 business

intelligence 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 be

instrumental in helping the company move to enable mobile commerce—a move that he believes

will transform the retail industry. “In the near future customers aren’t going to come into a store

to find their size. It will be more a scenario where on a whim, a customer decides he or she needs

something. They will use their mobile device to figure out what they are going to buy and where

they 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 of

films.“We have capacity to scale based on demand, and the only constraints are around the

physical 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 is

overloaded, we can switch to a different route to deliver content.” That, says Mr. Blakemore, will

dramatically 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 to

have more information and control of their energy use,” says Dr. Evans. While some of that

work is already happening with GE’s business-to business customers, Dr. Evans envisions a day

when a smart electricity grid provides transparent information to consumers so they can better

control the electricity consumed by the various appliances in their home.

16

Real-Time Data Analytics

Identified and Developed Case Studies

17

Case Study 1: KTH – Royal Institute of Technology

Overview

Analyzing large volumes of streaming data in real time is leading to smarter, more efficient and

environmentally friendly traffic in urban areas

Business Need: Researchers at KTH, Sweden’s leading technical university, gather real-time traffic data from a

variety 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 better

manage traffic is a difficult task.

Solution: Collected data is now flowing into a unique software tool that analyzes large volumes of

streaming, real-time data, both structured and unstructured. The data is then used to help

intelligently identify current conditions, and estimate how long it would take to travel from point

to point in the city, offer advice on various travel alternatives, such as routes, and eventually help

improve traffic in a metropolitan area.

Benefits: Uses diverse data, including GPS locations, weather conditions, speeds and flows from sensors

on 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, 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

Real 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

18

Case Study 2: Pepperl+Fuchs

Overview:

After evaluating different alternatives, Pepperl+Fuchs decided to implement the log-based

technology IBM® InfoSphere™ Change Data Capture (previously DataMirror) for data

replication in the heterogeneous database environment. This means that data changes within

heterogeneous data storage systems can be monitored in real time and only the changed data

transferred.

Business need: Following the introduction of a central Enterprise Resource Planning (ERP) system, data in

different worldwide locations had to be integrated and synchronized so that production processes

and reporting on site can continue to be optimally supported.

Solution: Pepperl+Fuchs implemented IBM InfoSphere Change Date Capture (previously DataMirror) for

real-time data synchronization between different database platforms. The solution provides

minimal load on the central ERP system, high performance and reliable data integrity.

Benefits: Error tolerance and stability reduce administrator costs. Efficient implementation by a small

team within just a few days helped ensure low implementation costs and transparent training

efforts. 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 receive

current 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.

19

Case Study 3: Ducati

Overview

Around the world motorcycle manufacturers are rolling out new models to delight bikers and

define success on and off the track. Ducati is responsible for the construction, importation and

distribution 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 is

being incorporated into the manufacturing of a motorcycle so that the company can rectify the

situation before the vehicle is manufactured, sold and driven.

Benefits: Projected 50% savings on revision campaigns and maintenance and monitoring of production

line ~ 30% decrease in time for scheduling and verification of products ~20% savings in

reclaimed scrap and missing parts

Tracking parts automatically Leveraging a strong preexisting business relationship, Ducati decided to design and implement

the RFID tag solution, which would enable real-time, automatic identification of each vehicle

and 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 efficiency The manufacturer deployed a total of three IBM® System x® 336 servers running the SUSE

Linux® V10 operating system to support IBM WebSphere® software and an IBM DB2® 9 data

server. Employees use wireless devices to scan RFID-tagged components and VINs into the

system. Approximately 100 users leverage the DB2 9 data server, 60 of whom can access it

concurrently. A custom-developed Java Platform, Enterprise Edition (Java EE) application—

powered by IBM WebSphere Application Server Network Deployment V6 software—will

enable employees to track inventory throughout the production life cycle.

The manufacturer anticipates that by replacing its legacy component-tracking system with an

RFID solution, it will reduce errors and improve production efficiency. For the company, fewer

20

errors translates into a higher-quality finished product, which will help improve customer

satisfaction and could potentially expand the company’s already large share of one European

motorcycle market. The new solution will also help improve the company’s ability to provide

suppliers with accurate inventory data. Using real-time data collected from the WebSphere-based

inventory-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 savings If a motorcycle is destined to be delivered to France and the factory installs a speedometer that

displays miles instead of kilometers, the solution alerts management in real time. “That way we

reduce the amount of errors directly in the assembly line,” says the coordinator of informatics

systems. “Another important innovation is replenishment of the components from the line to the

warehouse.” 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 and

verification of production, so products reach the market 30 percent faster. In addition, we’ve

seen a 20 percent decrease in the cost of missing parts and scrap. We’ve had improvements in

other indirect costs because of the efficiency of the production line. All of these have shown

improvements in quality and service.”

Preventing revision campaigns As yet, Ducati hasn’t had a revision campaign in which it has had to test its new technology―a

sign that prevention is happening in the factory. “By avoiding the cost of one revision campaign

we 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 export

this SOA technology, this business project management, to our outside suppliers, so that our

suppliers can have this information at the same time we do. We also want to know what our

suppliers are doing at a particular moment for us, and therefore we will be able to consider if

something will be just in time or if it’s delayed. We’re also eager to change over to the smart

factory with smart supervision. And we will be doing this all in real time using Web 2.0

technology and the latest innovations in SOA and business project management, thanks to IBM’s

help.”

The improvements to the assembly line and inventory system have brought about efficiencies

that were unthinkable just three to four years ago, and the coordinator of informatics systems

hopes there will be more benefits to enjoy three or four years from now

Smarter car manufacturing Ducati 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 to

correct the mistake. Now the company is working smarter and benefitting from savings in time

and money with the automatic monitoring and alerting performed by its RFID parts tracking

solution.

21

Case Study 4: Heathrow Airport

Overview:

In recent years, Heathrow faced a challenge not only keeping up with existing traffic but bracing

for major increases in traffic. Due to political and geographic issues, the airport had to live

within 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 to

tune ground crews to incoming traffic by linking to the Euro Air Space system to give crews a

full 30 minutes warning before a plane arrives. A business activity monitoring (BAM) capability

is also used to manage crews, gate assignments, docking and undocking, refueling, cleaning, and

more.

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, if

significant delays are expected for departing passengers, security systems are adjusted to keep

people in retail areas rather than gate areas for a longer period. This has resulted in happier

passengers and security workers and also had the serendipitous result of substantially raising

retail revenues.

“Today’s BPM handles expected exceptions but with newer tools you can start looking for

unexpected exceptions,” says Jim Sinur of Gartner Research. For instance, he indicated, you can

start to look for signs of inflation or deflation. You can look at complex event patterns and

certain recurring events and combine that with modeling.

“That is the top level—where someone managing the process gets notification either from

complex events or modeling tools that there is a good chance something unusual is happening,”

says Sinur.

Seek, model, adapt

Sinur 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 have

missed the “seeking” part, often because information exists across multiple processes.

He says the most successful users of BPM real time analytics should consider taking different

approaches, depending on whether they are a planning-oriented or reactive company and

depending on the specific analytic capabilities their system provides. Plan-oriented companies

often prepare for different scenarios and compose alternative responses for each—so when BPM

provides an early warning, they can leverage that information for quick action. Reactive

companies, on the other hand, may be able to leverage elements of real time analysis to “learn by

doing”—inventing new ways to operate quickly with the help of BPM analytics. “Some BPM

systems can provide modeling tools and automated process discovery to look for patterns,” says

Sinur. “This stuff is hot and will be for the next decade—companies are just starting to discover

the potential this represents,” he adds.

22

Case Study 5: KEMCO

Overview:

Korea Energy Management Corporation (KEMCO) is an organization that contributes the

national economy and develops and supplies future energy. KEMCO has created and provided

the 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 these

times and the related laws have been embodied, the organization felt the need of having the new

class of business efficiency. KEMCO selected the Voluntary Agreement (VA)-related business

system started from 1998 as the first target for innovation and then they got the development of

BI system going in the end of 2009. It has driven the BI project to maximize the ease-of-use

through the Microsoft's solution instead of the existing OLAP tools. Through this, having noticed

the business efficiency of the user environment where they can draw the statistics they want on

Excel without moving to lookup screen each time, KEMCO has a plan to extend the user-

oriented BI into their other system areas.

Situation

The statistical and analytical business is of the highest importance to Korea Entergy

Management Corporation (KEMCO) leading the charge as it tries to implement the green growth

and low carbon economy. Since it is the organization managing the demands for energy, the

statistics 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 by

key business system. Prior to creating reports, they had to log in the data lookup site of the

related business system and extract the data. And if there is any hard-to-acquire information from

the view, they had requested it to Information Statistics Division and KEMCO via e-mails or

phone calls. After collecting data, they arranged and analyzed it in Excel to create reports

ordinarily.

The need of new class of efficiency has not been on the rise in such a business process until the

'green' emerged as the keyword passing through this time. In a major move towards green growth

everywhere including economy and society, the government declared the low carbon green

growth as a national vision in 2008.

To support this institutionally, the preparation for the Framework Act on Low Carbon, Green

Growth (hereinafter, Green Act) is under discussion commonly. Naturally, as the KEMCO's

stature and role has grown, for statistics and analysis business as well, they need to prepare the

future changes with the stream of times.

VA (Voluntary Agreement)-related business was picked up as the first target for innovation. VA

is the non-compulsory energy saving measurement where a corporation sets itself the objective

of energy saving or greenhouse gas emission reduction for five years and makes an agreement

with government, and then the government helps the company fulfill this agreement to make the

objectives achieved with the collaborative effort of two parties.

23

KEMCO has collected and analyzed a number of information for evaluating and monitoring VA

related corporations' fulfillment. As the demand for this statistical and analytical information has

increased in an era of Green Act, the VA related business is chosen first as a target for

innovation.

Yeong-sang Yoon, Leader of Information Statistics Division, KEMCO, stated, "Analytic works

related to VA has increased. It is because of increased corporations participating to VA, and

Green 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 of

Green Act."

Actually, in addition to its existing job of establishing the energy saving goal and then

monitoring and evaluating its enforcement, its statistical and analytical activity will be extended

to the measurement, reporting and verification of greenhouse gas emission for the emission

trading scheme between corporations regarding the industrial development initiatives.

Solution

KEMCO selected the new approach of BI (Business Intelligence) for innovating statistical and

analytical business related to VA system. Making a decision that it was no more efficient to

develop the screen each time needed, KEMCO viewed BI the best way, since the kind and the

number 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 of

RealWeb which was responsible for this BI project, explained, "KEMCO has ever introduced the

PowerPlay of Cognos and OLAP tools developed by a domestic development firm. As it difficult

to use these tools without expertise, its usage was not too high. Thus these proprietary tools were

not 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 data

exported from VA systems operated on SQL Server 2005 and distributed and shared through the

Microsoft Office SharePoint Server 2007's Excel Services. The existing lookup screens were

removed except the page used to access by its staff outside the company. Besides, About 10 data

marts were created to reside on SQL Server 2008.

Finishing BI development for VA system successfully, KEMCO is considering extending BI into

its other systems step by step. Mr. Yoon stated, "We have noticed that BI was very useful from

the case of VA system. Now we are in discussions with Microsoft about applying BI to Climate

Change related business and have a plan of approaching to all sorts of statistical and analytical

business which we do from a enterprise-wide BI perspective in the mid- and long-term."

24

Benefits

Complete the data collection that used to take 3 days in real time Staffs dealing with data analysis related to VA are feeling that all the before and after of BI

implementation 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 request

the materials to the Information Statistics Division and it took at least 2-3 days to receive it. Now

we don't have to spend our time to wait materials. We can bring the data on data marts in system

into Excel."

Respond immediately to the request of a variety of statistical reports Analysts don't have to drill down to 10~20 lookup pages or request the materials they need to the

information statistics division, which increase their responsiveness to businesses.

Mr. Yoon said, "Unlike general enterprises, a public institution has to consider the statistics in

building business systems. It has to submit the data to superior authorities if they request. After

the SQL Server based BI system was implemented, we don't have to look up the law data case by

case, freeing us to focus on our own tasks including writing graphs, which makes our

responsiveness to statistical and analytical business increased."

Reduce cost about 60% compared to that of OLAP dedicated tool KEMCO broke stereotypes that BI is the expensive tool from this project. It was not a challenge

followed by big investment once the organization actually tried to do it. They found that their

will and practice were important.

Mr. Yoon explained, "While Introducing OLAP tool will need at least 100 million won to be

invested, 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 of

applying BI into the system that many people use in the future. Unlike the VA system that 50

analysts use, in case of the system that has to support more users, cost benefits of Microsoft's BI

solution is expected to be much higher compared to that of the traditional OLAP tool.

Ensure a smooth transition One of results of moving from lookup screen to BI is the smooth transition of VA related

business. Mr. Yoon said, "A person in charge is changed regularly in our organization, so there

was a difficulty in transition for the successor in analytical business. Analysts often need their

personalized environment due to the nature of statistical and analytical tasks. Thus in case of

delivering information based on lookup screen in the past, the new analyst often requested a new

screen to be developed. Now using Excel ensures a smooth transition

25

Case Study 6: Faro Hospital

Overview: Faro Hospital provides medical services for citizens in the Algarve region of Portugal. The

organisation wanted to improve decision making and increase efficiency. It deployed a business

intelligence 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 helped

the hospital reduce costs and improve patient care.

Business Needs Founded in 1979, Faro Hospital provides medical services for more than 250,000 people in the

Algarve. The hospital complex covers 46,500 square meters and has 530 beds. In December

1999, it became a state-managed organisation, with specific government targets. In addition to

providing high-quality medical care to patients, the hospital managers needed to streamline

processes and minimize costs.

Until recently, each department used its own applications to manage patient records and

employee schedules. Francisco Serra, Head of IT Strategy for Faro Hospital, says: “It was

slow—often impossible—to share information across departments quickly.” For example, any

time nurses needed outpatient treatment information they had to request a manual transfer of the

records from the outpatient clinic. This slowed patient treatments, and prevented the hospital

from providing the quality of care to which it aspired.”

This lack of integration also delayed decision making. “It took a long time to consolidate

sufficient information to achieve an overview of the entire hospital,” says Serra. “Managers had

no day-to-day view of the organization’s activities, so it was difficult for them to make crucial

decisions quickly. For example, if a problem occurred, such as a need to arrange cover for a

surgeon 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 time

collecting, organising, and re-entering information, instead of doing their jobs. For example, if an

employee needed information from another department, such as medical records, it would have

to be requested and manually delivered. We needed to increase employee efficiency and start

generating a better return on investment from our workforce,” says Serra.

Solution

In January 2009, Faro Hospital engaged with Microsoft Services to create a hospital management

solution based on Microsoft SQL Server 2008 data management software, Microsoft Office

SharePoint Server 2007, and Microsoft Office PerformancePoint Server 2007 business

intelligence software. The solution consolidates data, such as patient records, appointments, and

staff schedules, into a centralized repository.

26

Office SharePoint Server 2007 acts as the hospital intranet where employees can update and

view patient data. If a patient is transferred to a different department or specialist, this

information is always captured in Office SharePoint Server 2007, ensuring that treatment history

can be accessed by doctors and nurses in different departments.

Benefits

Since 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 employee

workloads. 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

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

28

Predictive Data Analytics

Identified and Developed Case Studies

29

Case Study 1: Ingersoll Rand

Overview:

Ingersoll Rand is a global, diversified industrial firm providing products, services, and solutions

that enhance the quality of air and comfort in homes, buildings, and transport and that protect

food and perishables, secure homes and commercial properties, and increase industrial

productivity and efficiency.

Given the breadth of its product offerings and its global reach, the company has a complex

network of manufacturing and distribution operations. To increase efficiency while ensuring that

it continues to deliver the superior service its customers expect, Ingersoll Rand implemented

Oracle Transportation Management to coordinate its outbound shipping across various lines of

business and facilities.

Even while the application was implemented currently at eight sites of the company’s more than

70 locations, it already exceeded expectations for improved productivity and cost savings. With

Oracle Transportation Management, users can view shipping plans across lines of business and

facilities, enabling them to adjust shipments to increase efficiency, such as making multiple stops

within one shipment. The application has already contributed to more than 2 million dollars in

savings 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 delays

Solutions

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%

30

Case Study 2: EMC Insurance Companies

Overview:

EMC Insurance Companies struggled with pinpointing the right amount of money to hold in

reserve against potential case payouts; holding back either too much or too little could be

disadvantageous 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 selected

PolyVista, advanced data analytics software built on Microsoft SQL Server Analysis Services, to

uncover anomalies, correlations, relationships, and patterns hidden within the firm’s warehouse

of claim data. After deploying the solution, EMC was able to improve financial reserve

management, identify claims requiring special attention, improve data quality, support executive

decision making with improved analysis, and better manage expenses.

Situation

EMC Insurance Companies sell workers’ compensation, property, and casualty insurance

through independent insurance agents in 16 branch offices across the United States. Founded in

1911, EMC is one of the largest insurance companies in its home state of Iowa. With assets of

approximately U.S.$3 billion, the insurance firm has the bulk of its volume in commercial lines

of business, with the remaining 15 percent in personal lines.

The Challenge of Financial Reserve Management

EMC manages its loss reserves—money set aside for paying future insurance claims—by setting

up a reserve for every case that the company handles. “When a claim walks in the door, we try to

determine what is the most likely outcome for that claim,” explains Rich Schulz, Senior Vice

President of Claims for EMC Insurance Companies. Strategic management of financial reserves

can be a key contributor to the success of any insurance firm, and they are important indicators

that 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 is

critical,” 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 realized

we had to get a better handle on the way we were setting our reserves, and I was brought in to

EMC 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 as

increased costs, additional services, or a lengthy resolution. “We had all this claim data and not

an effective way to look at it,” says Schulz. “We wanted a tool that would allow us to look at

trends that the company was experiencing and be able to figure out what was driving them.” He

looked at a number of predictive modeling tools available on the market, but found that initial

investment costs were high and the cost of maintaining those tools was far beyond what the

company was willing to pay.

31

The Need to Identify Trends EMC stored a great deal of data in its internal claim processing system, but managers needed the

ability to sift through that data more efficiently. For example, executives at EMC wanted to track

and understand company trends related to claim frequency, which is the ratio of claim volume to

premium volume, and severity, which is the ratio of claim cost to premium volume. Not only did

they need to track these measures in order to manage their business, but executives also were

expected 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 and

frequency of workers’ compensation claims,” says Schulz. “At the time, I had to call individual

branches to inquire as to what was going on.” EMC had no good system of reporting that could

be quickly used to discover the reasons behind unexpected changes.

Schulz also needed to be able to track expenses both across the company and by individual

branch, 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 ability

to understand where expenses were running high is important. All of our adjustment expenses

are open to analysis. [For example,] do we have one or two business units that are overusing

certain types of vendors to investigate claims?”

Solution

In need of an affordable and effective solution, EMC Insurance Companies heard about

PolyVista business intelligence software, which provides two important capabilities: A business

user 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 custom

code or predefining the anticipated results.

EMC was interested in determining whether the PolyVista software, which uses Microsoft SQL

Server 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 a

robust open development environment that allows us to build and enhance our unique

sophisticated 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 IT

personnel at EMC completed training.

After EMC completed the training, PolyVista sent a senior consultant to EMC to train key IT

staff 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 less

than 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 with

the software.

32

Benefits From the beginning, EMC Insurance Companies anticipated many uses for the business

intelligence solution. The initial goal was to improve reserve account management; however, the

insurance firm is also building predictive modeling into the claim system and has improved the

quality of data. Finally, EMC is improving decision making and controlling expenses.

Effective Management of Financial Reserves

EMC relies on PolyVista as its primary tool for managing case reserves. Since the

implementation of the business intelligence solution, EMC has managed its financial reserves

more accurately and the company has avoided the volatility that can occur when reserves are not

consistently and accurately established.

The software has helped the firm ensure that all branches are in compliance with the company’s

reserve guidelines.”

Predictive Modeling System for Improved Claim Outcomes

EMC is using PolyVista to add predictive capabilities to the automated claim system. Now, the

company can identify casualty and workers’ compensation claims that are likely to have a

negative outcome and assign these claims to the company’s most experienced people to help

mitigate potential risks. If a claim is not properly identified or assigned, it can take longer to

resolve and result in additional costs and services. By recognizing the complexity and risk

factors early on, an experienced claim manager can expedite the claim with better results for both

the claimant and EMC.

Better Data Quality Enhanced reliability and accuracy of data with the business intelligence solution was an

important realization for EMC.

In the highly regulated insurance industry, data quality is an important aspect of compliance. For

example, Schulz says, “Under the federal government’s Medicare Secondary Payer laws, we

have to have the age of the claimant or insured in our system. We now can run reports showing

the portion of claims that don’t have an age or have an age that doesn’t match other data or

doesn’t make sense.” These reports allow EMC to identify and fix errors before submitting data

to the Medicare program.

Improved Executive Analysis to Support Decision Making During periods of rapid change, EMC relies on its business intelligence solution to help

executives turn data into knowledge and make more informed decisions. “When you have

weather events or the economy is changing in a specific area, you need to understand how it is

affecting the business,” says Schulz.

Effective Control of Expenses By using the business intelligence solution, EMC is able to track expenses and control costs that

are tied to specific buys, and match data by business unit.

33

Case Study 3: National Science Foundation

Overview:

The NSF is helping to predict earthquakes across the world in real time. The system has been

developed by a consortium of universities and government agencies, with funding from NSF

(National Science Foundation), to provide an infrastructure of networked tools for research in

ocean science – constructing an internet-based system to collect and share data.

Real-time Seismic Monitoring Program

This is a large system with 16,000 land and sea-based sensors, each sending several channels of

seismic event data at a rate of 40 data points per second. The application executes in an Amazon

EC2 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 to

view the AMQP bus as a domain of input and output data streams. The initial prototype was

upgraded to a full-scale system running on a cluster of servers without any changes to the

streaming 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.

34

Seismic Monitoring:

The sensor network contains about 16,000 seismic sensors, organized into grids, covering large

parts of the North American continent and the adjacent oceans (see illustration below). Each

sensor measures the motion of the ground under it in three dimensions, and transmits its data as

several 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 a

seismic event are heuristics that apply to a time interval of several minutes: so the application has

to calculate some quantities from the raw data and to store these calculated values over a time

window.

But to detect an earthquake reliably, it’s better to monitor all the sensors at once, looking for a

disturbance in the signals that first appear in one place, and then appear in nearby places: a

disturbance signal that propagates and changes shape in a way consistent with the physics of a

seismic wave.

Real-time sensor network management in an EC2 Cloud Monitoring tens of thousands of signal channels arriving at 40 sample points per second is a

complicated problem, but it can be made simpler by breaking it into stages. We’re interested in

earthquakes, which are infrequent, so SQLstream first reduces the amount of data by scanning

each channel for patterns that suggest the beginning, the peak or the end of a quake: in other

words reduce the dense signal to a sequence of interesting events. Then we can look for events

detected on other channels that could be due to the same quake propagating in physical space and

time.

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 sensor

data infrastructure – using our AMQP adapter. This involved less than 100 lines of streaming

SQL.

In the second phase, the prototype was scaled up to a full-size system, dealing with 16,000

sensor channels, running on multiple SQLstream server nodes created automatically in an

Amazon Elastic Cloud. This expansion required no changes to the streaming SQL application

developed for the initial prototype – simply running the same streaming SQL application inside

an elastic container/manager.

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.

36

Case Study 4: Laboratorios Indas

Overview:

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 as

pharmacies, 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, and

the 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 precise

intelligence 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 uploaded

via sales reps’ PDAs, together with demographic data to build sophisticated statistical models

that provide deep insights into likely sales patterns, at the individual product level, for

pharmacies in different locations.

Results: Increased average sales per pharmacy by 37 percent over the prior year. Reduced the average

time it takes sales reps to compile sales data from store visits by 92 percent, from five minutes to

25 seconds.

Benefits: Began generating commercial benefits just four months after implementation by providing

insights into the likely product demand patterns for individual pharmacies. This enabled the

company to target key pharmacies more systematically with visits, emails and mailings to help

drive sales.

Instrumented Captures all sales data and order history, plus client and customer demographics, in a central data

warehouse for data mining.

Interconnected This technology integrates sales information from PDAs in the field with population size,

socioeconomic levels and number of pharmacies in the region, providing real-time predictive

marketing support.

37

Intelligent Increases sales per pharmacy by 37 percent by using statistical analysis to leverage real-time

order history and local demand patterns. It also reduces effort for sales teams while optimizing

sales visits, strengthening customer loyalty and expanding the business.