Foreword Big Data – The Future Is Here! The Big Data explosion has virtually scripted the obituary for traditional data management systems and processes. The volume of data getting generated by the millisecond is mind-boggling. An IDC report says that by 2020 the quantity of electronically stored data will reach 35 trillion gigabytes from 1.2 zettabytes at the end of 2010. That’s apparently enough data to fill a stack of DVDs reaching from the earth to the moon and back. The Big Data environment expectedly poses new challenge to business organizations that have recognized the significance of leveraging the data flow for building competitive strategies but do not possess the requisite tools to make the best use of it. Analytics, a fast emerging solution to this challenge is not just about unraveling the historical trends but is also a compelling tool to predict. Hence, predictive modeling has acquired new-found relevance in the corporate world. Similarly, analytics helps organizations to build efficient ‘sense and respond’ supply chains. For example, media companies can put a cable box in every customer’s home. That device records every action. Companies can analyze this data to search for granular viewing patterns. That can help them make better choices when selecting a new show, or segment its advertisements more effectively. Keeping in view the growing relevance of this subject and the opportunities it presents, in this edition we have put together a set of highly insightful articles on the usage of analytics. Our lead article discusses how enterprises use Big Data to make better business decisions. We discuss how to derive the best insights from the ‘data storm’. We have also turned the spotlight on the usage of analytics in key sectors – retail, manufacturing, and health & life sciences. We have taken special care to put together the most apt pieces connected with the theme, and we hope that you will derive great value from the articles presented here.
Rajan Kohli CMO - Wipro Global IT Business
02 CONVERSATION
CONVERSATION 03
Enterprises, physicians, stock traders, oil rig operators — everyone these days is awash in data. How do you make sense of all the information? How do you use it to make better business decisions and ultimately more money? Conversation with: K R Sanjiv, Senior Vice President for Analytics and Information Management Enterprises Beth Ellyn Rosenthal, Editor Outsourcing Center
Wipro has created a new service line: Analytics and Information Management to help its customers manage large volumes of unsegregated data. The service provider has included analytics in its offerings for years. But what’s different now is analytics “are a priority and aligned closely with business outcomes,” says K R Sanjiv, the new division’s senior vice president. The new service line addresses one of the trends Sanjiv called “Big Data.” Three attributes create the Big Data environment: 1.Volume: Sanjiv points out transaction databases could have hundreds of terabytes of data and exploding every year. 2. Types of data: In addition to structured data (the data in databases), companies now have to analyze unstructured data like texts, blogs and messages from the Web. It also includes feeds from companies like Reuters or Bloomberg and other syndication sources. It includes data from machines or sensors like those placed in an operational oil well or a patient monitoring operating room. Unstructured data is any data not in a relationship format. 3. Real-time usage: Today corporations want to analyze data in real time, not just at the end of the quarter. In fact, the need for speed and right timing outweighs even the need for accuracy in many scenarios.
Defining a sustainable and efficient enterprise architecture for all the above is a complex exercise for most organizations.
04 CONVERSATION
“Wipro recommends a three-tier architecture to handle Big Data scenarios, which can help keep the processing costs down.”
Wipro recommends a three-tier architecture: 1. Conventional databases and datamarts tier for handling the structured data. 2. Appliances tier that processes closer to the hardware and in memory for faster and realtime processing. These can be up to 10 times faster than conventional processing. 3. Unstructured data tier for processing, parsing and analyzing the unstructured content.
Next, they need to look at their business processes and link their key performance indicators (KPIs) and their required return on investment to what these new technologies can provide. For example, a media company puts a cable box in every customer’s home. That device records every action. Companies can analyze this data to search for granular viewing patterns. “This information has business value,” Sanjiv explains. The company can make better choices when selecting a new show, or segment its advertisements more effectively. “The data lets them monetize their insights because they now know how to optimize their spend,” he continues.
“Processing happens in the appropriate layer to keep the cost down,” says Sanjiv. Also a uniform layer to load data into the appropriate layer and integrate the outputs from each layer to provide an integrated look is essential for success.
In the healthcare field, elderly patients now have monitoring devices so they can age at home. These devices send information on regular basis. Analytical applications can detect abnormal patterns and alert the concerned physicians who can then look at the data to make the correct diagnosis from afar.
HOW CAN INFORMATION MANAGEMENT HELP OUTSOURCING BUYERS?
The three-tiered architecture allows outsourcing buyers “to implement cost-effective uses that were not possible before. These innovations create real impact,” Sanjiv says. He suggests corporations identify one or two processes to explore new options, options that were not possible before because either the technology was too expensive or just not available.
Sanjiv says this is a two-step process. First, outsourcing buyers need “to get a first hand understanding of this new technology so they don’t get carried away by the hype.” They can do this by doing a proof of technology or a pilot around these technologies.
CONVERSATION 05
06 CONVERSATION
5 KEYS TO SUCCESS IN INFORMATION MANAGEMENT
3. Build an analytics service that can take inputs from the transactional system and do analysis and mining on the large data. This is required where patterns or correlations have to be detected by using the whole data. As an example, a credit card company would want to drill down to analyze the data to detect fraud committed by one person. 4. Align Business Value: Start with ‘Business performance backwards’. The use cases that will be implemented should have a clear financial ROI, or a clear set of decisions that will be driven as an outcome of the implementation. 5. Create Strong governance, change management and security. Compliance has to be uniform across all three platforms.
after completing many pilots for clients , wipro ’ s executives 1. Create a standard reference architecture at the outset. Without standardization, “different lines of business will use the same technology in different ways, resulting in duplications and related quality issues.”
noticed they were creating new architectures for this technology that co - exist with their existing technology .
2. Build a summarization capability in the unstructured data layer. This is necessary because of the large volumes involved. This layer must have the ability to do the summarizations in the huge mass of data and then pass these on to the structured layer or other analytical applications for processing. To use the medical monitoring device example, the processing layer would extract the abnormal readings only and pass them on to the medical application or tell the physicians. That way the application would only have to deal with a manageable amount of relevant data.
“ companies
are building
around their legacy systems ,” sanjiv reports . that ’ s good news . no one has to jettison systems they are currently using to take advantage of the data they are producing . Reprinted with permission from Outsourcing Center, Copyright 2011
07
08 RETAIL
Analytics Driven Retailing A Thought Paper from Wipro and RIS News
Despite making steady investments in analytics over the past few years many retailers acknowledge they have not achieved their desired goals. A majority say they have difficulty overcoming barriers caused by corporate culture and an inability to demonstrate return on investment. And even many of those that have made recent software upgrades admit their analytical capabilities still need improvement. On the opposite end of the spectrum are a select few high-performing retailers that have made the discipline of analytics central to their execution of strategy. Proven experiences by these retailers have demonstrated that sophisticated, analytically driven strategies deliver strong, measurable results. Says Joe Skorupa, “One electronics retailer found that 7% of its customers were responsible for 43% of its sales. An apparel retailer quantified the value of employee engagement to shopper conversion and found that increased engagement is worth $100,000 in annual store revenue. A dollar-store retailer found that instead of having an average customer that spends $20 per visit it had two distinct customers that spend either less than $10 or more than $30. Each of these insights led to changes in strategy that produced dramatic results.”
“When retailers shift from working with averages to working with absolutes they discover a-ha moments that drive business results.” Joe Skorupa, Group Editor-in-Chief of RIS News.
RETAIL 09
decisions and results by making the discipline of analytics central to strategy retailers can take quick actions that solve problems and seize competitive advantage
10 RETAIL
BETTER INSIGHTS, BETTER DECISIONS Retailers know that data is proliferating in volume and type at an accelerating pace. According to research firm IDC, the estimated digital universe is doubling in size every 18 months, as reported in the influential white paper, “The Diverse and Exploding Digital Universe”. Unfortunately, many retailers view management of this resource as a challenge instead of an asset. The result is that data and analytics are underused today and underappreciated.
When a successful analytical ecosystem is in place a retailer can take quick actions that solve problems and launch first-mover initiatives that seize competitive advantage. On a strategic level a majority of retailers understand this, according to the RIS/Wipro report. Retailers say they believe analytics-driven insights will improve margins (77.4%), improve customer satisfaction (74.2%), increase market share (67.7%), and improve comparable store sales (61.3%).
On a tactical level a majority also believes analytics-driven According to a study done by Wipro and RIS News insights will improve forecasting and planning accuracy exclusively for this report, a majority (87.5%), retention/frequency of loyal of retailers have significant internal investment in customers (84.4%), and conversions and external obstacles to for ad/promotion spending (59.4%). analytics has overcome before they can achieve A retail enterprise that gets maximum maximum benefit from advanced steadily increased value out of its analytics capabilities analytical tools. These obstacles in information is one that has an integrated include training power users (58.1%), framework that employs quantitative technology budgets reliable return on investment methods to derive actionable insights (54.8%), changing the overall for approximately from data, and then uses those corporate mindset (54.8%), and 60% of retailers over insights to shape business decisions a clear roadmap to deployment to improve outcomes. (51.6%). (See “Analytics Drives the past three years . “To get to this stage a retailer will Success,” RIS, Aug. 2011) need to do three things,” says Joe Skorupa of RIS News. Retailers need to build a culture that understands and “Demonstrate a commitment to change through toppromotes the capabilities of analytics to shape favorable outcomes.
RETAIL 11
level leadership, break down silos that stifle multidepartment collaboration, and develop key talent to shift the organization to a more fact-based culture.� Training workers in IT skills has been a consuming necessity for retailers for the past few decades, however in the future it will be more about training employees in analytics skills and integrating analytics into everyday work. This raises a question about whether it is better to build this competency internally or work with an external IT services firm to provide the necessary analytical expertise. Since transformation projects like this have implications on corporate culture and organizational processes, the answer to the question is probably a combination of both — internal skill acquisition and external expertise. The end game is the creation of a retail organization where analytics capabilities solve problems, predict outcomes and deliver results. When the transformation is complete, analytics will be an engine that helps drive revenue growth, profitability, customer loyalty and innovation. Success in the hyper-competitive retailing environment will be dominated by merchants that have made analytics an integral capability in their organizations, and the sooner a retailer sets this plan in motion the better prepared it will be to succeed in the future.
12 RETAIL
Predictive Analytics: Driving the Next Wave of Business Disruption Over the next five years the secret to retail success will come from knowing the impact of decisions before they take effect
Using analytics to review past performance is no longer enough to be truly competitive. The reason is that most retailers have already deployed quantitative analytics to make performance improvements such as enabling rapid response to reports and alerts, leaning out supply chain costs and inventory, fine-tuning advertising budgets, reducing labor, and targeting merchandising plans to specific customer segments. But in today’s retail environment something more is needed. Retailers need to shift from using analytics to show what happened (reporting) and what is happening now (alerts) to what will happen. They need to progress from extrapolating how something happened to getting recommendations for optimal decisions and predicting best and worst outcomes. One retailer, for example, found out that its in-stock positions were at a 95% level, which was acceptable, but it dug deeper and discovered that fast-moving products were out of
stock 30% of the time. Then, using sophisticated statistical modeling, forecasting and optimization algorithms, the retailer was able to anticipate the impact of various actions before implementing the ones that reduced the problem and increased sales. Software solutions that make this happen are becoming more capable and more accessible, but only to a select few retailers that use them. According to a study Wipro produced with RIS News, most retailers are already using traditional analytics tools: 50% of retailers have up-to-date performance analytics capabilities in place and 43.5% have up-to-date customer/marketing analytics capabilities. But only 16.1% have up-to-date predictive analytics solutions in place. (See chart alongside). This means smart retailers that adopt predictive analytics now to shape decisions and outcomes will have a competitive advantage in the marketplace for the next three to five years.
PROGRESSING BEYOND HINDSIGHT Today’s rapidly changing, data-driven retail environment is characterized by increased competition, higher customer expectations, and a proliferation of customer channels and touch points that are driving the need for enterprise wide business agility. “Predicting the future during these uncertain times has been a very difficult and risky task. Retailers that sense and respond better and faster than their competition to the increasing demands of the consumer and rapid changes in the economic environment have excelled in our omnipresent world. Leading companies using analytics and leveraging “Big Data” today will outperform their peers tomorrow. We believe analytics-driven experimentation and performance improvement will be the key for successful retailers in the future. With our ability to drive predictive consumer insights and operational analytics, we are sure we will help
RETAIL 13
retailers of tomorrow to do business better by delivering insights that have a significant positive impact on customers business performance,� says Bhanumurthy B M, Senior Vice President, Retail, Consumer Goods, Transportation & Government at Wipro Technologies. Using analytics to deliver business intelligence, alerts and reports is reactive and falls short of the advanced capabilities of predictive analytics software. With current solutions retailers know how many new customers were drawn by a promotional campaign, but only in total. They can uncover some useful insights, such as the average sales lift per store or geographic region, but only in hindsight. However, using predictive analytics these same retailers can do complex, dynamic research using multiple variables that go far beyond simple linear what-if exercises. For example, a retailer can combine assortment and space analytics to see the effect of adjacencies on shopping basket sizes. They can also use purchase path analytics to determine the next likely purchase made by shoppers to adjust assortment plans and store designs. And they can do this by each demographic segment. Predictive analytics requires an integrated approach, one that develops a suite of analytical capabilities, embeds analytics-based
what is the status of your predictive analytics solutions?
29% Updating now 28.5% Will begin by end of year 22.6% No plans 16.1% Up-to-date tech in place 6.5% Plan to update in 2012
processes within the organization, and creates an analytics-based culture that extends throughout the workforce. Moving toward a predictive analytics approach involves shifting from questions that produce reports and alerts to more sophisticated questions like: What is the lifetime value of a customer segment? How can we tap social media metrics to redesign the customer experience across stores and digital channels? How can we use digital channels to drive traffic to stores? What’s the ROI of marketing spend in social media, digital coupon networks and location-based services? How can we make the best real estate purchases, energy reduction plans, and commodity price decisions?
By asking more complex questions and using predictive analytics to make recommendations retailers can test scenarios and assess the impact of decisions before they take effect. The improved results from these efforts will produce a significant marketplace advantage for these retailers over their competitors for the next five years.
14 PERSPECTIVE
Information As An Enterprise Asset the data is growing. why aren’t insights? By: Srini Pallia Senior VP & Global Head Business Application Services Jayant Prabhu Principal Consultant, Analytics Information Management
Worldwide, enterprises are generating vast amounts of data. Not only is the volume of data overwhelming, its complexity is staggering. In a knowledge and information driven economy, data has become an invaluable enterprise asset. A Gar tner repor t says that by year-end 2012 information assets will appear on the balance sheets of 25% of Global 2000 companies.
The data storm To put data availability in perspective, consider this: Google CEO Eric Schmidt recently estimated there were 5 exabytes (billion gigabytes) of information created between the dawn of civilization through 2003. That much information is now created every 2 days and the pace is increasing. * Last year, The Economist gave readers a sense of the crushing power of the data storm while reporting that “experiments at the Large Hadron Collider at CERN, Europe’s particle-physics laboratory near Geneva, generate 40 terabytes every second – orders of magnitude more than can be stored or analyzed. So scientists collect what they can and let the rest dissipate into the ether.”**
PERSPECTIVE 1515
Over the last decade the cost of storing the data has dropped dramatically. Technologies to move, manage, organize, retrieve, mask and secure the data are, however, becoming complex. Information has become so huge that it can bury entire departments. Luckily, the mountains of data also present opportunity. Extracting actionable knowledge and business insight using predictive analytics has become the Holy Grail of business leaders. They know that to do business better, they must create better customer insight. Today’s data availability and analytical tools can help deliver previously unimagined customer, market and business insight. The Daily Data Exhaust
“to-do” items that business leaders must address is creating the expertise to understand the data. With rapid changes in data management technologies, businesses will do well to partner with experts who can help create processes, technologies and best practices to manage their data. Understanding the Challenge With business processes becoming collaborative, enterprises are stumbling across the challenges in uncovering insights locked within their data. Data is in silos across the enterprise. Often the data is difficult to access. The quality of the data is uncertain, rendering it unreliable. Data formats are not necessarily standardized. Before being able to share and use the data across the enterprise, these hurdles must be crossed through effective integration.
A smart phone that is constantly exchanging data with the network is generating a stream of information about the user’s location. The information can be used to target organizations that location-aware advertising. The have established series of clicks on a website are being recorded to create data stacks systems to manage that can be used to decipher information can still behavior patterns of individuals and fail . this is because entire communities. The information they do not have a can be used to deliver context-aware strategy driving it . content driving e-commerce. Wal-Mart operates more than 8,000 stores in 15 countries with half a million Stock Keeping Units each. Every time a customer checks out, hidden in the data generated are relationships between customers, stores, products, services and offers. Analytics has the potential to unearth these relationships and drive accurate personalization leading to direct bottom line gains. Enterprises are eyeing and analyzing the data exhaust of their customers, users, vendors, partners, collaborators, competitors, and their own employees, assets, processes and systems, to get a faster and sharper view of changing business demands. Expectedly, the data types and volumes are so large that it raises questions about the availability of expertise to structure and analyze it. Amongst the top
The question every enterprise must ask itself is: Do we have a sound data and information management strategy? What is the ideal framework for the strategy within my industry? How can I derive the maximum business value from my data?
Information strategy is the key step. It establishes the mechanics that takes the vision of making information a business asset and turns it into an operational reality. Organizations that have established systems to manage information can still fail. This is because they do not have a strategy driving it. The extended nature of enterprises integrated with partners, customers, and competitors add to the challenges of integrating information. What stands between an enterprise and a successful methodology of managing data and information? The key factors are summed up here: • The deluge of both structured and unstructured data customers and employees produce & consume • Siloed storage • Integration complexities • The ever-changing technology landscape
* Google CEO Schmidt: “People Aren’t Ready for the Technology Revolution”, ReadWriteWeb, 4 August, 2010: http://www.readwriteweb.com/archives/google_ceo_schmidt_people_arent_ready_for_the_tech.php ** All too much, The Economist, 25 February, 2010: http://www.economist.com/node/15557421
16 PERSPECTIVE
Organizations have a fragmented response to the challenges. They struggle with their information architecture, they are uncertain about the tools to use, the IT complexity is daunting, expertise is lacking, data stewards are not identified or their roles are not sharply defined and the poor quality of data hampers successful deployment of intelligence. The outcome is a lack of trust in enterprise data. Bringing these initiatives together into a coherent strategy is vital for effective use of corporate information. An A.T. Kearny 2009 IT Innovation and Effectiveness study reported that executives consider two factors – inconsistent data and IT complexity – as the biggest IT-related barriers to company growth. That is hardly surprising. Businesses have been taken by surprise at the sudden growth of enterprise data. They are now searching for solutions.
Structuring Your Information Strategy Organizations need a strategy that ensures information of value is consistently available to the key stakeholders across the enterprise at the right time. An information strategy blueprint must move organizations from a network of silos to an integrated enterprise-wide viewpoint. The goal is to unleash value by leveraging information as a strategic enterprise asset. Wipro’s 4-Layer Information Strategy Blueprint shown below creates a logical approach by addressing: • Information life cycle management including master data, metadata, and unstructured content as the foundation layer of the blueprint. It uses domains of data management, quality management, governance, and archives to ensure high-quality data is available across the enterprise.
PERSPECTIVE 17
• Business intelligence that delivers operational and analytical reports to assist in improving operational efficiencies.
consumer products manufacturer savings of 8% in brand media spend was realized by reallocating media mix through Marketing Mix Modeling solution.
• Business analytics that help organizations plan and optimize business operations.
There are new models of taking effective decisions emerging and organizations having the vision to apply new approaches like predictive analytics stand to gain competitive advantage.
• Enterprise performance management that allows organizations to measure themselves against financials and other strategic parameters. How far can such a strategy guarantee outcomes? A specialty retailer adopted a streamlined information strategy, harvesting new insights that helped increase its market share by 4% and its gross margin by 2.6%. The information caused it to take action that reduced employee churn from 8.1% down to 6.1%. Predictive Analytics The incremental intelligence gathered by organizations traditionally has accelerated at a slower pace than the information growth itself. However, what will help leading organizations to be better prepared for unpredictable scenarios in the era of information explosion is their ability to effectively use predictive analytics. Actionable insights generated through advanced capability will be a key differentiator for continuous sustenance and competitive advantage. Thus, the art of doing business now needs continuous fusion of advanced science. We have observed that using “insights derived from predictive analytics” as a discipline has enabled organizations make more effective decisions. It delivers better outcomes to the bottom line and drives growth. One of the leading fashion retail clients for whom we help analyze the size profiles and optimal composition of size packs for seasonal apparels, was able to reduce lost sales by an average 31%. An early estimate of last year calculations show projected gains of approximate USD 14M through this initiative. In another example for leading
Management Sponsorship Makes the Difference Organizations that create a strategy to leverage data and information rapidly adopt platforms with information management tools. Their mature management approach addresses key non-technical issues like governance and stakeholder management. The impact of top-level executive sponsorship for data and information management cannot be undermined. It can make the difference between successful implementation of the strategy and frustrating failure. Having an astute and dependable information strategy has an impact on a company’s ability to do business better by improving its ability to make smarter decisions at every level.
18 MANUFACTURING
NOT JUST A COG IN THE WHEEL manufacturing excellence is increasingly guided by the usage of advanced analytical tools that also equip firms to deal with challenges related to innovations, green practices and regulatory compliance N S Bala Senior Vice President, Manufacturing & Hi-Tech
Globally, manufacturing has become a key talking point in government and industry circles in view of its critical role in GDP growth and sustainable development strategies. However, the sector’s share of GDP has declined considerably over the years in advanced and emerging economies alike owing to the rapid expansion of the services sector. Today, manufacturing is staging a strong comeback, powered by renewed policy focus, adoption of new technologies and intelligent systems, market demand for higher-end manufactures, and rise of new global markets.
MANUFACTURING 19
Certain manufacturing segments are also adopting a services approach, which is a distinct breakaway from conventional manufacturing delivery. For example, Xerox is leasing out machines as a service. This approach spells a paradigm shift in the manufacturing sector’s strategic thinking. Over the next decade, manufacturing competitiveness, growth and development will increasingly hinge on the quality of inputs that go into the strategic planning. Towards this end, the importance of analytics can hardly be overstated. Analytics locates itself in the data that is related to the particular business. With the adoption of new network and communication technologies, companies have access to several terabytes of data about their customers, suppliers and operations. The amount of data in the
manufacturing world currently tots up to nearly 1 exabyte per year and growing exponentially (compared to nearly 434 petabytes for healthcare and nearly 619 petabytes for banking). However, manufacturing firms in general have not leveraged the growth opportunity that data analytics creates. This is corroborated by a Ventana Research report titled ‘Business Analytics Still in Infancy Stage in Manufacturing’ which states that only 12% of all manufacturing companies function at a high level of maturity in their use of analytics. Nonetheless, availability of large and diverse data provides compelling opportunities for manufacturing companies to identify multiple levers that can help address the sector’s competitiveness, supply chain optimization and sustainable development goals.
20 MANUFACTURING
Key Growth Drivers From the competitiveness perspective, manufacturing analytics will be determined by data on three key areas: new growth centers, demand for high-end manufacturers, and adoption of IT.
• Adoption of IT & Intelligent Devices: In strengthening the knowledge content of manufacturing, major players worldwide are adopting best-of-breed networking, digital and communication technologies, supported with intelligent devices that promote efficient use of resources. Leveraging cloud computing capabilities to drive manufacturing excellence is a case in point.
• New Growth Centers: The current global GDP growth draws its sustenance from the performance of major emerging economies like China, India, Brazil, South Africa and others. The IMF forecasts suggest that investors will continue to invest in emerging markets for some time to come.
To build a robust manufacturing strategy, companies would have to invest heavily in data capture covering the key areas cited, as well as create an analytics organization to turn the data into insights for future action.
• High-end Manufacturing: Cost arbitrage is not the sole reason for global manufacturing firms to locate/ relocate their production units. Instead, key manufacturing players are looking up a variety of solutions to move up the global value chain. The accent on green manufacturing has also redefined the production landscape, with increasing focus on emission standards, carbon footprints, etc. Concomitantly, changing lifestyles are creating demand for new products.
Analytics-Driven Competitiveness The manufacturing sector is inundated with data – both structured and unstructured – such as, text and voice analytics, social media analysis and other predictive and prescriptive techniques. Typically, in the manufacturing space, Big Data (see Diagram below) is generated from machine2machine, people2machine and people2people interfaces. While machine2machine data is generated from devices like sensors, RFID tags, GPS, bar code scanners, and industrial machinery, people2machine
MANUFACTURING 21
data comes from the use of tools like smart cards, gadgets, and cars. Further, people2people interaction through social media generates a welter of domainspecific data and information. Smart manufacturing organizations are embedding analytics to transform this data into insights to improve efficiencies, drive optimization, develop niche insights, use predictive modeling and conduct simulations that support innovations and reduce costs on multiple fronts. They also provide key support for driving R&D innovations. For instance, analytics has helped car manufacturers to move from concept to production stage within a quarter of the time it took earlier. In the pursuit of business value, manufacturing organizations need to know what is happening now, what is likely to happen next and what actions should be taken to get the optimal results. So, modeling and
predicting the future, using structured and unstructured data, assumes key significance. The emerging scenario builds a strong case for the usage of advanced analytics in various forms (like business intelligence, modeling, simulations and predictive analytics). Advanced analytics can positively bring business value in terms of influencing cost, revenue and help firms offer better consumer services/ products (see Diagram above).
advanced analytics: key success factor Advanced analytics demands a high-performance data management infrastructure to handle data integration, statistical analysis, and other compute-intensive functions. There are innumerable examples of manufacturing firms
22 MANUFACTURING
that have profited immensely from usage of advanced analytics. For instance, using advanced analytics, the ‘concept to production’ of a new car is down from 60 months to 18 months. Now, instead of 250 engineers sitting at drafting tables in Flint, Michigan, to design GM’s now-extinct Oldsmobile, car designs come from some 17 different locations around the world. The cycle time to produce a new car is also about a fifth of what it used to be. Whereas it used to take as many as 10 years to get a new car designed and into production, now it’s one to two years. While this is indeed the age of big data, manufacturing analytics can also base itself on a limited range of data, meaning big data is not a sine quo non for manufacturing analytics. Key Challenges The barriers to analytics adoption faced by manufacturing organizations are mostly managerial and cultural rather than related to data and technology. The leading obstacle to its widespread adoption is lack of understanding of how to use analytics to improve the business. With business processes becoming collaborative, enterprises are facing new challenges in uncovering insights locked within their data. Data is stored in silos across the enterprise. There are also question marks on data quality, rendering them unreliable. Data formats
are also not necessarily standardized. Before being able to share and use the data across the enterprise, these hurdles must be crossed through effective integration efforts. Enterprises are eyeing and analyzing the data exhaust of their customers, users, vendors, partners, collaborators, competitors, and their own employees, assets, processes and systems, to get a faster and sharper view of changing business demands. Expectedly, the data types and volumes are so large that it raises questions about the availability of expertise to structure and analyze them. Today, corporations want to analyze data real time, and not just at the end of the quarter. In fact, the need for speed and right timing outweighs even the need for accuracy in many scenarios.
On a larger plane, the challenge would be to build an analytics ecosystem comprising providers of services and information, IT software vendors, offshore analytical outsourcers, data providers, etc. The ecosystem will be guided by an analytics roadmap. The next generation information management system will be predicated to optimization, real-time sensing, pro-active response, and leverage of analytics-driven business. Success in this lies in how simulations support optimization, alerts drive real-time sensing, and early warning predictions enable firms to pro-actively respond to emerging challenges. Conclusion Even as governments and industry around the world take firms steps to raise the manufacturing share of GDP, there is growing awareness of the need to use analytics to achieve competitiveness, cost efficiency and optimum resource utilization. Analytics is helping manufacturing firms to shorten the concept to production cycle and drive innovation. Widespread adoption of analytics will greatly help industry to switch to high-end manufacturing and move up the global value chain.
24 ENERGY
EXPLORING NEW FRONTIERS OF GROWTH oil & gas companies, particularly in the exploration & production (e&p) business, will profit immensely from the usage of integrated upstream data management systems and analytics Anand Padmanabhan Senior Vice President, Energy, Natural Resources & Utilities
ENERGY 25
Rising competitive pressures and new complexities in the global oil & gas sector mandate all key players to sharpen the focus on analytics to obtain fresh insights on the emerging challenges in business. It is important for oil & gas companies to visualize the business information to stay ahead of the growth curve. However, most oil & gas companies, especially those operating in the E&P segment, have met with rather limited success from investing in data management projects. The expectation gap is attributed to several underlying issues, such as, business users not being adequately engaged in analytics and their general distrust of the data in use, traditional databases under-delivering, relatively long gestation of data management projects, inefficient use of metrics, and the business leadership view that investments in data management projects do not deliver the expected ROI. Data management initiatives in oil & gas companies are also constrained by the management and project teams not sharing enough information on analytics with the business users. Consequently, the users tend to view data management and analytics as IT functions in which they have no direct role. They are also circumspect about basing their business decisions on data sets that they do not trust fully. The common view therefore is that the pain of change is greater than the gain of change.
There is a misplaced apprehension among business users toward analytics, since the more traditional databases have typically under-delivered and have been largely inefficient.
Flare-up of Issues In the pursuit of appropriate solutions to the intractable data management issues, companies need to dive deep into the symptoms and causes of the flare-up. At the outset, data management initiatives in the oil & gas sector are perceived to be complex with low visibility for the users. The metrics and milestones pertaining to data management are usually not well-defined, and the course checks are not robust. As a result, the business users fail to see the specific benefits they will derive from the data management initiative. The users often experience an initiative overload characterized by highly complex data that is difficult to deal with. Many a time, accessing the data itself proves difficult, resulting in loss of productivity. Inconsistency in the data captured across databases, lack of a unified system to track, maintain and govern data, and the data itself being siloed in different corporate locations add to the users’ predilection with the usage of analytics for decision making. Among other key issues, the industry has assessed that traditional databases are just not enough to power upstream data management. Data stored in multiple databases with no integration between them is costly to maintain and lacks the ability to impose data standards or governance. With no measurable and/ or visible milestones in place, the data management project will likely under-deliver. The initiative will also end up taking on too many data sets and/or developing architectures in too much detail.
26 ENERGY
Question of Value
Total Approach
The constraints cited above tend to influence the business leadership into believing that data management projects do not deliver value for money. Also, there are usually no metrics in place to measure data management initiative benefits or the attendant risks.
Given the challenges, companies would do well to undertake a deep assessment of the symptoms and root causes of common upstream data management problems. It is important that the business leadership takes ownership of the exercise, and the data management initiative should focus upon highly visible data that most stakeholders see, use, or build on for their workflows. This will attract the interest and participation of all business units in the initiative.
For want of an integrated approach, oil & gas companies deploy high-cost resources to clean up and manage complex data. For instance, geologists and geophysicists are seen to spend inordinately long time on managing seismic data (low value) rather than interpreting seismic data (high value). These companies also spend excess money on seismic data acquisition costs because of poor data management practices (data is lost, stored on local drives, purchased twice, stored in non-digital form, not accessible remotely, etc.). Usage of analytics will help these companies to visualize the prospects more efficiently and tide over the difficulty of there being too few qualified, experienced resources to evaluate reservoir prospects. The prevalence of legacy systems wherein data sources (structured and unstructured) are siloed impedes the companies’ ability to obtain an integrated view of upstream/production data. In such situations, the business users tend to abdicate the responsibility for data and leave the issue of data quality to be addressed by the IT or project teams. Further, in the absence of an integrated upstream data management system, oil & gas companies lean on manual methods that are riddled with challenges. As a case in point, the IT staff evaluating the existing data sets and, while running scripts, are wont to discover data issues such as inconsistencies and lack of standards. When these issues are found, high-cost business experts are brought in to fix the problems. They request for sound data to be further described in spreadsheets and go through the data line by line, reviewing thousands of well, drilling, production or exploration exceptions one by one. This creates a cyclical process whereby three separate data sets are used at the same time (i.e., original scripts derived by IT, spreadsheets which are reviewed and edited by business experts, and new business data generated at source). In the end, thousands of labor hours and dollars are spent on a project yielding results that are not maintainable, repeatable or reproducible.
It is equally important for the company to develop a high-level data architecture with standard common applications and integrated workflows. Use of open architecture and common standards will help companies develop solutions that can be modified as per changing business needs. A sharp governance and architectural vision will help firms in converting the data assets into powerful insights for building business strategies. Conclusion Looking into the future, oil & gas companies that initiate data management projects should carry out a reality check on the barriers to the adoption of analytics. This will greatly help the data management project team to deliver on the expected lines. Given the complexity of oil & gas business and the unique circumstances that dictate decision making, a case-by-case approach will be the preferred approach to analytics.
PERSPECTIVE 27
Predictive Analytics
Future-Proofing Business Predictive analytics helps enterprises ride the crest of a growth wave and deal with difficult market conditions
Nitesh Jain Head of Analytics Center of Excellence Srini Pallia Senior VP & Global Head Business Application Services
28 PERSPECTIVE
precisely than ever before. The report underscores that predictive analytics is all the more useful in difficult market conditions. Users achieved nearly1% improvement in their operating margins year-on-year in a shrinking market, and customer retention rose 6%. In contrast, firms that did not adopt predictive analytics saw a 2% dip in margins and 1% decline in customer retention rate. No wonder more than 30% of the global leaders in varied industries an analytics program today have an enterprise-wide should be designed information strategy to collect and to achieve useful analyze information and use it for predictive analytics. correlations to
Increased competition, globalization and the worldwide recession have decimated corporate profits of many of the world’s largest companies. As these companies plan for a new normal, they will seek to grow at an accelerated pace. Business insights derived from the information across the organization will be a strategic advantage to enable this growth.
According to independent research firms, information generated in enterprises will roughly double almost every 18 months. The world’s total storage of digital data already amounts to more than a trillion derive prescriptive Traditionally, business users have gigabytes. While a good portion relied more on operational and BI of this is being accumulated in value with large reporting backed with intuition and the form of structured data by volumes of data . experience of the business leaders. systems and data warehouses of However, this approach is not organizations, social media invention sufficient when confronting the far-reaching changes is causing exponential growth through textual and other driven by macroeconomic upheaval and the exigencies of unconventional forms of data. networked, collaborative and global business operations. The incremental intelligence gathered by organizations Advanced “predictive” analytics helps these enterprises traditionally has accelerated at a slower pace than the derive deep insights from the vast amounts of information growth itself. However, what will help available data for future decision making. However, not leading organizations to be better prepared for every company has the ability to leverage analytics unpredictable scenarios is their ability to effectively use as part of its strategy. predictive analytics. What’s in store for enterprises?
Art & science of predictive analytics
Using “insights derived from predictive analytics” as a discipline has enabled organizations make more effective decisions. For instance, a leading fashion retail business was able to reduce lost sales by nearly 31% by using predictive tools to analyze the size profiles and optimal composition of size packs for seasonal apparels. The projected gains from this initiative added up to nearly $14mn. In another example, a leading consumer products manufacturer made savings of 8% in brand media spend by reallocating its media mix through a marketing mix modeling solution.
A well executed predictive analytics initiative has to have a combination of good business understanding (art) and strong analytics skills (science). This interplay becomes crucial as organizations seek to leverage rich business insights to optimize operations and capitalize on new sources of revenue, and proactively manage risk while ensuring efficiency. However, globalization, complex networked operations and increased risk, combined with an explosion of information, pose new challenges needing a scientific approach.
Similar findings have been accumulated by a recent Aberdeen Group survey, co-sponsored by TIBCO Spotfire, showing that the top performing 20% of organizations use predictive analytics to predict the future behavior of customers and potential customers, and the internal performance of the organization more
In some instances, the ability to leverage predictive analytics gets severely impacted due to causal factors that behave differently across regions and cannot be standardized. A leading restaurant chain, for instance, realized the inability to forecast the demand for its merchandise as it did not possess adequate business knowledge on the eastern Asia region where it operates.
PERSPECTIVE 29
A fundamental shift to a smarter, intelligent, fact-based enterprise should be part of the information strategy of the organization. Predictive analytics leverages advanced statistical models, leading-edge mathematical assets and methodologies to develop solutions. Converting them into operational reality requires effective alignment of this science with business acumen. In the normal course, the CIO’s office has traditionally been responsible for core ERP, infrastructure and BI applications whereas predictive analytics initiatives are mostly driven by business. In the quest of intuitive insights, the pace and appetite of business users to go ahead by themselves with analytics introduces a latency between business and IT resulting in silos. What is needed is a cohesive effort by all teams to create effective business outcomes - the science coming from CIO organization and the art from business. Orchestrating the change Organizations realize substantial value by orchestrating the information and process environment in a simplified form. A complex information environment will require a unifying information strategy to start the next process of applying analytical methods. The organization should align this in a way that the outcomes will enable realization of organizational and individual functional goals. An analytics program should be designed to achieve useful correlations to derive prescriptive value with large volumes of data. Any analytics program should target to prioritize both ease-of-use of information and process optimization, simultaneously aligning them both to the business objectives. Finally, an effective change program to manage shift from present attitudes and approaches toward information is fundamental to the success of the program. The importance of information strategy and benefits from predictive analytics has to be communicated and demonstrated to business leaders. Conclusion In the emerging global business environment where new growth opportunities are counter-balanced by short business cycles, decision making cannot rest solely on knowledge predicated to historical data and legacy
information systems. Organizations need to look into the future, anticipate key trends and build strategies to stay ahead of the curve. Predictive analytics enables these enterprises to use complex mathematical tools – once the preserve of academics – to break down complex business problems and meet future challenges.
30 HEALTHCARE & LIFE SCIENCES
HEALTH MEANS MORE Healthcare & life sciences (HLS) organizations are leveraging analytics to tackle complex business situations Paul S Coggin Lead Principal Consultant, Advanced Analytics Center of Excellence
Growing health awareness among global societies and the rise of many developing economies have spawned new growth opportunities for healthcare and life sciences (HLS) organizations. The nature of this growth is impacted by key factors including (i) political and public pressure on payers and providers to contribute tangibly to health outcomes and reduce healthcare costs, (ii) patent expirations and the heightened need for innovation, (iii) regulatory changes impacting the interaction between medical firms, patients, providers and payers, (iv) changes to the competitive landscape, improvements in IT and medical communications, (v) the advent of specialized medicines and treatment plans, (vi) growth of patient-centered care, (vii) global sourcing and global customers and patients, and (viii) liability mitigation in view of ever-expanding needs to ensure compliance across multiple customer dimensions.
To meet the varied demands originating from the key constituencies – patients, care providers, payers, intermediaries, and facilities – and to address complex business challenges, HLS organizations are turning to analytics in a big way to gain new insights for appropriate action and interventions. The HLS sector is today awash with rich flow of data from multiple sources like point-of-care encounters, medical claims, pharmacy claims, lab values, genetic markers, biometrics, etc. Analytics helps HLS organizations leverage the vast swathe of data to build strategies on how to penetrate new markets, grow revenues, track competition, drive product and service differentiation, acquire and retain new customers, respond to market dynamics and regulations, and perhaps most importantly, help improve health outcomes.
HEALTHCARE & LIFE SCIENCES 31
32 HEALTHCARE & LIFE SCIENCES
Analytics: The Direction Finder More and more HLS firms are leveraging analytics – people, process and technology – to meet critical business goals embedded in the care continuum. Healthcare enterprises use analytics to address the innovation imperative, enhance product development and demonstration of product value, drive revenue flows, understand and meet customer expectations, and promote customer empowerment. These are elaborated below: • Innovation imperative: Proliferation of HLS products and services and the ensuing global competition mandates HLS firms to step up R&D innovation and product development. Competing firms are called upon to collect and analyze customer and market intelligence to finetune and optimize product lifecycle. • Pressure to demonstrate & defend product value: HLS firms come under great pressure to demonstrate and defend their products in the medical setting. They have not only to obtain US FDA approvals for new products that have been clinically tested but also convince the payers – both government payers and private insurance companies – to include these in their respective product lists. Analytics helps HLS firms to demonstrate the product value with new insights. • Manage revenue & margin pressure: Firms need to develop and leverage business intelligence across all aspects of the business – including manufacturing, supply chain and logistics, field sales, customer acquisition, contracting and revenue management – so as to meet their revenue goals. • Customer empowerment: As HLS firms step up the interaction with customers, its marketing and sales units need to fully understand and leverage interrelationships between health information availability, social media, the customer’s financial situation and healthcare decision making. Analytics gives the firms greater insights on customer preferences and patient experiences, their willingness to pay and other factors that impact spending. • Exception handling & crisis management: From unfortunate events in manufacturing and supply chain to clinical safety and promotional compliance, negative
press reporting catches the eye of regulators, healthcare institutions and patients alike. Robust data capture and analytics helps firms anticipate negative outcomes and take corrective measures. Diagram below illustrates how data and analytics are leveraged to drive total customer experience management and expand medically productive utilization of healthcare resources. Analytics serves as the foundation to understand and serve the needs of healthcare in our resource-finite world.
HEALTHCARE & LIFE SCIENCES 33
Addressable Areas Effective usage of business analytics enables HLS firms to develop the most appropriate solution for each segment of business. For instance, analytics helps firms make a better assessment of patient and physician behavior that may be employed to strengthen customer communications and the overall customer experience. Diagram below illustrates how data analytics and performance management support situation assessment and building of a competitive strategy, leading up to
campaign and account planning and execution. This scenario is common for medical and pharmaceutical manufacturers – on the heels of their latest innovation – trying to “get the word out” in a precise, qualified and impactful manner. Among a variety of other usages, analytics helps firms uncover potential undercurrents in the supply chain as well as identify key channels including GPOs, wholesalers and distributors for the distribution of diverse products and services. Firms are thus better placed to optimize
34 HEALTHCARE & LIFE SCIENCES
their financial metrics and concurrently get close to those that touch the end-customers. Further, analytics drives health outcomes measurement, enabling firms to operate efficiently in a healthcare environment that is increasingly looking to squeeze costs and reimbursements and concurrently improve health outcomes. Need-Based Approach The Healthcare Behavior Impact Pyramid (see the diagram below) indicates that the levels at the base of the pyramid, such as, generalized patient education, patient interventions, and utilization management need less intensive targeting, specificity, coordinated touchpoints and analytics, contrasted with the higher levels that cover personalized doctor interventions, disease category management, health management, and care management, at the apex. For example, generalized patient education (at the bottom) can be met with distribution of pamphlets in large numbers, whereas data used for care management will be
intensive and targeted at a small, defined group. The use of analytics, especially at the apex, can yield significant results in the allocation and management of healthcare resources.
In using business analytics, healthcare organizations naturally take critical steps to ensure data integrity from various interactive channels and verify their accuracy at any given point in time. This is especially important in a regulated industry where the safety and wellbeing of patients is at stake.
In meeting this challenge, organizations acquire or develop the skills required to perform the analytics, which are relatively difficult to find. The analyst skill set must encompass a wide range of areas such as clinical, pharmacy, claims, coding, data warehousing, database and healthcare informatics.
HEALTHCARE & LIFE SCIENCES 35
Use Case Scenario Analytics support HLS firms in a myriad of ways, and the scenarios below are merely several examples: • Scenario 1: Insurance companies which are key payers in the HLS domain use business analytics to conduct due diligence on utilization patterns, clinical outcomes and provider attributes. This data, for example, processed with analytical tools can yield key insights which in turn deliver significant business value. • Scenario 2: Drug manufacturing firms could leverage social analytics as an input to product development strategies and needs for patient and physician communications. However, the obligation to act on all information or suggestions gleaned from unstructured and informal social media sources, given certain legally binding provisions, could dissuade such firms from leveraging social data. • Scenario 3: Drug companies conduct clinical trials by closely monitoring provider and patient data and ensuring protocol adherence. Post-US FDA approval, however (when the product becomes available to all doctors, not just those conducting the trials for the drug), the key safety challenge is how to quickly and effectively familiarize tens of thousands of doctors regarding the drug and its appropriate treatment protocol. At launch, when physician experience with the product is limited, there is always a risk of inappropriate use of the drug, which could have serious ramifications for the medical manufacturer. If a product is sidelined (or withdrawn) because an inappropriate use has caused a safety concern, this will negatively affect all patients who would have benefited from its use. Predictive analytics helps drug companies anticipate these likely situations – and helps them build product development and marketing communications strategies accordingly. Conclusion Medical and healthcare regulatory norms are tightening worldwide, and all key stakeholders including payers, providers and patients are becoming more discerning with regard to the quality of products and services delivered. As a result the HLS sector is increasingly leaning on analytics to find solutions for improving the delivery of healthcare – and improving the business of healthcare. Usage of business analytics in this sector is still at a
nascent stage. As new data sources like social media and mobility gain further traction, the scope of analytics in this domain will broaden further and impel a greater number of players to adopt analytical tools to support their decision making.
We hope you enjoyed reading “WINSIGHTS” If you would like to read more, please visit our website www.wipro.com/insights — where we regularly publish our viewpoints and perspectives that can help companies sustain competitive advantage. We would love to hear your thoughts and suggestions that could go a long way in making this journal a valuable knowledge sharing tool for Senior Executives like yourself. Please write to us at wipro.insights@wipro.com Best Wishes, Karthik Nagendra Wipro Council for Industry Research
About Wipro Technologies Wipro Technologies, the global IT business of Wipro Limited (NYSE:WIT) is a leading Information Technology, Consulting and Outsourcing company, that delivers solutions to enable its clients do business better. Wipro Technologies delivers winning business outcomes through its deep industry experience and a 360 degree view of “Business through Technology” – helping clients create successful and adaptive businesses. A company recognised globally for its comprehensive portfolio of services, a practitioner’s approach to delivering innovation and an organization wide commitment to sustainability, Wipro Technologies has 120,000 employees and clients across 54 countries. For more information, please visit www.wipro.com