Getting Serious About Analytics: Preferred Letter Alternate ......to optimize their core processes....

Alternate Letter Brochure Title



Preferred Letter Brochure Title

Getting Serious About Analytics: Better Insights, Better Decisions, Better Outcomes By Brian McCarthy and Jeanne Harris

2 | Getting Serious About Analytics: Better Insights, Better Decisions, Better Outcomes

Many companies have analytical capabilities in a few pockets of their organization. The ambitious ones are incorporating analytics more broadly. They’re redesigning how analytics and fact-based insights get embedded in key processes, leading to smarter decisions. The goal, of course, is to produce better business outcomes.

The ability to use quantitative data to shape decisions and outcomes has become a key source of competitive advantage over the past decade. With information technology practically ubiquitous, firms of any size can harness data to get smarter about customer behavior, the supply chain, product development, talent management, and other areas of business.

Data is proliferating in volume and type, including video, audio, and web data that wasn’t readily extracted even five years ago. IDC estimates that the digital universe is doubling in size every 18 months.1 But for most companies, data remains an underused and underappreciated asset. Recent Accenture research, based on a survey of more than 600 C-level and other senior executives of large companies in the U.S. and U.K., describes the extent and contours of the problem.

Over the past few years, many companies have been investing in reporting and business intelligence technology solutions to improve decision-making. Accenture

research indicates that eight out of ten of those companies are not achieving desired goals, largely because they have not developed an analytical capability to manage the vast quantity of information available.

Worse, only one in 12 of the respondents had achieved their anticipated return on investments for these technology solutions. Fully two-thirds of U.S. companies surveyed acknowledge that they need to improve their analytical capabilities.

By contrast, a small but growing number of companies have developed advanced analytical capabilities across their organization. By analytics, we mean an integrated framework that employs quantitative methods to derive actionable insights from data, then uses those insights to shape business decisions and, ultimately, to improve outcomes (Figure 1.)

Accenture research confirms that high-performance businesses— those that substantially outperform

competitors over the long term and across economic, industry, and leadership cycles—are five times more likely to use analytics strategically compared with low performers.2

High-performing companies like Procter & Gamble and Tesco, for instance, have made the discipline of analytics central to the execution of their strategy. Non-profit organizations can also parlay analytics into vital public information. Google.org uses search engine results to predict flu outbreaks by county across the U.S. and provides that information to the federal Centers for Disease Control, which can dispatch flu vaccines to those areas.

Most high-performers are using analytics to optimize their core processes. Scheduling analytics at Cemex, for instance, helps the firm to deliver cement to construction sites within a specified 15-minute window. That allows Cemex to charge premium pricing in a commodity market.

3

Data, data everywhere, but to what purpose?

Figure 1. The analytics value chain

Data Methods Decisions

Tools &Technology

People, Leadership & Organization

• Insights driving processes and actions to impact results

• Access to and extraction of data• Moving from data to insight

Focus area for most enterprisesand service providers today

Process

Outcomes

High performers do it differentlyAnalytics have other uses as well. It can integrate datasets from separate functions in order to balance apparently competing objectives. For example, a procurement employee may be pushing for early payments to get discounts from suppliers. But a finance colleague wants to hang on to cash as long as possible. With an analytics capability in place, a company can find the precise point at which the benefits on both sides of the equation are balanced.

Risk mitigation is another area that benefits from cross-functional analytics. A manufacturer that wants an early warning sign of problems on the factory floor can look at warranty transaction data—how many customers are returning products and for what reasons. A sudden increase in warranty repairs for a particular part should trigger a timely response from the engineers, designers, and suppliers overseeing that part.

The discipline of analytics is well suited to a multi-polar world, where different regions are growing at different paces, and have quite different regulatory and political risks. Managing a multinational portfolio of businesses requires executives to anticipate changing supply and demand conditions in many different markets at once. This means developing data-based insights that answer the question “What are our best next steps?”

3


Figure 2. Closed loop decision making process

Technology Enablers

Marketing

BI and Packaged Workbench

Root Cause AnalysisTools

Statistical ModelsOptimization Tools

Supply Chain HR Manufacturing Finance

Performance Monitoring

Questions on Key Metrics

Value Realization

Insight Validation

InsightGeneration

Weekly Forecast Adjustment Monthly Business Review Quarterly Health Check

Execution

Core Analytics Team

Functional Analytics

Cross-Functional Analytics Team

Analytics Center of Excellence

How to make better decisions

Some high-performing companies are establishing real differentiation in analytics by deliberately closing the decision process loop as shown in Figure 2. They turn raw data into insight, and use the insight to shape business processes, which then generate better decisions. Unless managers follow through with the execution steps, insights by themselves have little business value and are merely nice to know.

Redesigning the decision process involves taking each process relevant to solving a problem, embedding analytics in the processes, and linking the reengineered processes more tightly together. In most organizations, the decision process has three components:

•Choosingtherightmetrics—thosethatwill help to improve execution of the strategy

•Generatingandvalidatinginsightsthatcan be acted on by all the relevant units of the organization

•Allocatingresourcestoturninsightsintospecific actions

The next step is a checkpoint to determine the effect of the decision: Did the organization achieve the right outcomes? If not, where did something go wrong in one of the three components? It’s essential to close the loop this way, because if you can quickly discern where the problem occurred, you can make rapid adjustments by changing a key metric, an analytical method, or the deployment of resources.

Analytics become even more powerful with a cross-functional approach, since most business problems touch multiple areas of a company. At a health payer, for example, the traditional batch claim processing

provides poor customer service, increases administrative costs, and drives poor cash management. A better solution, real-time adjudication, is complex enough that it requires analytics and collaboration among several areas—from prioritizing transactions, to retraining technicians, to shifting resources away from adjustments and appeals and toward customer service at the provider’s office.

Building an advanced analytical capability to support improved decision-making is not easy. Companies may struggle to generate insights from their technology investments, connect the insights to the relevant processes, and then link them to tangible business outcomes. While each company has its own unique set of challenges, decision process optimization tends to run up against several common problem areas:

5

Focusing on the wrong metrics or too many metrics. Most firms establish a large set of metrics, but they often lack a causal mapping of the key drivers of their business, which a small set of metrics should track.

Inability to validate insights. Data analysis usually generates many possible insights. What’s tricky is to validate them across functions, in order to identify the most useful insights. There is no formula for validation, but rather a review of the statistical analysis through the filter of management experience. A business-to-business manufacturing company, for instance, was preparing to launch a new product. Detailed customer research showed a shift in preferences among small to medium-size businesses toward preparing more of their own products for export. Also, recent cuts in staff meant that the end users put more value on the product’s ease of application. Each of these insights were double-checked and deemed important enough to reformulate the product.

Faulty execution. Once a key insight has been selected, there are many possible actions to take. Should you change pricing, or reconfigure the sales force, or adjust the supply chain? The most common failure in this regard is that managers choose not to make a decision at all, fearing the implications of making a wrong decision. Yet choosing the right action in a timely way is essential to success. Accenture research shows that one attribute shared by high-performing companies is the speed with which managers make decisions, typically in close proximity to their customers. High performers get the right information into the hands of the right people who can act quickly at a local level.3 That reinforces the need for analytical capabilities close to the customer— which in practice means distributed in many parts of an organization, if not throughout the entire enterprise.

Cultural resistance. Many managers, while reluctant to say so, rely primarily on intuition and experience rather than fact-based analysis. A recent Accenture survey found that 40 percent of business decisions are still made based on judgment alone, partly because of the absence of good data.4 While experience and intuition are valuable assets, they remain limited until combined with relevant data.

Executives and managers thus must increasingly be fluent with analytics. They should understand the models underlying decisions, as well as the assumptions behind the models. And at all levels of the organization, the structure of incentives and rewards should encourage people to use analytics in their day-to-day business processes.


From ‘so what?’ to ‘now what?’As managers become more fluent in analytics and rigorous decision making, they can address progressively more sophisticated questions. (See Figure 3.) They can more accurately predict what customers will buy, where supply chains could break down, and which employee will become a top performer.

UBS Investment Research, for example, has started incorporating proprietary analysis of satellite images of the parking lots of Wal-Mart into its earnings estimates. UBS bought its satellite data and analysis from a startup called Remote Sensing Metrics LLC, which built a model for how customer flow correlates to quarterly earnings. By counting the cars in Wal-Mart’s parking lots month in and month out, Remote Sensing Metrics analysts were able to predict company’s customer flow. From there, they worked up a mathematical

regression to come up with a more accurate prediction of the company’s quarterly revenue each month.5

Similarly, a major U.S. music distributor used predictive analytics to address a sudden spike in demand for the CDs of one of the artists in its back catalog. How could it ramp up production to meet immediate needs without creating excess inventory in the future? The company’s analytics engine used data from the supply chain, finance, and the internal customer relationship management system. In short order, the music distributor had pinpointed the source of greatest demand, and could make timely decisions about where to boost production and where the most cost-effective and profitable locations were to locate inventory.

Every corner of the business stands to benefit from predictive analytics and more informed decisions. So how can companies achieve that level of sophistication?

The route to building an analytical capability that can improve decision making will depend on the level of analytical maturity currently within the organization. A consumer packaged goods company accustomed to innovating through market basket analytics will have a different set of issues, challenges, and questions than will a bank that may not even know its credit exposure on a daily basis. An electric utility accustomed to doing two physical meter readings per year will likely not be prepared to take advantage of the rollout of a smart grid that allows for several meter reads per hour. Therefore, a useful first step is a diagnostic to determine the company’s current maturity and where the gaps lie, as shown in Figure 4.

Analytics novices, those at Stage 1, probably should focus on improving the quality of data and technical tools. Poor quality of underlying data remains a major problem worldwide, and investing in analytics while the underlying data remains dirty is a

Figure 3. Moving up the analytical curve2

Competitive Advantage

Sophistication of Intelligence

Optimization What is the best that can happen?

Predictive Modeling What will happen next?

Forecasting/extrapolation What if these trends continue?

Statistical Analysis Why is this happening?

Alerts What actions are needed?

Query/drill down What exactly is the problem?

Ad Hoc Reports How many, how often, where?

Standard Reports What happened?

PredictiveAnalytics

DescriptiveAnalytics

Changing the Questions

Now What?

So What?

What?

7

An initial diagnostic

waste of money. In these circumstances, it’s essential to determine what is the highest priority data for executing the core strategy, and then to validate, clean, and consolidate that data. Stage 1 or 2 firms also tend to lack people with advanced analytical skills, the specialists who can make a huge difference. These firms should set about hiring selectively or exploring how to outsource at least some of their analytics function.

More advanced companies have already improved the quality of data, brought analytics specialists on board, and incorporated analytics into decision processes. But even these firms have mostly concentrated on functional

point solutions. Few have moved to the next level by joining up these capabilities—embedding analytics in all of their business and decision processes. For them, the greatest gains now will come from expanding analytics across functions wherever possible, aligning specialists to the highest-value projects and the right roles. (For a case study, see the sidebar, “Cross-Functional Analytics: A Retailer Woos Back Lapsed Customers.”)

Figure 4. The stages of analytical maturity2,6

Stages

People

Process

Technology

Organization

Stage 1Analytical NovicesAnalytical skills do not exist

Analytical process does not exist

Missing/poor quality data, multiple defines, un-integrated systems

Limited insight into customers, markets and competitors

Stage 2Localized AnalyticsPockets of isolated analysts (may be in Finance, SCM or Marketing/CRM, etc.)

Disconnected, very narrow focus

Recent transaction data un-integrated, missing important information. Isolated BI/analytic efforts

Autonomous activity builds experience and confidence using analytics; creates new analytically based insights

Stage 3Analytical AspirationsAnalysts in multiple areas of business but with limited interaction

Mostly separate analytics processes. Building enterprise level plan

Proliferation of business intelligence (BI) tools. Data marts/ data warehouse established/expands

Coordinated; establish enterprise performance metrics, build analyt-ically based insights

Stage 4Analytical CompaniesSkills exist, but often not aligned to right level/right role

Some embedded analytics processes

High quality data. Have an enterprise BI plan/strategy, IT processes, and governance principles in place

Change program to develop integrated analytical processes and applications and build analytical capabilities

Stage 5Analytical CompetitorsHighly skilled, leveraged, mobilized, centralized, out-sourced grunt work

Fully embedded and much more integrated analytics processes

Enterprise-wide BI/BA architecture largely implemented

Deep strategic insights, continuous renewal and improvement


Cross-functional analytics: A retailer woos back lapsed customers Most business problems touch multiple parts of a company, so an analytically based solution will be most effective when it involves each of the relevant units.

The case of a major retail chain illustrates the point. The retailer had a very high market share in the U.S. and faced slowing growth. More than one-third of customers were lapsing, that is, after shopping at the chain they did not return for more than one year. Customer surveys uncovered four main reasons for this situation:

• The product often was not on the shelf in stores.

• Sales staff often was not available to help the customer.

• The product was not assorted, bundled, or priced appropriately.

• The overall experience and interaction with employees was poor.

With the survey results pointing the way, the retailer built an analytics program and reengineered its decision processes to address key metrics around each of the four issues.

Looking at product availability, the metric tended to be fairly static at 95 percent on average. Digging deeper revealed a more volatile picture: Slow-moving products were always available but fast-moving products had just 30 percent availability. Further analysis by product, geographic region, and channel resulted in changes to demand planning processes. The firm examined fast-moving products according to geographic market, using web analytics to assess what customers were looking for on a weekly basis, in order to better predict demand.To get a handle on labor scheduling, the retailer analyzed sales by customer segment, store, and hour of the day.

That revealed patterns such as the fact that affluent males tended to shop on Tuesday, Wednesday, or Thursday nights between 6 and 9 p.m. During those times, the retailer would schedule more staff for certain sections of the stores. Staffing was tweaked for eight segments and also by season.

For pricing and assortment issues, the retailer combined assortment and space analytics to look at the effectiveness of product adjacencies and the overall shopping basket. Purchase path analytics were deployed to understand the next likely purchase by each demographic segment and the time lag between those purchases. The team also analyzed profit margin leakage. For instance, analysis of the statistical distribution of cashier price overrides looked for the exceptions that were more than a reasonable tolerance from the mean. These cases resulted from either fraud (cashiers giving away the product to friends) or poor training (cashiers were too customer-friendly, to the detriment of the business).

Unacceptably high employee turnover signaled a lack of employee engagement, which degraded the overall customer experience. Reducing turnover was a difficult challenge across a large employee base. So the team broke down the analysis down by key role and examined a number of attributes by role, such as recruiting costs, training costs, sales productivity, tenure, and supervisory relationship. They calculated a “time to break even” measure by role that incorporated the costs and the sales generated by employee as well as the capital cost hurdle for making that investment. Two of the most skilled roles in the stores were turning over faster than they were breaking even, and the economic impact for these two roles made

up 60 percent of the total impact of employee turnover on customer retention and financial results.

What’s instructive about this case is how the retailer successfully navigated the common obstacles in optimizing key daily, weekly and monthly decision processes. First, it focused on the metrics that really mattered, such as availability of fast-moving products. Second, it identified the most useful insights by probing deep into the customer’s experience and then validated them through other operational data. By narrowing the turnover problem to two key roles, the retailer could be very targeted in its remedy. And third, the retailer winnowed possible actions to a few that would make the greatest impact: training on discount policies, and firing for fraud. The company embedded actionable insights into key decision processes on a systematic basis.

Multi-function analytics have given the retailer a fact-based and feasible program to improve the retention, shopping frequency and growth of customers. By highlighting the areas of greatest significance, analytics supplement management intuition about where the opportunity and solution might lie and where scarce resources should be focused. Most importantly, this comprehensive approach has delivered the desired outcomes, achieving a 20 percent improvement in customer retention that translates into a 10 percent improvement in annual operating income.

9

Becoming thoroughly analyticalWith the results of the diagnostic at hand, companies can set about establishing a basic, robust, or truly advanced analytical capability to improve decision-making processes. Our observations of high-performing companies suggest that effective analytics are built on a three-part foundation: disciplined processes to ensure that valuable insights and recommendations are generated, acted on, and their effectiveness measured; the right people with the right skills, organized in the right way to put insights to work; and technology that ensures data integrity, quality, and accessibility. For most companies, technology has received the lion’s share of attention, with the process and people aspects getting short shrift. Let’s examine each in turn.

Disciplined, repeatable analytical processesHigh-performing companies make analysis an integral part of everyday business processes—the methods by which work gets done and value gets created. In Accenture’s recent survey, fewer than one in five respondents report that their company uses a repeatable analytical approach. Developing a repeatable decision-making process that leverages data and analytical methods thus should be a high priority for most companies.

It’s important to consider and include all the functions or stakeholders that need to be involved in order to mitigate risks and make the greatest positive impact. For instance, if a retailer sees an opportunity around an ad promotion for a “smart” kitchen appliance, the retailer will want to consider whether it has enough components in place on the shelf; enough sales people trained in this

relatively complex product and staffed at the right stores at the right times; and enough service staff trained to handle customers’ follow-up questions. Without the active participation of the people who run those functions, the retailer risks not just limiting sales but also tarnishing its brand.

Keep in mind that the power of analytics derives from making connections—recognizing patterns in customer demand or business activities, isolating the drivers of performance, and anticipating the effects of decisions. To make connections, you have to look beyond the immediate task and appreciate what happens upstream and downstream. Consider the challenge of improving the return on advertising spend. The solution will be most compelling when spend can be optimized across different channels, geographies, and the full range of products.


The right people, properly ledHaving the right people focused on the right set of problems is one of the most important components of developing an effective analytical capability. Taking a broad analytical approach to business calls for big changes in culture, process, behavior, and skills for many employees.

High-performing firms typically have their analytics capabilities centrally managed by one accountable leader. Likely candidates for this role are the CFO, CMO, COO, or CIO, but the title is less important than personal characteristics, including a strong analytical acumen and a bent to use analytical information to guide strategic decisions.

Analytical staff should be recruited and organized based on skill sets and the type of analysis they will complete. Companies will need a mix of analytical professionals to know how to find the relevant data and perform the modeling, but also analytical amateurs who can apply the models throughout the business. Finding people who have the requisite technical skills and also some combination of coaching, consulting, and business design skills represents a challenge. Companies will likely have to hire in order to scale up their analytical capabilities. Retooling, say, accountants to become versed in advanced financial analytics rarely works.

TechnologyWe don’t mean to give technology short shrift; it’s just that the decisions tend to be more straightforward and widely discussed. The type of tools that should be deployed across the organization will depend on the type of analysis that needs to be completed, the skill-sets of the analyst, and data availability and quality.

Technology should be fit for purpose, meaning that machines may be reallocated as needed if the analytics call for varying the volume or location of production. To assess the needs of an organization, identify and categorize tools based on basic, intermediate, and advanced analytical capabilities, the spectrum in Figure 4 can serve as a guide.

It’s usually most effective to embed analytics through the technologies that employees routinely use, on top of an industrial-strength architecture, rather than through special, standalone applications. Analytics can be woven into decision-making processes in several ways:

Strategic decision-making. To get the most out of analytics, it should be an integral part of high-value, low-frequency strategic decisions such as a product launch or acquisition.

Applications for tactical and operational decision-making. In the tactical realm, managers rely on analytical applications that are integrated directly into web applications or enterprise systems for tasks such as supply chain optimization, sales forecasting, replenishment, and advertising effectiveness. For operations, planning and “what if” applications can incorporate nearly real-time information and multiple models to reach an optimal solution.

Automated decision applications. These applications sense online data or conditions, apply codified knowledge or logic, and make decisions—all with minimal human intervention except when the system flags exceptions that require human judgment. Automated decision-making is being applied in a variety of settings, from labor scheduling in retail stores, to monitoring oil wells for preventive maintenance.

Applications to improve innovation. These more advanced analytics include prediction markets, text mining of unstructured data such as SEC filings of competitors, and web traffic monitoring to predict litigation action or retail consumer behavior.

The proliferation of data and more powerful computing technologies to crunch the data are propelling analytics to a more prominent role in business. But technology is just part of the story. Data becomes valuable only after it’s shaped into insights, and when those insights inform the key decision processes that lead to better outcomes.

The end game should be an enterprise analytics capability, where the piece parts combine to solve problems. No doubt this may require more effort at first, more sponsorship from the senior ranks, and buy-in from more people. But enterprise-scale results—in revenue growth, profitability, return on capital, customer loyalty, or other measures of value—make the effort worthwhile. Our experience suggests that annual operating profit improvements worth 1-2% of revenue are well within reach. That’s an outcome to rally behind.

11

References1 “The Diverse and Exploding Digital Universe,” IDC, May 2009.

2 Thomas H. Davenport and Jeanne G. Harris. “Competing on Analytics: The New Science of Winning” (Boston: Harvard Business School Press) 2007, pp 46-47.

3 “Managing in Uncertain Times: Strategies and Practices for High Performance,” Accenture, 2008.

4 Source: “Competing Through Business Analytics to Achieve High Performance,” Accenture Information Management Services, December 2008.

5 New Big Brother: Market-Moving Satellite Images”, cnbc.com, August 16, 2010.

6 Thomas Davenport, Jeanne Harris, and Robert Morison, “Analytics at Work: Smarter Decisions, Better Results,” Harvard Business School Press, January 2010

Barcode Placement

About AccentureAccenture is a global management consulting, technology services and outsourcing company, with approximately 257,000 people serving clients in more than 120 countries. Combining unparalleled experience, comprehensive capabilities across all industries and business functions, and extensive research on the world’s most successful companies, Accenture collaborates with clients to help them become high-performance businesses and governments. The company generated net revenues of US$27.9 billion for the fiscal year ended Aug. 31, 2012. Its home page is www.accenture.com.

About Accenture AnalyticsAccenture Analytics delivers insight- driven outcomes at scale to help organizations improve performance. Our extensive capabilities range from accessing and reporting on data to advanced mathematical modeling, forecasting and sophisticated statistical analysis. We draw on over 12,000 professionals with deep functional, business process and technical experience to develop innovative consulting and outsourcing services for our clients in the health, public service and private sectors. For more information about Accenture Analytics, visit www.accenture.com/analytics.

AcknowledgementThe authors would like to thank Tiffany Brown for her leadership of the Enterprise Analytics research cited in this paper, and acknowledge the contributions of George Marcotte and Chris Yager to the research effort.

About the authorsBrian McCarthy is the executive director, Strategy, for Accenture Analytics. Mr. McCarthy has extensive experience in value-based performance management and analytics engagements with clients across multiple industries. His primary focus is helping organizations address key performance management and analytics challenges and align around increasing shareholder and stakeholder value. He currently leads the Enterprise Management Analytics offering and has overall responsibility for the Accenture Analytics innovation agenda.

Jeanne Harris is an executive research fellow and director of research at the Accenture Institute for High Performance, and co-author of the books Analytics at Work: Smarter Decisions, Better Results and Competing on Analytics: The New Science of Winning. During more than 30 years at Accenture, Jeanne has worked extensively with clients seeking to improve their managerial information, decision-making, analytical and knowledge management capabilities.

Copyright © 2012 Accenture All rights reserved.

Accenture, its logo, and High Performance Delivered are trademarks of Accenture.

Getting Serious About Analytics: Preferred Letter Alternate ......to optimize their core processes....

Documents

Transcript of Getting Serious About Analytics: Preferred Letter Alternate ......to optimize their core processes....