Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities...

20
Unlocking the Value of Analytics Accenture’s Digitally Enabled Grid program

Transcript of Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities...

Page 1: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

Unlocking the Value of AnalyticsAccenture’s Digitally Enabled Grid program

Page 2: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

2

Contents

Introduction 3

Opportunities for utilities to improve performance 4

Lessons learned from other industries 5

Analytics adoption to date 6

Analytics opportunities for high performance 6

Becoming an information-driven organization 9

How analytics drive real value 10

1. Analytics can radically improve decision making at every level. 12

2. Real insight comes from business intelligence and analytics that correlate data 13 across the enterprise and externally.

3. New insights are possible by moving from business intelligence to comprehensive 13 analytics to fundamentally manage business differently.

4. Achieving sustainable value from analytics requires new approaches to 14 information management.

5. Changes to the operating model are required to enable benefit realization 16 from new insights.

6. Rapid learning improves models and increases value from insights. 17

Conclusion 18

Page 3: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

3

Utilities face significant challenges to meet their commitments to shareholders and regulators, while safely managing the health and reliability of the network. Utility executives continue to make important promises to shareholders to deliver earnings growth. In most regions, according to Accenture estimates, 30 to 40 percent of the share price of top-performing utilities is based on expectations for future earnings growth. Yet these growth goals have become significantly more difficult to achieve due to lack of load growth, rising operating costs and the volatility of energy supply markets. Over the next five years, disruptive changes are expected to put additional pressure on utilities’ economic models, including increasing focus on energy efficiency, the growth of demand-response programs and the impacts from increasing adoption of distributed generation.

In response, many utilities are deploying smart technologies to enable new capabilities and extract greater value from their assets. Utilities have learned from their own experience and lessons from other industries that delivering value from these new capabilities requires fundamental changes to the business and operating practices. Many of these changes can only be achieved by deriving critical insights from the voluminous data captured by smart devices and correlating with data in current operational systems and with external sources. Considering the potential value of this data, using advanced analytics to help drive these operational changes has become a critical imperative for most utility executives.

This paper provides insights and examples of how utilities can use advanced analytics to transform operating practices to achieve better operating results. It considers areas where utilities can focus their efforts on ways they can become more effective, information-driven organizations. It also offers a view into industries that have experienced, survived and thrived in times of disruptive change.

Accenture’s analysis suggests the potential value of using smart grid analytics to transform operating results, at a conservative estimate, could approach $40 to $70 per electric meter per year. While value estimates vary geographically, Accenture believes that the potential for value is compelling leaders in all markets to take steps to industrialize the discipline of driving operating change through analytics to an organization-wide scale, including integrating enterprise analytics with operational analytics. Successful companies will achieve a cultural shift to foster an information-driven organization that uses data-driven decision making to achieve high performance.

Introduction

Page 4: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

4

Opportunities for utilities to improve performance

Page 5: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

5

Lessons learned from other industriesTransmission and distribution utilities continue to deploy advanced technologies into their networks—a fact that has been true for decades. However, the wave of technology changes affecting utilities—not just advanced meter systems, but also intelligent devices, advanced mobility, smart consumer devices and social media—is driving much greater impacts. More than just providing opportunities to improve existing processes, these technology changes are providing new capabilities that will change how processes are executed. One example is the detection of and response to outages, which can be transformed when network diagnosis and response activities can be more closely tied through interactions with consumers via customer technologies.

Other industries such as telecommunications, retail and banking have experienced similar disruption, and the companies that thrived invariably embedded analytics into business processes in a robust, industrialized way. For example, traditional wireline telecommunication companies (telecoms) were blindsided by the rapid emergence of competitors using Internet protocol (IP) telephony, while at the same time, wireless was displacing the use of local lines. Successful companies used analytic models to examine and change almost everything from where to invest in packet-switched technology based on call routing to what to charge consumers for new product bundles that emerged. Had these landline companies not made these changes, their survival would have been in jeopardy.

In the area of fraud detection, banks have developed some of the most robust algorithms and processes to identify and prevent credit card fraud. Today, an ordinary practice involves identifying suspicious activity in real time and sending a text message to the consumer

notifying them of the charge before it is approved. Revenue departments at numerous government revenue agencies have employed data-driven insights to increase identification of noncompliance and increase collections.

The Moneyball analogy (see sidebar below) is relevant to illustrate the opportunity utilities have in facing their economic challenges. If “salary” is replaced with capital and “wins” replaced with system reliability, then it replicates one of the big challenges facing utility executives: how to achieve greater system reliability and asset health using less capital. This trade-off is generally defined by the “efficient frontier,” which is that most efficient combination of the amount of money spent and how much reliability can be achieved. Like Billy Beane, Accenture believes that utilities can move to a new frontier and achieve the same or greater reliability using less capital by changing the way they allocate capital using more granular data and insights about the network and statistical models that assess asset outcomes.

Analytics and the Moneyball example1

One ideal illustration of the transformational potential of analytics actually comes from the world of sports, more specifically, US Major League Baseball. In his book, Moneyball, Michael Lewis describes the unorthodox methods employed by Billy Beane, general manager of the Oakland Athletics. Beane had the daunting task of determining how to beat other baseball teams that were spending more than three times what his team was able to afford in salaries. Using analytics, the results he achieved were stunning and fundamentally changed how baseball teams are managed.

Traditionally, baseball team management and scouts relied on their extensive experience to evaluate players. They loaded their rosters with expensive players who looked good on paper (hits, batting average), but whose performance showed little correlation with actual results, in this case, wins. Beane used analytics to turn this thinking on its head, by tying every decision about players and playing to statistical models. In 2002, the New York Yankees paid $1.2 million for each game they won (dividing total games won by total salary). That same year, Beane and the Oakland Athletics paid just $273,000, or

4.4 times less per game won, and they won more games. Two years later, the Boston Red Sox adopted Beane’s methods and won the World Series championship for the first time in 87 years.

Page 6: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

6

Analytics adoption to dateUtilities currently make decisions based on extensive experience and information that lives in “silos” in different parts of the organization. Much of that data is of questionable quality and requires a high degree of manual manipulation, often leading to debate among leaders about the relevancy of data, rather than the actual significant decisions. Siloed data leads to difficulty in understanding cross-organizational impact. Decisions made using operational data can have unexpected consequences if there is little correlation between these separate types of data—consequences that are sometimes only realized after decisions are made and implemented. For example: a decision to defer replacement of an asset could result in increased costs as a result of increased failure rates. Another example: decisions to defer operations and maintenance (O&M) spend to adjust placement of capacitor banks, which could defer potential benefits in terms of voltage optimization. In many cases, extensive time is required to perform cross-organizational analysis, limiting these exercises to once or maybe twice a year, which inherently limits their value.

Conversely, an information-driven utility, one that can aggregate and correlate information across these data silos, would be able to derive information-driven insights that could potentially fundamentally change decision models, performance insights and process changes.

Accenture defines an information-driven utility as having the following characteristics:

•Systematicallyharveststheflowofinformation and knowledge, and provides it to the appropriate people at the appropriate time to help make optimal decisions to drive value

• Institutionalizesandindustrializescollective knowledge of the enterprise

•Architectsdatamodelstoleverageallavailable data (internal and external)

•Organizesandempowerspeopletomakedecisions based on the use of information models

These characteristics can help utilities achieve sustainable improvements to their business results, including:

•Enhancingreturntoshareholdersbymoreeffectively modeling capital and operating spend within a rate case period.

•Providingbetterreturnoninvestedcapital from better allocation of capital to projects.

• Increasingcontributionmarginonpowerdelivered by better optimizing the grid operations.

• Improvingrevenueonpowerdeliveredbyreducing “leakage” from theft and errors.

The adoption of analytics in the utilities industry has not been widespread compared to other industries. To date, most of the focus has been in the customer/retail areas, which has primarily been driven through competition. According to recent research on the utilities analytics landscape, conducted forAccentureby10EQS,oneformerCEOofan integrated electricity company summed up the situation for many utilities this way: “From a retail perspective, we’re realizing less than 10 percent of the potential of [advanced] analytics, however, we’re around 40 percent of the way to understand what we need to do to get there. In supply and transmission [i.e., networks] we’re realizing less than 5 percent of the potential, but we don’t really know how to get to the next level.”

Accenture estimates the current use of analytics across the various operational areas within a utility as shown in Table 1.

Gas utilities need to overcome critical infrastructure integrity issues and satisfy the need to engage with consumers in a more meaningful way in what is, for some, an increasingly competitive retail market through the continued adoption of advanced analytics. In water utilities, analytics adoption is driven by a need to improve asset management and address nonrevenue water challenges where there is heightened focus on resource management. Based on internal and external research, most gas and water utilities (with a few notable exceptions) are still far from using analytics extensively, even though doing so could help address a number of critical priorities.

Analytics opportunities for high performanceWith the widespread adoption of advanced metering infrastructure (AMI) and other sensors, electric utilities have begun to produce big data, or data at the scale necessary to leverage analytics. This production of big data is mostly applicable for distribution (networks) and retail and, to a lesser extent, for power generation.

While the application of analytics is not yet mainstream, estimates of planned investments show that utilities are clearly recognizing the potential of analytics. Pike Research estimates the compound annual growth in spending on analytics at 24.5 percent for the years 2012 to 2020, with a steady-state growth and total spending of more than $34 billion during this time period.2 GTM Research expects aggregate expenditure on automated metering analytics to reach an annual spend of $9.7 billion by 2020.3

Table 1. Estimated current use of analytics by operational areas.

Utility Operations Areas Use of AnalyticsType of Utility

Water Networks 5%

Networks 5%

Networks 5%Electricity Supply 5%

Supply 5%

Supply 5%

Gas

Retail 20%

Retail 10%

Retail 5%

Page 7: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

7

When asked where they see analytics as having the greatest impact for smart solutions deployment, respondents to Accenture’s recent executive survey note that there is broad support for analytics across all areas. Of note, however, is that the majority of executives point to analytics related to better management of the network itself. Perhaps not surprisingly, North American respondents tend to favor outage management and AMI operationsanalyticscomparedtoEuropeanrespondents (see Figure 1).

In addition, results from Accenture’s survey find that utilities clearly assess their current level of maturity on analytics capabilities in need of improvement. This characterization is true for all the facets of the capability, including data quality, tools and skills (see Figure 2).

InsightsfromAccenture’sDigitallyEnabledGridprogram: 2013 executive survey

Figure 1. Analytic areas of greatest value to smart solutions deployment.

Which analytical areas would represent the greatest value to smart solutions deployment for your network?

96%

69% 92%

85%

81%

77%

77% 77%

Grid operations analytics

Asset management analytics

Outage management analytics

Volt/VAR analytics

Communications network operations analytics

System planning analytics

AMI operations analytics

Demand-response analytics

Customer operations analytics

Customer segmentation and behavioral analytics

Revenue protection/theft reduction

AMI deployment analytics

Distributed generation analytics

EuropeNorth America

93%

56% Europe

North America

83%

56% Europe

North America

87%

Select all that apply

73%

73%

58% 52%

50% 48%

Base: All respondents, analytics section.Source: Accenture’s Digitally Enabled Grid program, 2013 executive survey.

Accenture’s Digitally Enabled Grid program: 2013 executive survey methodology

Accenture conducted an executive survey among utilities executives worldwide involved in the decision-making process for smart grid-related matters in their company. The survey results are based on questionnaire-led interviews with 54 utilities executives in 13 countries, conducted via telephone in 2013 for Accenture by Kadence.*

Page 8: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

8

Figure 2. Maturity assessment of current analytics capabilities.

Significant need for improvementSome need for improvementNo need for improvement

Data latency (time delay of

systems)

Access to real-time data

Statistical analysis/data scientist skills

Analysis toolsets (coverage/

completeness)

Data governance

IT/OT integration (i.e., integration between real-

time systems in the field and enterprise

IT systems)

Data integration across sources (integrate data from consumer, weather, etc.)

Data visualization

Data availability (detail, gaps, timeliness/ currency)

Data quality (accuracy and

precision)

53%

35%

12%

46%

46%

8%

42%

48%

10%

35%

65%

33%

63%

33%

55%

12%

31%

63%

27%

65%

8%

23%

64%

13%

15%

77%

8%4% 6%

Base: All respondents, analytics section.Source: Accenture’s Digitally Enabled Grid program, 2013 executive survey.

How would you assess the maturity of your current analytic capabilities in each of the following areas?

Page 9: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

9

Becoming an information-driven organization

Successful use of analytics requires great focus, having a clear view of what business activity or business processes they impact and what particular business outcomes they might improve. This focus helps identify and prioritize the analytics most relevant to a utility’s business goals. Figure 3 depicts some operational areas of analytics and the operating areas they affect. Other analytic models not included in the figure include corporate services (finance, supply chain, human resources), advanced asset maintenance and capital planning management, and work and field force.

We see this focus reflected in examples from around the world: one large Midwestern US utility is pursuing analytics initiatives to support load planning and forecasting; in the Netherlands, another utility is testing how the use of granular usage data could change the capital required for distribution. In Canada, a large utility has focused on developing asset management regimes that incorporate sophisticated asset risk heuristics and visualizations (as are utilities in China, Singapore and Australia).

Revenue protection and consumer programs are also high on utilities’ priority lists, with consumer programs having a wide range of objectives. Additionally, we also see increasing demand for analytics supporting network operations, specifically for the AMI network (examples include a large Northeastern US utility and a large Canadian utility).

Figure 3. Key operational analytical areas and affected operating areas.

Note: NOC: network operations center, SOC: smart operations center.Source: Accenture.

AMI Deployment NOC/SOC

Meter deployment metrics

Deployment exceptions

Meter discovery and certification

Common deployment performance

Installer performance

Deployment optimization

Comms. status Event management

Grid topology (as operated)

Security incidents

AMI Operations Outage/Fault Intelligence

Meter read performance

Remote meter operations

Meter data load performance

Meter quality Meter data consistency

Meter comms.

Outage identification, resolution and optimization

FLISR Outage comms.

KPI reporting and optimization

Revenue Protection and Billing Distributed Gen.

Exception handling performance

Billing data availability

Revenue protection (including theft)

Billing performance

Meter data accuracy

Unaccounted losses

Availability Capacity Voltage sag and swell

Power factors

Power Quality

Customer External Customer Internal Asset Management

Usage and consumption

Supplier and third-party transactions

Energy efficiency

Customer segmentation

Targeted marketing

Social media Usage aggregation (transformer, feeder, etc.)

Transformer load analysis

Demand- response effectiveness

Demand- response planning

Demand Response

Page 10: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

10

These examples are just some of the ways we see analytics transforming how the business operates. As illustrated by the Moneyball example, there are numerous areas where analytics can provide a transformative impact for decision making within utilities. Table 2 provides a list of 10 opportunity areas that we see driving significant value and are priority focus areas for utilities.

Accenture research and experience in multiple industries confirms that insights alone will not make any difference to operating results. Analytics create no inherent value if they are simply focused on reporting or involve putting large quantities of data into sophisticated repositories. Insights from analytics must be used as a way to fundamentally change decisions that impact utilities’ operating practices.

This is a lesson derived from years of investments in industries that were early movers in analytic capabilities, but often struggled in early stages to realize results from their investments.

In Accenture’s view, there are six insights that demonstrate how analytics can be used to transform operating practices and help achieve more fundamental improvements in results.

How analytics drive real value1. Analytics can radically improve decision

making at every level.

2. Real insight comes from business intelligence and analytics that correlate data across the enterprise and externally.

3. New insights are possible by moving from business intelligence to comprehensive analytics to fundamentally manage business differently.

4. Achieving sustainable value from analytics requires new approaches to information management.

5. Changes to the operating model are required to enable benefit realization from new insights.

6. Rapid learning improves models and increases value from insights.

Table 2. 10 priority use cases that can drive significant value for utilities.

Opportunity Areas Drivers

Revenue protection Detecting unauthorized use and configuration errors and recovering lost revenue

Voltage optimization Using asset-condition models to refine operational settings of assets to save on power costs

Demand-response effectiveness Increasing participation in demand-response programs and improving savings achieved from load control

Load forecasting and planning Improving long-term investment planning based on bottom-up demand and asset load and condition indexes

Outage detection and response Reducing outage costs from enhanced response to outages (detection, isolation and restoration)

Outage prevention Reducing equipment outages by focusing on assets with highest risk of failure

Investment planning Revising priorities of asset investments based on analysis of asset risk and consumer impact

Maintenance strategies Revising maintenance strategies, policies and programs based on condition and risk analytics

Energyefficiency Identifying and helping consumers improve value from energy and energy efficiency

Energyservices Targeting consumers for services and pricing to help improve value from energy and with adoption of new uses (e.g., use of distributed generation, photovoltaics)

Page 11: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

11

As previously stated, Accenture’s analysis suggests that the potential value of smart grid analytics, at a conservative estimate, could approach $40 to $70 per electric meter per year, with estimated benefits generally split with 60 percent benefit to the consumer and 40 percent to the utility. Figure 4 illustrates potential annual savings

from AMI analytics in various business areas, created for a representative US utility.

The bulk of the benefits to utilities, up to 85 percent, are derived from four main categories:

•Assetmanagement

•Powerquality

•Revenueprotectionandbilling

•Outageandfaultintelligence

Using data from Figure 4, the estimated share of annual savings per meter from AMI analytics, by business area, is shown as percentages in Figure 5.

Case in point: The US estimated potential value of smart grid analytics

Figure 4. Estimated average annual savings from AMI analytics, by business area, on a US dollar-per-meter basis (representative US utility).

1.10.11.31.52.54.32.5

6.24.1

7.24.5

10.2

Note: NOC: network operations center; SOC: smart operations center.Source: Accenture analysis, 2013.

AMI deployment

AMIoperations

0.1

Outage and fault

intelligence

NOC and SOCPower quality

7.8

Revenue protection and

billing

4.5

Customerinternal

0.2

Asset management

26.6

Customerexternal

Demand response

1.11

32.4

9.9

Maximum benefitMinimum benefit

Figure 5. Estimated share of annual savings per meter from AMI analytics, by business area, as a percentage (representative US utility).

Note: Total net benefit 100% = $64.1 per meter.Source: Accenture analysis, 2013.

Outage and fault intelligence

Power quality

Asset management47%

14%

11%

9.1%

8%

Revenue protection and billing

Customer external

Demand response 5.2% NOC and SOC 2.7% Customer internal1.2%

AMI operations0.9%

AMI deployment0.8%

Page 12: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

12

1. Analytics can radically improve decision making at every levelThe widespread adoption of AMI and other sensors in electric utilities is providing massive amounts of real-time data, which can help utilities make better-informed decisions. This data can be used to provide “bottom-up” insights that are more frequent, more granular and more accurate than the information they are currently using. For example, at one electric utility, distribution planners use a single snapshot of peak system load taken twice a year to base all decisions about the planning of assets, and plans for each area are reviewed in multiyear cycles. By using meter-interval data combined with asset information, the utility would be able to use daily and hourly information about the actual load on specific assets. In an early prototype, the utility discovered that many transformers were substantially underutilized.

For these types of changes to be possible, new models need to be adopted for reviewing system performance, evaluating investments and measuring outcomes. In the planning example, reversing the flow of insight creates a dramatic change in the quality and timeliness of the decisions possible (from top-down based on experience or guidelines to bottom-up).

Figure 6 illustrates how the joining of bottom-up insights and top-down objectives can transform the management model and impact all levels of decision making.

It is the availability of bottom-up information, when tied to top-down objectives and goals that can dramatically improve decisions and performance. In utilities, bottom-up data can impact how decisions are made in multiple areas—assets, workforce and field operations, grid operations, and customer. In the late 1990s, the struggling US retailer Sears Roebuck & Co. developed a sophisticated statistical model that correlated bottom-up store performance data with store profitability.4

Using the model to manage everything from staffing to merchandise mix, Sears began a half-decade of resurgent growth and profitability. The sponsor of the model, AlanLacy,waslaternamedCEO.

In the same way, the existence of detailed big data information can transform how utilities manage the business. Three specific capabilities arise from big data that can lead to the adoption of new management models:

•Predictive models: Big data analytics provide the basis for probabilistic models and scenario-based planning approaches that can fundamentally change the nature of planning. For example, in asset investment planning, by integrating data sources automatically and embedding institutional knowledge into analytic models, overloaded and underutilized assets are more easily identifiable. This insight provides a more detailed basis for ranking investment projects and also makes planning easier by automating asset planning analysis.

Figure 6. Illustrative view of potential analytics areas impact in the decision-making process.

From grid to C-suite

Analytics and business intelligence

drive decision

and actions

• One vision of entire system

• P&L management

• Capital management

• Portfolio risk management

• Prioritization of asset management

• Proactive asset intervention

• Workforce optimization

• Enabling faster decisions

• System health and status availability

• Real-time decisions

• Predictable restoration times

Data and complex event management

• Automate complex activities

• Recording of events

• Integration and visualization of data

• Monitor and record asset performance

• Monitor customer demand and delivered quality

Executiveintelligence

Business application and decision support

Real-time operations monitoring and analytics

Intelligent sensor anddevice data generation

Source: Accenture.

Page 13: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

13

•Real-time optimization: The availability of real-time data at the asset level provides a basis for day-to-day decisions that can significantly improve the optimization of the network and energy use (example: Volt/VAR control and voltage optimization). With greater adoption of distributed generation and storage, having real-time data about the grid greatly increases the operator’s ability to manage the grid effectively within acceptable parameters and for the best economic outcome. This is the same capability that helped telecoms manage network investments as their Class 5 telephony systems became the obsolete (to data and packet-switched systems).

• Performance reporting: Big data provides far more granular insight about performance against specific objectives. With the availability of more frequent and granular performance data, it becomes possible to correlate objectives with specific performance metrics, making it easier to make changes that help meet performance goals.

With big data, operators and managers are able to make more informed decisions and, more importantly, understand and validate the effect of their actions on overall objectives in reasonably short time frames. Analytic models make it possible to carry out detailed scenario analysis to explore different options, providing leadership with greater information and situational awareness than they had before.

2. Real insight comes from business intelligence and analytics that correlate data across the enterprise and externallyToday, most operational systems are managed within organizational silos, with data owned and managed by different parts of the organization and toward different objectives. Overcoming these barriers can reveal statistical relationships that lead to real insights on how to optimize processes. As an example, combining meter-interval data with data from the distribution management system (DMS) and SCADA

control systems enables algorithms that could allow a utility to discover where opportunities for voltage optimization exist. Another example is prioritization of capital investments. Combining interval data and data from asset information and work management allows for the creation of asset health metrics based on asset utilization and load. The same approach could help optimize crew location and assignments based on information about consumers reconnected after an outage. Numerous other examples illustrate the value of bringing data from disparate systems together.

Canada-based Hydro One recognized the criticality of data quality and of bringing multiple sources of data together to improve how the utility plans investments in its transmission and distribution assets. The utility has deployed an asset analytics solution to access and integrate historical asset data and information from multiple databases and business applications that were either inaccessible or not integrated into the utility’s information technology and operational technology networks. With the implementation of this solution, the utility’s asset management, engineering and planning workforces now have information on the performance of transmission and distribution assets that is time-based and locational. Hydro One’s managers are able to access integrated asset information to help them further identify and plan for short- and long-term investment scenarios and deliver performance outcomes that mitigate power system and ongoing investment risk. By pulling together the data and information, Hydro One expects to have improved data quality and advanced predictive analysis capabilities to define the requirements and forecast expenditures needed to maintain the electrical grid at an optimal level.

Correlating internal data with external data sources provides additional insights. Consumer usage profiles based on hourly or 15-minute interval data can become far more useful when combined with external weather data or with real estate data that could provide insight on the square footage of a house, which could help benchmark the home in terms of energy efficiency.

3. New insights are possible by moving from business intelligence to comprehensive analytics to fundamentally manage business differentlyMany utilities have made significant investments in traditional forms of business intelligence (BI), including reporting systems, data warehouses and data marts of historical data. Trends in analytics highlight some key differences between traditional BI and more comprehensive analytics.

Data that drives comprehensive analytics is distinguished from traditional BI approaches in at least two ways:

•Time:Comprehensiveanalyticswillincorporate very low latency data (seconds, minutes, hours) as well as high latency (days, weeks, months) analytics more common in BI.

•Dataorganizationtypes:Advancedanalytics will consider many unstructured sources (social media, e-mails, websites, customer comments) as well as structured data.

The emergence of these higher-velocity sources of data coupled with large-scale data from existing operational systems explains what makes analytic data so “big.” Using sophisticated computational models and statistical techniques on these new combined data sources make new insights possible on a more regular basis—effectively creating a new engine for insights.

Other industries have been leveraging these complex sources—some for more than 20 years. Many insurance companies have developed sophisticated data quality processes and systems to verify analytics are using highly reliable data for decision making. Credit card companies are able to flag cards that are used in geographical locations that do not appear in that card’s prior history and confirm the transaction with the customer in real time.

Page 14: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

14

Likewise, utilities can similarly use unusual patterns of usage to identify theft and security breaches. Oil companies are now able to monitor the real-time performance of well assets in making decisions to optimize production. Companies including Apple (iTunes), Pandora Media (song recommendations) and Amazon (purchase recommendations) illustrate the incredible power of inference engines used in other industries to help identify customer needs and make targeted offers.

4. Achieving sustainable value from analytics requires new approaches to information management Because of the collection of data across the enterprise, big data analytics require more discipline around information management. In turn, more collaboration is required between different parts of the organization that had not previously worked together. For example, since meters are related to billing, AMI meter data is often owned by the customer organization. However, AMI meters are an extension of the distribution network and can, among other benefits, provide significant information for better grid operation, increased reliability, higher levels of safety, as well as contribute considerably to improved asset management. The tension between who “owns”’ the data can prevent it from being effectively utilized across the enterprise.

Another significant need is for a common data model that allows data from all these silos to come together to provide meaningful insights. The electrical distribution grid changes continuously with the operation of switches, reclosers and the like. A typical utility’s grid will include millions of components, including AMI meters. The ability to understand the layout of the electrical connectivity model at any point in time is necessary for the effective use of some analytics, especially those around grid operations, outage management and asset management.

New regimes for information management have proven to be the foundation for many industries that successfully leveraged analytics. These industries, including financial services, telecoms and health, are generally much more mature in their

information management capabilities than the utilities industry. Their approaches to information management typically include data strategy and life cycle, master data management and data governance and, with the advent of the deluge of data made possible by smart devices, big data.

We know that providing the right data, to the right person, at the right time is critical; however, as we previously noted, data quality is often a major barrier and remains a major issue for utilities. For analytics to be useful and dependable, accountabilities for data quality need to be in place and data capture activities must be integrated into process activities. Adopting measures of data quality that can be used to drive accountability is also a key change critical to improving data assets.

Big data and big analytics

The key characteristics of big data are volume, velocity and variety, and big data goes beyond the ability to simply store large amounts of data. An effective big data solution needs to:

•Deriveinsightfrombigdatainnearreal time for certain types of analytics needed for utilities. Traditional business intelligence typically deals with high volumes of data but it is often used for after-the-fact analysis.

•Managelargevolumesofdatawhiletaking advantage of more cost-efficient scalable architectures.

From a performance perspective, typical data analytical engines are often tuned to do one or the other of these functions effectively. Solutions for big data are now leveraging open-source platforms to support massive transactional systems that support social media platforms (like Cassandra, the distributed data system behind Twitter and Facebook). These platforms will quickly change the price/performance ratio of big data, again transforming what is possible at reasonable costs. The capability will not be achieved simply by buying the fastest or largest infrastructure. Instead, the challenge lies in careful and considered attention to smart design of data models, data integration, data representation during transit, data storage and analytics applications.

Most utilities to date have managed their big data by providing analytics in a niche area; e.g., for theft or meter deployment, rather than holistic considerations. This model has been a successful one for achieving specific use cases, reinforcing our previously outlined perspective. In addition, Accenture believes that a foundation consisting of robust data integration architecture along with a comprehensive utility data model is essential to bringing greater value to the utility and to the long-term viability of an analytics solution; in particular, with regard to scalability, performance, flexibility and extendibility across the enterprise.

A comprehensive analytics solution should also be able to handle structured and unstructured data along with having the ability to deal with low- and high-latency analytics. More recent entrants in this space,primarilyC3EnergyandSiemens,have these key big data model features built into their solutions.

The cloud has also changed investment models driving analytics across many industries. For a utility, building and managing its infrastructure to meet these demands is expensive and can take focus away from the more important activities around data management and implementing the necessary change to achieve the business case. A key characteristic of the cloud is its ability to scale up or down on demand, which makes it an attractive alternative to building and maintaining an in-house infrastructure for analytics. Key considerations in a cloud-based solution include:

•Dataprivacyandsecurity.

•Costofdatatransportation(toandfrom)the cloud infrastructure and associated data latency.

Page 15: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

15

Accenture’s recommended conceptual solution architecture is depicted in Figure 7.

Data definition—management of master data

Effectivemanagementofmasterdataacross an enterprise is a significant endeavor and utilities should treat it as such. For example, operational meter metadata lives in multiple systems (head-end, meter data management system, asset inventory, workforce management) and could be modified in more than one of thosesystems.Enforcingownershipandconfirming data is consistent across these systems requires robust up-front design and data governance. Consumer products companies, which started much earlier in managing analytics for insight, are now leading the way in developing capabilities to manage master data. These companies have realized the need for a unified capability to manage and unify data across their global operations.

Governance

Data ownership, maintenance and accountability need to be clearly documented and enforced for disparate data to come together and be effectively maintained, which is critical to enable the mining of deep insights. Governance models should clearly lay out how decisions and accountabilities for data quality and data management will work, including roles and accountabilities for data quality. Many of the discussions about the need for a chief data officer are really focused on the need for strong and clear data governance.

Data security and privacy

Granular data is a double-edged sword. For example, while interval data provides previously unavailable insights about the consumer and the grid, it can also provide definitive information about whether someone is home or away, what appliances they have and how they use them. All of these insights can be considered an intrusion for some, especially if the data has the potential to

be used for nefarious purposes. Handled incorrectly, these deep insights about the customer can violate consumers’ privacy and fall foul of regulations.

As a result, regulations managing data privacy and data policies are continuously evolving as regulators and policymakers try to keep up with advances in technology and analytics. Consumer sentiment regarding these types of insights is evolving as well. For example, when handled carefully with permissions, consumers are generally willing to allow very intimate data about their lifestyles to be managed for them. But, breaches or inappropriate uses will be punished severely by consumers and by the public media. For example, in 2006, search data from a large set of AOL users was inadvertently posted online, leading to the resignation of the chief technology officer and members of the product team.5 Careful planning, management and communication of data policies and extremely rigorous processes for client data protection are necessities.

Figure 7. Conceptual architecture for analytics.

Web Reports/dashboards/visualization tools

Mid-/high-latency analytics

Real-time analytics(CEP; in memory)

Integration/business process management

Grid/enterprise systems

Analytics data store (meter/device data repository - structured/unstructured data)

• Operational data

• Events

• AMI data

• Topology data

Note: OMS: outage management system, DMS: distribution management system, EAM: enterprise asset management, GIS: geographic information system, MWMS: mobile workforce management system.Source: Accenture.

• Asset data

• Grid connectivity data

• Meter data

• Other data (weather, census, etc.)

• OMS

• DMS

• EAM

Actions/status

• GIS

• MWMS

• Others

Page 16: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

16

5. Changes to the operating model are required to enable benefit realization from new insightsDespite the incredible power of the insights that can be extracted from big data practices, value can only be achieved if those insights are used to improve or transform a process or set of activities. Providing insight to employees is not enough—insights should be coupled with overall goals to drive changes to how work is actually performed.

For example, insights about revenue protection opportunities may require changes to how field investigations are scheduled and changes to billing operating practices; for example, policies that determine actions taken when theft is discovered. A situation that has a very strong likelihood of theft could automatically create a work order for field investigation. On the other hand, less-certain situations might be routed to an experienced analyst for further investigation.Evenmoreimportantistheneed to track the outcome of the analytics

and its success rate to continuously improve on the analytic algorithms, and stay ahead of intentional misuse.

Figure 8 depicts the two critical aspects of realizing value from analytics. On the left, insights come from sophisticated computation models. On the right, revised operating model and changes in operating practices allow for value to be realized as management models and work practices are modified. While this is a straightforward implication, there are numerous examples in many industries of companies investing in the left side and failing to facilitate the change that actually delivers the results. Telecom companies were notorious in the early stages of business intelligence for building data warehouses and providing analytic workstations without much impact on the business.

At a minimum, new insights will require tweaks to how roles are defined and how certain activities within a process are performed. Some insights require a more fundamental rethinking of the operating model, the basic architecture of how the business operates. For example, one Europeantelecomcompanyrecently

completed a major effort to gather insight on its enterprise customers. In doing so, it realized it had to restructure the role of the product teams to more effectively deal with the insights—the current model did not provide the appropriate accountability for its teams to be able to use the insights.

Important insights can lead to new ways to manage teams and work, and characteristics of the operating model will need to be refined or reconstituted. We define an operating model as having the following characteristics:

•Leadershiprolesandaccountabilities

•Organizationandworkgroupstructure

•Valuemetrics—whatarethecriticalmeasures of success?

•Governance—howarekeydecisionsmade?

•Processes—howisprocessperformanceand accountability managed?

•Capabilities—whodoeswhat,where?

•Culture—whatareexpectedbehaviorsinkey roles?

Figure 8. Realizing value from analytics: From insights through process transformation.

Insights from analytics

Analytics platform

Process transformation

Value

Data management

• Data integration

• Data cleanse and transformation

• System integration (e.g., interfaces)

• Data operations

Computational models/analytics

• Computational models

• Propensity models

• Stochastic risk models

• Consumer value/ segmentation models

Insights

• Decision- support tools

• Scenario modeling

• Target lists

• Query/report visualization tools

Roles

• Revised roles

• Revised decision models

• Revised performance criteria

• New skills

Process

• Change operating practices

• Process changes

• New ways to work

• New process performance metrics

Operating model

• Organizing structure

• Leadership roles and accountabilities

• Value metrics

• Governance

• Behaviors/culture

Page 17: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

17

The insight about how to optimize voltage on the system is an interesting example. In a typical utility, voltage guidelines are the responsibility of area engineering managers and operations. But savings from conservation voltage reductions would accrue to generation or the commercial group responsible for sourcing power. How can this important trade-off decision be made when the costs and benefits reside in different parts of the organization? In reality, a new value metric is probably required—net margin on power delivered. This would encourage more optimal choices regarding voltage and power factors. Currently, this metric might only be visible to the utility CFO, or to the head of the commercial group who manages the plants, but most of the changes to enable the improvement to this factor would be made in operations. Clearly, some refinements to the operating model would help encourage the right decisions.

Another key consideration is the availability of the appropriate skills to unlock the potential of analytics. According to Accenture’s survey, 90 percent of respondents cited access to the right IT skills as critical or important for their

companies to manage the increasingly large data volumes and integration. Constraints on the analytics talent pool cross industries and geographies. In the United States and United Kingdom alone, jobs demanding STEMskills—advancedknowledgeinsciencetechnology engineering and mathematics—are projected to grow five times faster as jobs in other occupations by 2018. Utilities require talent with both engineering knowledge about the distribution networks and a deep understanding and ability to manage and analyze large amounts of data as well. Utilities need “data scientists” who have deep mathematical and statistical capabilities.6

The severe shortage in the pool of analytics professionals makes it imperative for analytics users to develop comprehensive strategies for building, hiring or sourcing therequiredanalyticsresources.Emergingeconomies are producing larger numbers of STEMtalentthandevelopingeconomics,butthis is still not sufficient to meet demand. In addition, Accenture’s survey shows that only 25 percent of respondents feel their utilities are very well positioned to compete for analytics skills in the market (see Figure 9).

6. Rapid learning improves models and increases value from insightsMost utilities are still at a fairly immature capability level to achieve full value from the types of insights highlighted in this paper. The reality is that analytic models have limitations, and can only improve and get more sophisticated (and therefore more reliable in terms of insight) using a good feedback loop with experienced business managers and executives. For example, improving statistical models that predict equipment failures requires insight on that experience over time with failure rates by asset class. The most effective insights come from predictive models, or models with smart heuristics to define when something is more likely to happen (for example, which consumers are more likely to participate in demand-response programs).

“Try and refine” is a critical approach to improving analytic models and insights. Focusing on a few specific use models where analytics can make a difference is the most effective way to refine the models while realizing some immediate value. Given the challenge of data management, it is important to have a view of the overall roadmap and total value of the journey to becoming an information-driven utility. This broader view is key to developing a holistic data roadmap and getting buy-in for the greater value that analytics can provide, without having to solve the challenges of data governance and quality all at once. The other benefit to starting early is to begin the longer journey to changing behaviors and expectations in the business as soon as possible. Speed to market is critical, as key behaviors can be better understood, modeled and predicted.

Figure 9. Ability of utilities to compete for analytics skills in the market.

How well positioned do you believe your company is to compete for analytics skills in the market?

25% 25%

50%

Poorly positioned Very well

positioned

On par with other companies

Base: All respondents, analytics section.Source: Accenture's Digitally Enabled Grid program, 2013 executive survey.

Page 18: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

18

Conclusion

There is significant value to be derived by applying analytics to the utility business. While the technical aspects are important, the real challenge is to infuse insight from analytics into operations—embedding analytic insights into business processes and decisions and to drive a more fundamental change for the enterprise. Prior efforts have demonstrated that, while technical challenges can be tangible, the business and management changes are more challenging because they require a change in how people think and act

across the organization. Using analytics is, therefore, one of the most important ways for utilities to achieve new results to support strategic imperatives and improve shareholder returns. High-performance businesses make analytics an integral part of their daily business process, and this should be a high priority for the utilities industry. The most important action is to find value propositions and use cases that are aligned with a utility’s strategic objectives and begin the journey.

Page 19: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

19

1. Moneyball: The art of winning the unfair game, Michael Lewis, 2003, W.W. Norton & Company, Inc.

2. “Smart grid data analytics spending to total more than $34 billion through 2020, forecasts Pike Research,” Business Wire, September 17, 2012, http://global.factiva.com.

3. ”Utility AMI analytics market to hit $9.7 billion by 2020, according to new GTM Research report,” GlobeNewswire, June 13, 2013, http://global.factiva.com.

4. Anthony J. Rucci, Steven P. Kirn, Richard T.Quinn,“Employee-customer-profitchain at Sears,” Harvard Business Review, January-February 1998.

5. “AOL executive quits after posting of search data,” International Herald Tribune, August 23, 2006, http://global.factiva.com.

6. Analytics in Action: Breakthroughs and Barriers on the Journey to ROI, Accenture, 2013, www.accenture.com.

* Countries in scope for Accenture’s DigitallyEnabledGridprogramexecutivesurvey: Argentina, Australia, Brazil, Canada, France, Germany, Italy, Japan, Netherlands, Spain, Singapore, United Kingdom, United States.

References

Page 20: Unlocking the Value of Analytics | Accenture...Analytics adoption to date 6 Analytics opportunities for high performance 6 Becoming an information-driven organization 9 How analytics

Copyright © 2013 Accenture All rights reserved.

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

Copyright © 2013 Accenture All rights reserved.

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

Other names may be trademarks, copyrights and/or service marks of their respective owners. 13-3457/11-6785

Senior executive sponsorJack Azagury Global managing director, Accenture Smart Grid Services [email protected]

AcknowledgementsWe would like to thank the following individuals for their contributions to The Digitally Enabled Grid program:

Andre Begosso Gregory Bolino Stephanie Bronchard Jonathan Burton Jenn Coldren Robin Dicker David Haak Tony Histon Robert Hopkin Vasan Krishnaswamy Wade Malcolm Trygve Skjøtskift Ashok Sundaram Brent Zylmans

Accenture Research: Jason Allen Lasse Kari Charlotte Raut Carmen Uys

For more information on Accenture’s DigitallyEnabledGridprogram,goto www.accenture.com/digitallyenabledgrid.

About Accenture Smart Grid ServicesAccenture Smart Grid Services focuses on delivering innovative business solutions supporting the modernization of electric, gas and water network infrastructures to improve capital efficiency and effectiveness, increase crew safety and productivity, optimize the operations of the grid and achieve the full value from advanced metering infrastructure (AMI) data and capabilities. It includes four offering areas which cover consulting, technology and managed solutions: Work, Field Resource Management; Transmission & Distribution Asset Management; Advanced Metering Infrastructure and Grid Operations.

About AccentureAccenture is a global management consulting, technology services and outsourcing company, with approximately 275,000 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$28.6 billion for the fiscal year ended Aug. 31, 2013. Its home page is www.accenture.com.