Delivering Value Out of Smart Metering: A Fully Operational Case Study of Combining Big Data Analytics with Process Change

Oracle Utilities and We Energies discuss how to create real customer engagement and measured benefit realization through the use of Big Data analytics software.
Published: Mon 31 Mar 2014

Data analytics should be fundamental to any utility. However, according to Rick Brakken, VP of Utility Analytics, Oracle Utilities, some utilities say they simply don’t have the time to use this data to its full potential.

In a recent webcast, Mr Brakken dicusses this point by saying that “big data is meaningless if it sits on the shelf.” He says that utilities should take advantage of the wealth of data on offer today. “The actual data itself is not valuable but if you know the algorithms to process that data in an automated fashion then what we're getting is the value of the data without having to engage in all the data."

Big data and predictive analytics are transforming utilities. Information-enabled decision-making is where the modern utility must move towards in order to create more personal, effective and long-term relationships with their customers.

Click here to watch Oracle Utilities recent webcast On-Demand

More than an IT solution

Data analytics is more than an IT solution as it has the potential to culturally change what a business is doing or aiming to do. The leveraging of data can improve operational efficiencies, reduce costs and improve customer service. It is a progression, an evolution-not a revolution in terms of what’s going on with analytics and big data.

Data analytics is becoming a hot topic, thanks to smart meter devices which are generating terabytes of data which never existed before. The smart data is new to utilities, providers, and distribution companies and retailers. There is a growing need to analyze this new smarter information as it promises to improve business processes and customer satisfaction.

Analytic opportunities

Across the utility chain, there are numerous opportunities for the grid, meter, customer and finance department:

  • Grid-Analytics can enhance distribution operations, system planning and energy procurement. An increasing number of utilities are beginning to rely more on data to manage the revenue stream.

  • Customer-Analytics can improve billing accuracy, service efficiency, and the meter to bill process. Call centers will have access to more relevant information so that they can communicate more efficiently and effectively with the customer. Call centers will have available data to resolve billing complains a lot quicker than before.

  • Meter-Utilities can use data to find aspects that are unsafe such as gas leaks, water leaks and meter overheating. They will have access to real-time information to identify anomalies. Demand response and energy efficiency will make assets more efficient. Utilities will be getting more out of the feeders.

  • Finance-Analytics assists with combatting theft and aids revenue protection. In addition to this, utilities will have access to data which will help them compile more accurate financials. Utilities can use real-time data to improve forecasting and rate development. They will also be able to identify meter tampering and theft, as well as behavioral trends.

There are many ways that data can be used as it covers the entire value chain. This also applies in deregulated markets. Many of these areas are applied to the different components-be it the retail or distribution company, explains Mr Brakken.


The challenges of data analytics

Utilities face key challenges in realizing the full value of analytics:

  • Information silos and different systems-Utilities should unify the analytic platform scaled for big data processing. The data should be combined in a meaningful way. For instance, the combination of customer, distribution and network data will result in meaningful results. All the data sources should be brought together and the end result will prove to be more useful.

  • Static legacy business processes-Legacy systems are not designed to absorb and process large volumes of data. Utilities no longer need large volumes of data-they need actionable results. There is a growing need for transparent transactions and real-time meaningful data.

  • Skill set gaps and lack of change management-There need to be executive and organizational commitment. It’s not only about the technology-it takes proactive behavior and a desire to do things better. With analytics, it’s all about searching for opportunities to improve efficiency and finding solutions to resolve problems. Analytics is discretionary-it’s not mandated. A supportive management team and executive sponsorship must be in place in order for the goals to be achieved.

It takes time to trust and understand the data and implementation can take time initially, explains Mr Brakken. There has to be a change in traditional mindsets within the utility business. Moving into a new direction will take time.


Analytics Myths

Analytics myths have led to failed initiatives:

  • A single system can solve all analytical needs-This is untrue since it’s not just a system that you plug in. It’s not just a new calculator.

  • The “calculator” is the most important part of solving complex math problems-This is wrong since the mathematician and algorithms are critical to the analytics process. The technology alone doesn’t solve the problem.

  • Each utility’s needs are totally unique-Industry is basically 80% similar and 20% different. Utilities in most places are competitive and tend not to share things. On the distribution side, there is a tendency to share things. Therein lies a good opportunity to share best practices and learn from each other. Utilities should share data models and algorithms instead of reinventing the wheel since this takes up time and money.

  • Build it and they shall come-It’s not just about getting IT to develop a solution to handle the data. The business needs to engage throughout the entire process. In order to draw more value from analytics, it involves more engagement.

  • There is a finish line-Analytics is a capability that goes on and on. It is an evolution and there is no finish line. The idea is to expand the capability and keep improving business processes. There is no finish line since the aim is to expand the value beyond 2 or 3 areas at least.


Operational Analytics Framework

Calls for a practical approach which will yield real results. By integrating various data sources-premise, network connectivity, customer attributes, asset data, AMI meter data, and weather data, the utility will achieve actionable insights for the following business operations:

  • Customer service and support

  • Meter operations and billing support

  • Efficiency and demand response

  • Revenue assurance

  • Load forecasting

  • Network planning and operations

  • Asset optimization

Sample outputs would include:

  • Meter malfunctions

  • Customer tampering

  • Overloaded transformers

  • Billing corrections

  • Maintenance priorities

  • Energy efficiency and demand response targets

If utilities carry out these operations in a smart way using data analytics, they will be able to solve different problems, resulting in real results:

  • Reduce back-office processing

  • Increase customer satisfaction

  • Prevent public safety hazards

  • Improve operational maintenance-phase balancing, and prevent overloads


Case Study: We Energies

Oracle Utilities has worked alongside We Energies for several years and Mr Brakken says that the company represents the best example of a “top-down” approach to having great executive sponsorship. He says they take a serious approach to data analytics and that they understand the need for change management. It is for this reason that We Energies has been chosen for our case study.

We Energies’ main focus has been on “meter to bill” for the last two years. The company is responsible for an area which consists of a mixture of meter reading systems for both gas and electric customers. Their biggest challenge was to establish a system that would integrate data from all the different metering systems.

Peggy Clippert, Manager of Analytics, We Energies, lists the drivers of a data analytic solution as follows:

  • A high volume of billing exceptions every month

  • Length of “break to fix” too long (time from meter malfunction to rebill). This was leading to poor customer satisfaction

  • Too many re-bills because of inaccuracy

  • Not taking advantage of daily reads

  • Not finding all the metering issues and losing the opportunity to get value out of the automated metering system. Issues weren’t been dealt with effectively or efficiently enough

  • Limited field resources since the company was trying to keep costs at a minimum


Project Description

Ms Clippert gives a brief description of the company’s analytics project:

  • Replace the existing customer information system (CIS) monthly generated exceptions with meter data analytics exceptions-high volume, high risk, or high complexity.

  • Determine a new approach for a comprehensive, dynamic, manageable and prioritized list of exceptions

  • Send high probability meter error exceptions directly to the field while others work queues in the customer information system using existing processes

  • Use business/IT/vendor collaboration


Systematic, integrated approach

The customer information system basically sends meter and customer data to Oracle Utilities via algorithms. Exceptions are then returned to We Energies and are then processed via an existing work tracking system. The data is sent via two paths:

  • A work order management system for field staff; and

  • Back office processing for billers


Completed 5 Phases

The project was named “High Low Slow No and Jo” (High consumption, low consumption, slow consumption, no consumption and Jo is the metering engineer).

Jo worked with Oracle Utilities for a year, looking at large electric meter error and diagnostic flags. Those were the first ones that were put into effect. If these had high voltages, high current, or other things that he had been looking at manually, he would send those through to be worked on.

“In April 2012, we built some of those algorithms to do that work for him and get those down to a more manageable level,” explains Ms Clippert.

The project was made up of three cycles:

Cycle 1

In October 2012, We Energy put into effect a new algorithm for gas slowing consumption. “This was one of our most successful ones and it was good that we had such a successful algorithm from the beginning of the project as it really helped gain a lot of traction with our employees.” With this new exception queue, employees could pinpoint problems immediately. With this system in place, the company didn’t have to wait for meters to go to a “no consumption” failure since these meters could be dealt with during the slowing down process.

By using analytics, over 3000 slowing meters were uncovered before shifting to a “no consumption” stage. As a result, US$934, 000-plus was rebilled. The company can catch these meters up to six months earlier now. “We can now replace these meters before they stop altogether.”

Cycle 2

In December 2012, the company focused on electric/gas “no consumption” and AMR Spikes and Reverses (Orders). Spikes and reverses were hidden in a large queue but have now been identified and are being worked on.

Approximately 13,000 orders were sent straight to the field (very few are sent to administration and billers) and saved a minimum of US$54,000. There was also a 75% reduction when comparing March and April 2013 volumes. Today, the volume remains steady.

“We redefined when and what gets flagged in order to get a better understanding of consumption,” explains Ms Clippert.

Cycle 3

In June 2013, the company focused on electric/gas low consumption, electric high/low demand. Previous queues were time-consuming but now, outgoing bills have been reduced and staff are left with less work to contend with. Ongoing tests help to verify and confirm correct usage, thereby reducing the number of queues.

Low consumption tests show an 83% decrease after one month-this includes both residential and commercial. This has resulted in a US$30000 savings. There has been a huge reduction in overtime work hours (to work on queues) which has saved the company a lot of money.

“The high consumption old queue was a “hodge-podge” of different things. In response, four algorithms were created to accommodate analysis. We are working on this one to try and get the attributes right. Unlike the other tests, no real reductions have occurred to date,” explains Ms Clippert.


Results of the project

Advanced metering diagnostics-large electric meters:

  • 86% decrease from 2011 to 2013.

  • 100 plus orders were sent straight to the field without having a biller look at them

  • This has helped to reduce “break to fix time” and has reduce the number of queues generated. This company is aiming for the lowest time period to attain higher customer satisfaction. “This makes us look smarter in their eyes and that we are on top of things. It is important for customers to have that trust in us,” says Ms Clippert.

The company realizes that while response time is critical to customer satisfaction, it remains a “delicate” balancing act. High priority problems are dealt with quickly whereas meters on a “low priority” status will be monitored for a bit longer before being dealt with.

There has been a significant reduction in queues over the project’s two year period. The backlog has been attended to and the numbers continue to drop with tests.


Success factors

Ms Clippert lists the project’s success factors:

  • Change management at all levels within organization and billing staff undergo extensive training

  • Diverse, committed project team following project management methodologies

  • Continuing performance monitoring and performance. The executive team showed a great deal of support

  • A flexible tool and partnership with internal IT

  • Internal climate on Big Data and Business Intelligence

  • Continuing training and communication (shift in how work is measured and skills needed)

“The human eye is still needed to analyze the real problems at hand. The work is now probably even more challenging for staff but at least the sheer volume of unnecessary work has been removed,” says Ms Clippert.


What’s Next for Meter Data Analytics?

Ms Clippert lists the company’s plans for meter data analytics:

  • Stop customer problems before they occur and avoid revenue leakage

  • Add algorithms for primary meters

  • Deploy meter health test during all installations

  • Evaluate and refine new queues at least once a year

  • Develop additional algorithms in the next highest complexity or volume areas

Ms Clippert points out that Meter Data Analytics is a high-value game changer for We Energies. She lists the key areas of value for the company:

  • Increased efficiency-faster identification and resolution of a variety of issues

  • Improved customer satisfaction-catching problems before customers become aware of them

  • Avoided IT Costs-Meter Data Analytics can take on reports, development and analysis that previously required IT times and costs.

With regards to employing new skill sets, Ms Clippert points out that her company used skill sets from Oracle. Therefore, there was no need to employ new staff members. “Existing teams learnt from Oracle’s suggestions. It is difficult to hire this kind of skill set so it is great to work with a vendor that has the skill set onboard already.”

In closing, Ms Clippert points out that meter data analytics has been the “be all and end all for us.” To attain real value from this process, she suggests that companies focus on one aspect to begin with: “Make as many inroads and go deeper into it in order to prove the value. Expand and explore horizontally. Deal with the obvious issues first and then move onto the next one.”