The global utilities sector is mobilizing to implement smart meters, with different countries at different points in their implementation. In the UK, the implementation of the government’s 2020 plan for completing the rollout of smart meters to every household is well under way and is gaining momentum. Smart meters have started changing customer expectations and business models for utilities.
From data scarcity to data profusion, from the periodic frequency of data collection to real-time data feeds, from benign customer engagement with companies to active customer interest in their energy consumption behaviour, many factors are bringing in a revolution to the industry. This presents an opportunity for the utility industry to capitalize on a datarich environment to enhance customer experience, improve their operational cost structure and enhance overall industry performance.
The Big Data challenge in utilities presents possibilities to grow the both top line and bottom line of energy retailers, along with providing additional value to the customers. Whether it’s accurate billing or advanced features enabled by smart metering like connected homes, there is a clear win for the customers. According to an estimate1 , connected homes, work, and digital cities will create a $731.79B market opportunity for utilities by 2020. Smart meter data can be monetized by addressing a few key areas of impact:
- Customer experience improvements through better customer insights, better understanding of consumption behavior and reduced billing challenges and by providing more control to consumers over their consumption decisions.
- Financial performance enhanced by growing the topline through tailored propositions, improved revenue assurance, reduced cost of operations and the innovative use of smart data for monetization.
- Enhanced governance mechanisms that support for integrated grid and improvement in overall efficiency of the grid/industry.
Monetization for enhanced customer experience
Customer outlooks will go through an evolution because of an increase in expectation from what smart meters can deliver. Current expectations of basic service levels like accurate and timely billing would become hygiene factors. Advanced propositions, flexible tariff plans, information and communication to support informed choices will become important to deliver higher value and help differentiate in the market place.
Data analytics will be the key to understanding and meeting customer expectations through appropriate products and services. The key to success would be to provide an actionable, real-time, yet simplistic view to customers to facilitate their decisions.
At a broad level, customer experience related to evolution of smart meters can be categorized into two parts.
1. There will be a need for a host of offerings around real-time visualization and control that help customers analyze and customize different factors like tariff plans, consumption behavior, appliance control, dynamic proposition selection, participation in demand management and local generation
2. There will be a need for rewards and loyalty programs that incentivize good behavior, promote loyalty and provide an integrated, consistent experience comparable across other industries
Real-time visualization and control
Smart meters are expected to transform customer expectations for accessible real-time appliance level consumption, customized recommendations, dynamic pricing, automated breakdown predictions and corresponding actions and information about efficient products in order to reduce their energy bill by controlling and managing their consumption behavior.
Appliance control and customization A significant change brought on by smart meters will be the increased adoption of home management solutions facilitating connected homes. Customers will expect several benefits from these networked homes:
The customers would want to have the ability to control and customize power consumption at their homes remotely using an app installed on their smartphones. Smart meter data analysis makes it possible to determine the resource usage of individual appliances, lights and HVAC system by analyzing their data patterns. Alerts or alarms could be activated on the app if an appliance is consuming more than average consumption of similar appliances, enabling the customer to get the appliance checked for any malfunctions. This in turn will help customers to keep their bills within a desirable limit and save money through the increased longevity of their appliances.
Utilities can use this opportunity to generate new revenue streams by offering insurance and service products where appliances are tracked real-time. Furthermore, sophisticated backend models like principal component analysis (PCA) can be deployed to cluster breakdown types and help engineers reach the site with the precise equipment required to fix the issue, resulting in savings by eliminating the cost of multiple trips. Also, renewal prices can be refined significantly by looking at previous performance data of appliances and charging a premium for cases where multiple breakdowns are forecasted.
Machine learning can be applied to improve energy efficiency, reliability and comfort by monitoring operations and using algorithms to adjust it for external weather while avoiding any manual interventions. Configuring smart meters to the real-time GPS location of a customer can help activate devices depending on the estimated arrival time after taking traffic condition into account. This would help provide a seamless, cost-effective experience for customers. A four step approach for implementing algorithms can be adopted to reach this goal:
1. Identifying consumption histogram,
2. A three-line algorithm for understanding the effect of external temperature on consumption
3. A periodic auto-regression (PAR) algorithm to extract typical daily profiles
4. Time series similarity search to find similar consumers
While the first three algorithms analyze electricity consumption of each household in terms of its distribution, temperature sensitivity and daily patterns; the fourth algorithm finds similarities among different consumers for adjustments and recommendations. This tracking and model can be extremely useful for utilities by helping them to accurately forecast customer consumption.
The price consumers pay do not reflect true cost of production, and also do not incentivize customers to shed during peak loads. The ability to “shave” peak demand could allow a utility to reduce built capacity and save on the cost of generation. Utilities can pass some of these savings to the customer through dynamic pricing. Smart metering promises several avenues for realistic pricing which can drive beneficial results.
Time of usage (ToU) pricing institutes a price schedule for electricity usage, under which electricity is least expensive when loads are low and most expensive during peak hours. ToU is based on the fact that by altering rates at different times of the day, providers can incentivize customers to adjust their loads, either manually or through home energy management systems.
Singapore employed a dynamic pricing pilot on its expanding smart grid in 2012, with dynamic pricing on 30 minute intervals. The pilot reduced peak residential loads by 3.9% and total energy consumption by 2.4%.
Breakdown prediction and security
By capturing real-time consumption data, utilities can identify breakdowns at an early stage. This can be achieved by looking at the consumption pattern of an appliance, as well as overall household usage levels to identify significant changes in trends from an expected value. While being able to proactively reach out to the customers would help save utilities on inbound call volume, customers would also be able to save on costs through early intervention before the damage reaches an irreparable level.
Smart meters also present an opportunity for data breaches. The risk of hackers infiltrating systems can be mitigated through systematic configurations. Utilities can effectively detect and eliminate the risk of tampering through advanced metering infrastructure. Utilities can consider providing several enhanced add-on features along with basic gas and electricity connections, and open doors for new revenue sources similar to telecom and digital service providers. For example, smart meters can be made to work in two separate modes: ‘home’ and ‘away’ mode. When a customer is not at home for a long time, he or she can choose a few appliances that should continue to run and key in a password that will make the meter work in ‘away’ mode. When there is an unexpected usage or a wrong password keyed to activate an appliance, the smart meter can send a text message to the consumer.
Puerto Rico started the smart meter roll-out along with installation of a 1GW renewable energy with an aim to reduce its dependency on expensive oil-fired grid. The smart meter pilot, in conjunction with other measures, has already saved USD 17 million per year in reduced electricity theft within the first three years. This figure is expected to rise to USD 50 million when the rollout is expanded.
Rewards and loyalty
With real-time data feeds from smart meters, utilities can roll out several reward and loyalty programs similar to that offered by retail chains and airlines. Reward allocation done in real-time provides an opportunity to make these programs extremely engaging and generate a higher level of participation. These programs can be broadly classified into two sets. While first set of offers would correspond to incentives rolled-out to promote loyalty while driving an appropriate behavior, the second set corresponds to collaborative partnership with other industries to capture cross-sell opportunities.
Incentive roll-out to promote loyalty and ideal behavior
By integrating behavioural economics, gamification and loyalty programs, utilities can rollout multiple loyalty programs and incentive schemes to keep customers engaged and alter their consumption patterns. Advanced marketing mix models (MMM) can be deployed to estimate and track the impact of various rollouts and optimize promotional tactics with respect to profit. Since these programs will involve investments from utilities, a test-andlearn process, where RoI is evaluated for multiple variants of programs rolled out in small batches is recommended to decide on the optimal scheme for a large scale rollout. The deployment of region based gamification based on rewarding efficient household consumption levels during peak hours can serve as a useful tool to enhance retention by keeping customers engaged while curtailing peak loads.
A pilot program initiated by National Grid in New York and Rhode Island offers loyalty points to customers based on the amount of energy saved. Rewards can be redeemed at Home Depot, Amazon. com etc. or donated to charities. Since the launch of the pilot in 2009, National Grid customers in those states have saved - $73.7M and 800M kilowatt-hours of electricity.
In order to fully leverage the advantages of smart meters, utilities should explore partnerships with other industries. While a partnership with a telecom company can be extremely useful to set up the process and get devices connected to smart meter, partnership with other industry players can help rollout advanced gamification along with loyalty programs.
Utilities usage data at appliance level is a useful asset that can provide insights for customer segmentation and optimal marketing channel. Hence, it should be utilized as a potential revenue source while working with other industries. As an example discounted IoT (Internet of Things) connections and no network charges for accessing smart consumption on smartphone apps, in lieu of customer segmentation insights can be considered as potential collaboration terms with a telecom player.
Monetization of increased efficiency gains Providers should aim to make the best use of the AMI (Advanced Metering Infrastructure) data to increase efficiencies at all levels of services and offerings to the customer. A close examination of different operational elements in the smart grid will open avenues for growth in top line and bottom line for utilities.
Customer segmentation based propositions
While the AMI data opens up many possibilities, a key area where utilities need to work is improving customer segmentation based on actual consumption data from smart meters. With the roll-out of smart meters, the use of data mining to generate customer segments to increase marketing effectiveness and boost potential RoI can be refined significantly. Segments generated through predictive analytics programs can be used to generate target messages with high precision.
Further, a traditional customer’s lifetime value can be fine-tuned to look into cost to serve and revenue at a far more granular level (e.g. peak vs. off peak, hourly, daily etc.) as compared to the current monthly level approach which can lead to inaccurate classification and segmentations. Accurate definition of customer’s lifetime values can help prioritize, target and acquire the most valuable customers first while deprioritizing the ones who might have a low or negative lifetime value.
With more accurate segments, several manual activities can be automated to improve cost efficiency, performance levels and accuracy. For example, the current process to identify customers likely to experience a breakdown is based the discretion of call center agents. This can be replaced with an auto alert system for specific device failures in a household that trigger home emergencies. In order to realize these benefits, utilities should be prepared to leverage analytics by developing:
- Accurate load curves for every customer, thus segmenting them based on:
- + Customer attributes
- + Energy consumption pattern
- + Areas triggering peak load at grid level
- + Areas of energy efficiency
- + Renewable energy usage
- The capability to process data at extremely high velocities in order to quickly respond to pricing signals and identify the set of customers who can be targeted for energy efficiency programs
Demand response management
When a production facility is not used, it represents less efficient use of capital. However, utilities need to plan their capacity to be able to meet the peak demand with a sufficient buffer to deal with unanticipated events. Hence, any opportunity to shed peak demand can help release significant capital for utilities. It is estimated that a 5% lowering of peak electricity demand would result in a 50% price reduction2 to the end consumers.
The dynamic Time of usage (ToU) pricing, rewards and incentive mechanisms covered in earlier sections are some of the methods which can be deployed for demand response. Further, several new technologies are available to automate the process of demand response. Such technologies forecast the need for load shedding, communicate the demand to participating users, automate load shedding, and verify compliance with demand-response programs. Utilities can automate appliances connected to its users that can reduce consumption at times of peak demand by delaying draw marginally, such as turning up refrigeration and lowering the temperature of hot water during peak hours. Such programs have a considerable scope to reduce peak demands. Utilities can substantially cut costs through these schemes, and then pass on some of these benefits to participating customer based on megawatt power.
The Fayetteville Public Works Commission, the largest municipal electric provider in the state of North Carolina, has rolled out demand response/home energy management service commercially. The demand response service has installed smart meters with an integrated gateway module and a programmable networked thermostat in homes and small business premises. Initial results show that consumers have saved as much as 15-20% of their overall electricity usage compared to previous years.
A similar advantage for large scale customers with smart meters and generation capacity could be the ability to closely monitor, shift, and balance load in a way that allows them to trade what they have saved in an energy market. This will involve sophisticated energy management systems, incentives, and a viable trading market.
Additional opportunities across the utilities sector Regulatory bodies across countries want to know how investments in smart meters are helping improve operational efficiencies and deliver enhanced levels of customer service. AMI data and the overall adoption of smart propositions are expected to drive the benefits at a sector level by closing gaps between regulatory bodies and customers, along with incentivising the use of alternate energy.
Managing demand with analytics
Smart propositions will facilitate a higher adoption of distributed energy systems enabling consumers to generate on-premise energy that can be fed back into the distribution grid. Distributed energy resources often use renewable energy (RE) sources, including biomass, biogas, solar power, wind power and geothermal power, enabling access to cleaner energy on the grid.
Smart meters can provide near-instant data on supply and demand levels. By combining this information with real-time markets, an analytics engine can make the grid intelligent enough to manage loads and provide convenient reconciliation of anything that is produced and consumed anywhere.
This will enable the distributed energy to be precisely billed, benefiting customers for every unit of energy produced in their premises. Utilities can better manage peak loads by having access to multiple energy sources with reduced carbon foot print for the grid. However, due to a higher variability in generation versus demand, dynamic price negotiations within the grid would be required, necessitating demand response capabilities. Energy storage will become quite critical due to the variability of renewable sources.
Predictive analytics can be applied to the problem of energy storage to forecast demand spikes and optimize energy storage and distribution systems for renewable sources. Combined resources can then be used for managing demand response or determining how surplus energy can be traded in a broader electricity market. Analytics can also forecast how energy storage systems are used on a daily basis so the systems can be properly sized for a building’s energy demand, thus avoiding any underutilized storage capacity.
Analytics will act as an enabler for better response mechanisms through integrating and reconciliation of various data sources:
- Accurate forecasting: Widespread instrumentation and advanced computer models allow system operators to better predict and manage renewable energy variability and uncertainty.
- Smart inverters with auto-switch: Inverters and other power electronics can provide control to system operators to automatically provide some level of grid support. Auto-switching between sources can be modelled on historic consumption trends and weather forecast, such as switching to solar on sunny days or wind turbines on windy days to help meet the peak demand from alternate sources.
- Integrated storage: Smart storage can help reduce short-term variations in renewable output, and also manage mismatches in supply and demand.
- Real-time system management: Instrumentation and control equipment across the transmission and distributions networks will allow system operators to have real-time awareness of system conditions. This will also enable the ability to actively manage grid behaviour, as well as identify and resolve losses and theft.
- Distribution network planning: Combining data from grid meters, smart meters and inline sensors along with geographical data can provide a real-time network plot illustrating key line parameters including voltage, real and reactive power, percentage loading and other variables. This enables better monitoring, reliable system operation and better customer service in the form of faster outage restoration and automated alerts. This will eventually lead to fewer customer contacts and complaints.
- Differential pricing basis generation mode: It is imperative that renewable energy sources have different pricing per unit. This depends on the number of sources and their mode of operation, which is why it is critical to have differential pricing. Analytical systems will enable establishing a weight-based model for this differential pricing after due consideration of all factors. Such pricing models will incentivize more consumers to actively participate in distributed generation, and eventually make it easier for utilities to reduce dependency on traditional sources
Gapa Island in South Korea is an example of self-sufficient renewable energy deployment in smart grid. The island, with a size of 8.5km² and a resident population of 281 in 2012, had wind and solar generation systems of 500 kW and 111 kW, respectively, complimented by a 1-MW lithium-ion battery. This in turn replaced a 450-KW diesel generator. This combination has made Gapa a carbon-free electricity system. The project benefits are estimated at $415,000 avoided fuel costs and an annual reduction of more than 750 tons of CO2 emissions.
Information flow across the industry value chain
Smart metering and connected homes will empower customers, but only if they are kept engaged. Also, regulatory pressure is mounting for utilities to enhance customer experiences by keeping things simple without compromising on adopting technology. High levels of stakeholder engagements will require delivering seamless experiences across utility channels by providing choice, consistency, context and continuity for everyone.
Analytics can enhance transparency, communication and accessibility among various stakeholders in the utility chain:
- Faster communication of breakdown and restoration: Utilities can determine the location of breakdowns using GPS coordinates, along with which crew and equipment is well-suited for the necessary repairs. This information could be shared with its service team. This type of information and analysis enables utility to accurately provide an estimated time of restoration to the customers impacted, often even before they realize there is a problem.
- Complaints/query turnaround time: Customers can receive auto alerts of a breakdown to update them of the expected resolution time, effort and grievance compensation for similar breakdowns.
- Tariffs propositions transparency: Details corresponding to cost for customers of competitors with similar consumption levels will be available and accessible across the board for tariff quotes.
- Improved safety: Real-time analysis of customer usage enables detection on unusual spikes that indicate safety risk, thus allowing quick identification and action. Effective monitoring and proactive maintenance of utilities assets is made possible with predictive analytics models incorporating the make, maintenance schedules and energy usage of assets in the grid.
- Vulnerable customers: Superior customer segmentation can be facilitated by smart data to enable accurate identification of vulnerable customers and proactively address their concerns, suggesting suitable devices for their individual needs. This would also help providers in proactively addressing the targets set by regulators regarding social obligations.
- Governance reports: Smart metering data enables the automated preparation of detailed governance reports. By employing suitable data architecture for the smart meter data, aggregate figures can be pulled automatically, eliminating manual effort for creating such detailed reports.
- Fines and penalty reduction: Timely and proactively addressing issues, enabled by data-driven insights from smart meter, will bring down the number of escalated issues and regulatory interventions. This reduces the cost of fines and penalties.
To conclude, the advent of smart meters provides significant opportunities for utilities stakeholders and other industries through effective, collaborative partnership. However, in order to realize these benefits, there are a few challenges which utilities players need to successfully address.
The potential benefits for successfully implementing smart meter programs significantly outweigh these challenges. Be it enhanced customer experience, superior financial performance or better industry governance, assessing the scale of these opportunities helps utilize smart meters more effectively.