Billing system complexity makes billing errors likely

Changes in rate structures and rate plan assignments require additional quality assurance checks

Need to quickly check bills before sending

Need to reconcile billing input values against source meter data

Delivered parallel billing system to check system-of-record billing system

Reconciled every line item of every bill before being sent to customer

Rate editor enables real-time changes to rate structures and rate plans

Billing inputs reconciled against source meter data and any discrepancies flagged

Rate assignments on every account checked to insure that account qualifies for that rate

Eliminated need for $35 million billing system rewrite

Multiple sources of meter data can lead to inconsistencies

Customers need to see a consistent data view

Billing calculations must produce matching results regardless of meter data source

Meter communication issues create data issues

Delivered parallel meter data management system to check system-of-record

Reconciled all interval data against all daily data

Reconciled all daily and interval data against billing system data

Machine learning module to identify unusual consumption patterns

Eliminated need for $10 million meter data management system re-write

Large intermittent loads such as EV charging can cause distribution transformer overload

Variable distributed generation can cause distribution transformer overload

Installing interval meters on all transformers would be prohibitively expensive

Virtual meters defined for every transformer that aggregate all of the transformer’s end point loads

Aggregation performed on both 15-minute interval and daily data

Alerting and alarming based on transformer rated capacity

Identify problems and take corrective action before transformer failure

Enables proactive action to prevent transformer failure and associated liability

More reliable delivery of power to customers

Electric vehicle (EV) charging impacts the grid causing potentially dangerous transformer overloads

There is no easy way for utilities to identify these locations

Need to know who has electric vehicles so that customer programs can be more targeted

Electric meter data evaluated for back years and current months to identify load profiles indicative of EV charging

Utility was able to make much more accurate forecasts of electric demand for EVs

Distribution grid hot spots identified and distribution transformer failures avoided

Prevented transformer failures by identifying distribution hot spots and proactively upgrading transformers

Utility was able to more aggressively roll out EV and distributed generation programs

Need to analyze aggregate behavior by customer groups

Need a flexible/easy way to define customer groups

Need to correlate behavior with weather, program participation, and other non-billing attributes

Need to aggregate at the scale of entire customer base

Virtual meter capability enables easy definition of customer groups

Aggregation performed on both daily and interval data

Aggregation by rate class, demographics, consumption patterns, program participation, on premise equipment, and building type

Analyze group response to events such as demand response, weather changes, etc.

Deeper insight into customer behavior

Deeper insight into customer program results

Need to monitor PV arrays across entire distribution grid

Need detect equipment problems (especially invertor failures)

Need to check output against equipment ratings

Need to correlate output over all PV arrays to account for cloud cover and weather patterns

Virtual meter capability enables easy definition of PV groups

Analysis performed on both daily and interval data

Analysis by equipment type and program participation

Automated alerting when equipment issues detected

Increased customer satisfaction due to identification and alerting of equipment issues

Increased and more predictable PV output

Effectiveness of utility program difficult to measure and analyze

The scale of large programs make data processing difficult

Need to analyze sub-groups within programs

Need to adjust group definitions based on changing conditions

Virtual meters defined for program participants to analyze aggregate behavior

Each individual customer’s behavior also analyzed

Virtual meters defined for each sub-group within the program

Virtual meters also defined for reference groups that do not participate in the program

Programs can be analyzed at overall program level and sub-groups to identify where program is most effective

Program resources can be allocated based on effectiveness to maximize program impact and results

Forecasting system must run within time constraints

Using major data center, legacy system did not always finish in time

Utility also needed to increase system size and capacity

Utility needed to accurately bid in futures market

Fractal Programming enabled new demand forecasting system to be written and deployed in a single business quarter

Dramatically increased accuracy and number of scenarios that can be run

Uses full customer data set instead of just a data sub-set

System runs in minutes instead of hours

Added accuracy of forecasting 100% of the data, not samples, impacts millions of dollars of futures contracts each year

Key customer relationship are not always obvious – especially when customers have a portfolio of locations and multiple billable entities

Ownership of commercial properties frequently via special purpose vehicles, making it unclear who the portfolio customer relationship is with

Qualification for key account status based on spending levels and program participation can be complicated to identify and track

Implement qualification rules for key accounts

Implement aggregation rules to identify portfolio relationships

Classify every customer at both individual account and aggregate customer portfolio level

Enable utility profitability analysis by key account relationship

Identify key account customers by evaluating total portfolio consumption and behavior

Enables more focused programs by identifying impactful portfolio level decision makers in customer base

Key customer insight data is often located outside of systems-of-record

Valuable customer information is located department-level computers or desktop spreadsheets that are not accessible by systems-of-record

Customer entered and customer maintained data needs to be incorporated into customer insight apps

Public domain information (eg. tax records) must be incorporated into customer insights

Implement continuous real-time data import from systems-of-record

Implement continuous real-time data import from department-level computers and desktop applications

Implement continuous real-time data import from public domain sources

Implement continuous real-time data import directly from customers

Deeper key customer insight

Enables all customer information, regardless of where it resides, to be incorporated into 360 degree view of customer relationship and customer activities

Assigning rates to accounts can be time consuming and error prone

Rate assignment complicated due to wind, solar, EV and other green energy and distributed generation initiatives

Qualification requirements for rates change over time requiring reassessment of all customer accounts

Implement qualification rules for each rate class

Test every rate on every customer account

Enable “what if” scenario analysis for key account customers for qualification for different rate

Forecast customer budget impact and utility revenue changes for new rate assignments

Enable utility profitability analysis by rate plan

Saved half a million dollars in consulting fees for validating rate assignments

100% of customers on correct rate plans

Need to plan different rates across millions of customers every year

Rates complicated due to wind, solar and other green energy and distributed generation initiatives

Costs millions of dollars to simulate rates, create financial forecasts, and estimate customer impact

Customers may get billing surprises when rates structures change

Implement real time rate planning engine

Test every rate, across every customer, using previous bills

Forecast annual revenue to the penny

Eliminate all new rate surprises

Enable utility profitability analysis by rate plan

Saved $2 million paid to rate consultants for single rate plan

Saved $8 million for subsequent rate planning

Data provider service level agreements can be challenging to monitor, analyze and manage

Data service level shortfalls can lead to billing problems with customers

Quick early detection of data provider problems is needed to prevent downstream issues

Parallel meter data management system

Analyze all interval and daily data to locate any missing data elements and test against service level agreement metrics

Reconcile all interval data against all daily data to check for consistency

Machine learning module to identify unusual consumption patterns that can indicate meter reading issues

Increase in bill quality

Reduction is customer billing complaints and issues

Contact information for notification of outages or other problems can be different than billing contact

The scale of customer base can make data management and automated communication challenging

Customers need to be able to update contact information and preferences to track changes in their internal organization

Customer portal that enables customers to update their contact information and preferences

Internal support portal for use by call center and key accounts team to track and update customer contact information

Identification and classification of critical, priority, and key accounts loads for real-time detection of site-specific outages and other issues

Customers proactively notified and updated on issues

Organizational and contact changes continuously tracked and updated

Customers want access to energy and billing data on digital devices

Customers want to do scenario analysis for rate plans

Customers want budget forecasts

Customers want visibility to solar and wind energy usage metrics

Customer want to supplement their data

Parallel customer care and billing system that is accessible from customer’s digital devices

Instant customer visibility to their real time bill

Customer can test different rate plans in real time

Customers have easy access to budget forecasts

Customers can self-service their data needs which frees up key account and call center resources

$10 million development cost savings and customer support cost reduction for customer information portal

Customers increasingly requesting programmatic access to their data

Large customers want convenient access to their entire portfolio of accounts and associated information

An API is necessary in order to make many third party services available to customers

Web services API’s are exposed as components of parallel systems for managing meter data, billing, and customer information

Customers can access the data both via interactive web portals and via programmatic APIs

Multiple data format supported (JSON, XML, CSV) as well as custom data formats

Increased customer satisfaction

No additional load on IT resources

Third party services easier to enable for customers

Weather data is required for many business functions and applications

Consistent weather assumptions must be made across multiple business processes and applications

Required weather data may need to be integrated from multiple data sources

Weather data management system

Integrates data from multiple data sources

Provides data at multiple levels of time granularity (5-minute, hourly, daily, weekly, monthly)

Data accessible via interactive graphical user interface(s) as well as application programming interface(s) API

Increased consistency across business processes

Increased quality and consistency in analysis and forecasting processes and applications

Billing system too complex to easily modify

Needed 360 view of major corporate customers

Give customers ability to self-service their data needs

Check each bill for 100% accuracy before sending

Give key account team flexible customer info

Delivered parallel customer care and billing system costing 1/10th of legacy app

Reconciled every bill before being sent to customer

Delivered customer data portal to all customers

Entire solution in production in 90 days

Gained deeper understanding of largest customers

Key accounts team and call center increased customer satisfaction

Eliminated need for $35 million billing system rewrite