Skip to content
Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

CalTRACK Executive Summary


CalTRACK is a set of methods for calculating site-based, weather-normalized, metered energy savings from an existing conditions baseline and applied to single family residential retrofits using data from utility meters.


CalTRACK can be used by Program Administrators or third party implementers for managing energy efficiency programs. CalTRACK also supports Pay-for-Performance programs by tracking metered savings using calculations that are transparent and replicable. When both parties - purchaser and vendor - use the same standardized set of methods for calculating energy savings, a mature and robust energy efficiency market is possible.

CalTRACK attempts to comply with AB-802 (Williams, 2015) and SB-350 (de León, 2015).

CalTRACK calculates the whole-building site-based savings that result from any mix of measures, building types, and consumer behavior. CalTRACK does not estimate the amount of savings that can be attributed to a particular measure.

CalTRACK does not take the place of program evaluation. Calculating site-based, weather normalized metered energy savings is often a first step in a program evaluation, but a full evaluation that seeks to control for other effects, such as exogenous factors like economic growth or technology adoption or endogenous factors like socioeconomic status would be expected to consider additional variables.

These methods focus specifically on calculating site-based, weather-normalized metered energy savings for determining payments under a Pay-for-Performance program, rather than the more evaluation oriented guidance that is found in ASHRAE Guideline 14 or the Uniform Methods Project (UMP). CalTRACK methods do not require energy consumption data from a population of energy users beyond the treatment group.

CalTRACK allows third-party energy efficiency program implementers to conduct energy savings analysis on their own customers themselves, provided they are given access to their customers’ energy consumption data via Green Button or other means. This ability to track program performance in real time gives implementers an essential tool to monitor their performance and adjust their implementation practices. It also allows aggregators to quantify their expected yields for Pay-for-Performance programs.


The CalTRACK methods were informed by and developed in concert with empirical testing by members of the CalTRACK technical Working Group listed below.

The empirical testing and refinement of the methods was performed using historical program data supplied by PG&E from an existing home upgrade program and included data from 4,777 homes that were retrofitted during 2014 and 2015. The CalTRACK methods include instructions for data cleaning and formatting as well as a two-stage model for calculating savings and aggregating site based savings into portfolios. The results of the methods testing as well as the deliberations of the working group are documented in GitHub.

The CalTRACK technical working group’s empirical testing of these methods is archived in a Github repository. The repository contains the results of a variety of tests related to the methods choices. In general, the group limited its discussions to focus on specific technical issues rather than on broader, policy-oriented issues. These discussions can be found in the "Issues" section of the CalTRACK Github repository. While summary statistics are presented for both monthly and daily analysis, the intent was to use testing to inform methods guidance rather than to provide for a software equivalency testing process.


The technical working group arrived at two sets of methods specifications, the first for calculating savings using monthly data, and the second for calculating savings using daily data. These methods are referred to as Monthly and Daily, respectively, and have each been tagged with a version number of 1.0.

The purpose of versioning the CalTRACK methods specification that informed the initial guidance is to lay the foundation for further efforts to refine and expand upon these methods. We invite collaboration from stakeholders to perform their own empirical tests of these methods and to offer suggestions for improvement, ideas for how to handle edge cases, or provide further tests to identify potential issues.

We have taken this first step towards an empirically-informed, professionally-tested, and transparent set of methods. We encourage others to work together to continue to make progress.

CalTRACK Working Group

Leif Magnuson - PG&E Project Lead

Matt Golden - Open Energy Efficiency*

Matt Gee - Open Energy Efficiency*

McGee Young - Open Energy Efficiency*

Ken Agnew – DNV GL*

Jarred Metoyer – DNV GL*

Jonathan Farland – DNV GL*

Ben Polly - NREL

Brian A. Smith - PG&E

Charlene Chi-Johnston - PG&E

Cynthia Swaim - Sempra Utilities​

Denise Parker - SoCal Edison

Gamaliel Lodge - Optimiser Energy

Jake Oster - EnergySavvy*

John Backus Mayes – EnergySavvy*

Blake Hough – EnergySavvy*

Lisa Schmidt - Home Energy Analyzer

Martha Brook - CEC

Richard Ridge, PhD - PG&E

Robert Hansen - CPUC

Ryan Bullard - SoCal Edison

Torsten Glidden - Build it Green

Alfredo Gutierrez - ICF

Able Gomez - RHA

*Open Energy Efficiency, DNV GL, and EnergySavvy participated in methods development and testing for both monthly and daily methods.


Shared repository for documentation and testing of CalTRACK methods






No releases published


No packages published