SUMMARY Since 1999, the Residential Energy Efficiency Project, through Green Communities Canada, has been a delivery agent for EnerGuide for Houses and ecoENERGY programs in the Waterloo Region. This research analyzed 10,208 initial and 2,383 final evaluations made by the Residential Energy Efficiency Project between May 1999 and February 2009 to understand what energy efficiency improvements homeowners have made over time. Because of changes to federal and provincial programs, the data are analyzed under four main periods: the EnerGuide program with no financial incentives (1999 to 2003), the EnerGuide program with financial incentives based on performance (2003 to April 2006), no official program (April 2006 to April 2007), and the ecoENERGY program with financial incentives by measure (April 2007 to February 2009). Participating households were divided into cohorts associated with each of these periods. For each of these cohorts, the impact of retrofit changes was assessed as the amount of energy and greenhouse gas emissions saved. It was also assessed whether homeowners treated the ‘house as a system’, that is, whether they made multiple energy efficiency changes (e.g., increasing insulation in the ceiling, foundation or main walls, upgrading windows and doors, air sealing, improving heating, ventilation, air conditioning, or domestic water heating systems) or focused on a single change associated with large energy savings, such as changing their furnace. The findings show that recommended and achieved energy and greenhouse gas emissions savings differed across cohorts, and that achieved savings were generally less than recommended. The final cohort of homeowners who achieved major energy savings was twice as likely (25%) to take the single action of replacing the furnace as during earlier periods. Changes to furnaces were associated with higher levels of energy savings, but overall, 85% of the homes which achieved the highest levels of savings made multiple changes, thereby adopting the ‘house as a system’ approach to energy efficiency improvements. Finally, homeowners were less likely to make improvements to the building envelope, and these improvements were generally less extensive than the identified technical potential for improvement. Copyright © 2011 John Wiley & Sons, Ltd.