Time of Use Algorithm Fix & Monetization
The refined TOU algorithm optimized thermostat pre-heating and cooling, leading up to a 20% reduction in energy usage
The Problem
The Time of Use (TOU) algorithm was underperforming and a developer discovered a bug which resulted in suboptimal energy savings. I was tasked with heading up the fix as the TOU algorithm was relativley outdated and changing it involved product facing decisions. The TOU algorithm was also black box to many product, marketing, and sales team members so I worked with the ML team to put together a info doc that circulated to utility partners and within the team. Additionally with a sales lead we identified an opportunity to pilot a TOU monetization program with a utility partner, a new oppoerunity for energy services
My Contributions
- Discovery - Conducted a detailed analysis of user data to identify patterns and pinpoint areas for improvement in the TOU optimization. Looked into ways to provide more accurate TOU rates
- Technical Scoping - Led the project to fix the bug and update the TOU algorithm, ensuring it aligned with our energy efficiency goals and user needs. Collaborated with the ML team to enhance the algorithm’s functionality and accuracy
- Implementation and Testing - Led the implementation of the updated TOU algorithm, ensuring seamless integration with existing systems. Conducted extensive testing to validate the algorithm’s performance, focusing on accuracy and reliability
- Internal and External Communication - Developed an informational document with the ML team to explain the TOU algorithm, which was circulated to utility partners and internal teams
- Monetization - Worked with the sales team to develop a pilot program for TOU monetization, collaborating with utility partners to define the program scope and requirements. Conducted market research to identify potential revenue streams and growth opportunities
About ecobee
For ecobee owners whose retail electricity rate varies by hour of the day, the TOU algorithm shifts energy use from high price hours to lower price hours while maintaining the desired comfort levels through customized pre-cooling and temporary temperature setbacks (*The above link is not my research but is related to TOU)
Context
As part of ecobee's initiative to optimize energy usage for our customers, we focused on fixing and enhancing the Time of Use (TOU) algorithm. This algorithm aimed to adjust thermostat settings based on peak and off-peak energy pricing, thereby reducing overall energy consumption
Problem
A developer discovered a bug that resulted in suboptimal energy savings. Additionally, the algorithm was outdated and its functionality was a "black box" to many members of the product, marketing, and sales teams. This lack of understanding hindered effective communication and utilization of the algorithm’s potential
Opportunities
By fixing the bug and updating the TOU algorithm, we had the opportunity to significantly enhance energy savings for users, improve internal team knowledge and communication, and explore new revenue streams through a TOU monetization program with utility partners. Additionally the dedicated focus on TOU allowed me to gather data reincorcing the benefit of gettng more users on TOU rates as a business goal
Action
We fixed the bug this allowed me to highlight the issues with user facing product dependancies that are soley known by developers. As a personal goal of being a technical PM I worked with a dev to create an educational document which in turn allowed the teams to have more discussions about TOU impact and the market shift towards real time pricing. Additionally I worked with the sales team to develop a pilot program for TOU monetization, collaborating with utility partners to define the program scope and requirements