Last Sunday I installed a Sense device in small office panel (3-phase - I picked 2 cables at random for the CT clamps knowing I wouldn’t get the full picture). 4 days later I already have 6 devices detected.
Back in May I installed Sense on a sub-panel at my home (main breaker panel has bus bars from the meter so no way to attach CT clamps). 4 months later I have 7 devices detected (only 2 devices are on the main panel so not a factor).
Having been an avid follower of the forum for 4 months, I know that there are many varied experiences and levels of success with Sense. I also know that the machine-learning component is central to the way Sense works, and I am good with that. I am not looking to bring up the “why can’t we train this?” conversation.
That said, I think this a worthy discussion. A noisy office (printers, HVAC, monitors, office kitchen, servers, printers, you name it) on the one hand; and a house with regular usage patterns and - presumably - common devices on the other hand. Or is it the sub-panel installation, or the single phase, or the very old wiring in my house (no ground on many outlets and sometimes no neutral in the wall receptacle)?
While I am fully aware that Sense would not want to eliminate potential customers on the basis of “you may have a less than optimal experience because you have x and not y”, it may be cause for some recognition of optimal conditions.