Performance of my Sense unit after a year. Poor!

After a year I have given up expecting Sense to reliably detect or identify devices. I use it now as a total power used and compare that to my bill which gives, daily, weekly and monthly use numbers.

The few things it did detect are all either intermittent, partial or were detected and then later lost. Sometimes a devices on indication vs the “bubble” vs the timeline don’t match up.
Example; Sometimes using the microwave for 23 seconds everything will match exactly. Other times I use it for 2 minutes the “bubble” pops up for two minutes but the timeline says it was in use for 10 or 20 minutes. Best I can tell the power use reported was accurate.

I have 15 total smart plugs and several Hue devices. If I were to remove those the numbers you see in the chart below would be even more stark. They account for almost All the detections. My other and always on are consistently in the 60%+ percentile every month.

I know many have had a better experience with Sense but for me not so much!
Just my observation.


Hi @ron111157 ,

Thanks for your assessment. Three thoughts as a fellow user that has also sought to get the unknowns down in my house:

  1. I have similar thoughts on the runtime and times on summaries for each device - the numbers can be wonky. I just avoid those numbers. But Timeline data for my Microwaves seems to be pretty good.
  2. I do believe that Sense should set the expectation that there will be a number of common devices that Sense won’t detect, but that one can judiciously use smart plugs and other integrations to get a much clearer pictures, especially with some of the smart plugs coming in below 10$ a pop (when I first started buying them they were 25$ a pop on sale).
  3. In my books, Other and Always On are very different - lumping them together in your 60% is wrong. You have the tools to itemize most of the Always On, but maybe you need better methodology for doing so ? You do know that Sense has a way to itemize ??

I do know Others & Always on are different but other than all the smart plugs & a couple of somewhat detected items (mini split and window unit) It all I have to go by. Adding all the smart plugs usage amounts to 15 maybe 20%, if that much.

If you mean itemizing by using “Settings, My Home, Device Inventory” I have listed every single electrical item (about 50) I have to include a model and part #.

I also try to add a name to what’s connected on the smart plug.

Sounds like you are doing the right thing. But if you are itemizing most of Always On, then that category of usage should be mostly known as well. It would be great to share a screenshot of your itemization - sounds like you have been really thorough.

ps: I do think that Always On at the house level will exceed the sum of Always On devices by a little due to overlapping, low-level power draws that do go totally off occasionally.

Itemization of Always On is different than the Device Inventory. Read about the Always On itemization at this link: What’s new in v37 (iOS/Android) . Some of the data for that list is pre-filled by Sense, while the rest can be obtained using a traveling smart plug or (sometimes) by looking in the owner’s manual for the device.

1 Like

I’ve been using Sense since 2019 and I agree, I’m disappointed with the device detection accuracy, especially over the last year or so. At least I can reliably use the meter, which is my main focus. It sucks that the device detections no longer show on the Meter view, I really liked that, but since they’re not accurate anymore anyway, I guess it’s not that big of a loss.

I’m not going to write a ticket for everytime I see that my laser printer gets misidentified, or it thinks my stove turns on when the dryer does…it’s just not worth the effort when I know the detections won’t get better.

I’m happy I have the monitor so I can track my overall usage; that’s the electricity I pay for, regardless of what device is using it.



Here’s a partial listing. Thats a big list to get in a snapshot. How much detail do you want?

@ron111157 , whoops I misinterpreted what you were saying. You aren’t doing all you could ! @jefflayman correctly outlines how to take aim at itemizing Always On. The Always On itemization looks like this, and automatically brings in the Always On from all your devices on permanent smart plugs. But you need to fill in all the other devices. That’s what I use a Roaming Smartplug for.

I wish the Home Inventory and Always On were more integrated, but today there’s probably more value in device level Always On identification and itemization than there is in meticulously filling in the whole Sense Home Inventory.

ps: My approach has been to do a superset Home Inventory outside of Sense that contains all the info I need to fill in the Home Inventory, the Always On itemization plus my strategy for covering all the major usage in my house.



I remember when I first got Sense I made that list using manuals, smart plugs and guesstimates. It probably needs an update since the total is about half the estimated unknown. And then the Always On bubble is even higher than the estimated.

Even so doesn’t really help my initial assessment. As the day gets hotter and the Other and Always on bubble grows double or more. Logically that usage is the Mini-split.
Actually I even turn off most of the smart plugs that control unneeded items from 1400-1900.

As I just stated to @jefflayman below I have done all that as well.

It may need a little tweaking after a year in use.

1 Like

@ron111157, Two thoughts:

  1. If you are turning off devices during 5 hours during the day using smart plugs, then your Always On, both at the house level and in the itemizations will likely go down. That’s great.

  2. Seems like half of your Always On is known, so Other Always On is maybe only 14% of your house total usage. But you can go further - a roaming smart-plug enables you to go beyond guesstimates and also cover all the obvious plug-in devices in your house.

@kevin1 , why do you say half? I would figure closer to 90+%

I have just about every item listed I can test. Many of the usual suspects are connected to UPS units but all those UPSs are plugged into a smart plug. The UPS load display closely matches what the Smart plug energy use shows.

I do have a dedicated “Roam” smart plug and it is currently connected to the Fridge since it was never fully “sensed” by the Sense unit.

I’m guessing that Kevin says half because the posted screen shot lists the total of known Always On Devices as 233W while the estimated unknown is 514W. I’m guessing that Ron says 90% because he tracks a different value for Always On, which may not match the point-in-time value of his screen shot.

In my case, my Always On usually reports a larger value in the summer than in other seasons. I believe this is because even if one air conditioner is resting, another is likely working (I have window units). Thus when I think of Always On, I mentally reference the non-summer value.

That’s my thinking exactly @jefflayman. House-level Always On is a top-down view and the itemization is a bottom-up view from each device. @ron111157, has about 50% (really 233/514 = 45%) of the house-level Always On accounted for. But as you suggest, a component of house-level Always On can also emanate from pairs (or more) of devices that have overlapping on cycles - they might turn off sometimes, but the other one is going during the shutdown interval.

The flip side is that there are also a number of wired devices and devices I can’t get to in my house to measure, so that can be part of the gap as well. Things like wired furnaces, ceiling fans, lighting with electronic dimmers, doorbell transformers that are powering Ring an other cameras, etc.

@jefflayman , There is some truth in what you say about both statements.

I don’t have a mental reference value for Always On, Actually I mostly ignore it. I do know it grows quite a bit as the day gets hotter.

I did some updating since that list was a year old. As it stands right now Always On is 350, estimated unknown is 359 and the bubble shows 709.

Your statement about windows units did get me thinking about window units. I only have one but that reminded me it was never fully detected. I only see it in Sense when the compressor is running but it’s always on and the fan running so that may account for some of Always on estimates. Also the same can be said about the Mini-split system to a lesser extent. it is partially detected as well. It is always on and a fan or something going. As it gets hotter such as now @1730 it is 108 outside and the Other bubble is HUGE + - 4000W and climbing. Sense thinks the mini-split is in an off cycle, LOL. The “parts of mini-split” detection when it appears normally runs 200-800 Watts.

1 Like

@kevin1, All of this it interesting and food for thought but is off topic from my original problem. Detecting devices!

I do have one last question.
If I have all my smart plugs listed in The Always On devices and they also are detected ie in a bubble is that not double reporting?

@ron111157 ,
First off, I agree that there are wide range of devices that Sense’s core native detection cannot find, mainly because those devices don’t present any on/off transitions (constant consumption), slow transitions, irregular / inconsistent transitions, or transitions that get lost in the noise.

From what I can tell, you’re doing the right thing to fill in the unknown usage. But I view Always On as an identified class of usage, and you have been able to pin it down more with your itemization. Instead of 60% Other and Always On, you really have 30% Other, 15% Unknown Always On and 15% Fully Identified Always On.

On your question - yes, the Always On is double counted in the Bubbles and totals though the numbers are usually relatively small (Always On for just the devices on smart plugs). I’m not sure if the double counting bubbles through into Other or whether the Always On components get pulled out of the subtraction to calculate Other.


Ok, so after much updating and adding to the Always On list I’m at about 408/260. I have 38 items listed. I am comfortably sure that most of that remaining 260 is the AC window unit fan running all the time and whatever is left running in the Mini-split system. It is one big unit outside and three units inside that are truly never “off”.

1 Like

@ron111157 , sounds like you have tracked all you can. Is the mini-split wired 240V ? And is the AC window unit plugged 240V ? Just trying to think through solutions for those. Right now, the biggest gap in Sense’s hybrid coverage strategy is the paucity of 240V sensing devices. The additional Flex Sensor and direct circuit measurement (DCM) is useful for those who want to look at 1-2 challenging 120V or 240V circuits in their house, but is a little pricey (25$ / per circuit), and isn’t useful for people who are already using the Flex sensor (second Sense port) for Solar, 400A or a Generator.

I do have a second Sense unit that I use solely for the most problematic 240V circuits in my house. Wouldn’t recommend from a cost efficiency perspective, but it is useful since I have solar (the Flex Sensor on my main unit is I use for solar), but it is very useful for me. Just like smart plugs, DCM breaks out Always On for the 240V devices plus provides Device Power Meters.

Good luck !


The Window Unit is a small 120V unit.
The Mini-split is a Mitsubishi 240V unit.

Actually I do have a second set of Flex sensors. When I bought the Sense I got the solar option thinking I might add solar power at a later date. Still in my future plan.
I guess I could put it on that AC 240V circuit until I do.

1 Like