Yes, its a complain, but not a rant, which I’ve been postponing to talk about far too long.
We live in Quebec, Canada. Here winters are extreme and houses are mainly heated with electricity for that it is the cheapest in North America and is a very reliable heating system.
In the “Trends” tab, there is the “Compare” card.
The issue here is that this comparison is absolutely unfair because it takes into account all Sense users, which I believe are mostly in the southern region, who either are in a warmer climate or use other kind of technology to heat their houses.
Compared to other houses in Quebec, we believe to be actually in the very average range or energy usage, maybe on the lower side since we also have a fireplace.
The issue is that this “Compare” card should be a helpful tool for me to check how my consumption is doing compared to others, but as it is, it keeps giving me the feeling that I am an irresponsible person.
It’s pretty much doing the opposite: it’s making me feel bad.
So, could you please adjust this card to do comparisons with houses that have similar characteristics?
The Compare feature is using Sense statistics from three regions
Similar homes in your area
Sense homes in “your State”
All Sense users
If you want “a fair” comparison (in your terms) you need to consider only #1 … however, the Sense user-base in your area is probably quite low.
Personally, I wouldn’t worry about using the Compare feature in your case since it will be some time before there are enough Sense users in your immediate area for it to be a statistically meaningful metric.
Use Sense, now, for what it’s great at: Finding variations in your home. Regardless of your relative energy usage, Sense is a great way to target and reduce your absolute energy consumption. Over time you will see more relevance in comparing device and overall usage with other Sense users.
This brings up a good question about what other platforms/devices/services can also be used for benchmark comparisons for residences. Ecobee users of course have access to comparisons (but only really for specifically heating/cooling performance) to their user base, although I tend to prefer to use beestat.io for looking at my ecobee data. For comparisons for whole-house energy use, the options seem to be fewer. My utilities offers comparisons to other customers through their web portal, which is somewhat useful considering all other customers are reasonably nearby, but they lack transparency and I question the results myself. Through my day-to-day working in commercial energy management, I’m well familiar with EnergyStar Portfolio Manager for benchmarking energy performance for primarily large commercial/institutional buildings, and they do not offer comparisons or EnergyStar scores for houses.
So I have yet to find a residential-only comparison platform that I believe has both the sufficient accuracy in home details (that affect energy use) provided by users to allow meaningful normalization between various systems choices/setups AND a large enough user base of nearby, or otherwise climate-adjusted users to provide truly meaningful and accurate comparisons.
Has anyone else found something like this that they trust for household energy use comparisons? If so, please share!
I agree with this completely, but also because, I suspect that I am not a typical Sense user, and here’s why. The comparisons for #1 show me doing pretty well (almost in the top performing [lowest wattage] third), but comparisons offered by #2/#3 show me doing much much better (top performing 15%-20% average.) The reason I say I don’t think I’m a typical Sense user, is that:
People with high electricity costs (i.e. relative to other areas in the country) are more likely to be concerned with tracking/lowering usage, and more willing to pay $300/$350 for a device to help them do this. My annual average blended rate in MN is about $0.14/kWh which is far lower than many other areas of the country.
Related to what I just said, people with higher use of electricity (again, resulting in higher cost and greater motivation to use less) are again more likely to buy a Sense. My urban house (~2000 sqft) is relatively modest in size compared to the McMansions that are popular and are more likely to use greater energy (especially in plug loads, assuming greater occupants) as people with larger houses tend to have greater buying power, more likely to have more things plugged in, etc.
Many Sense users also have PV, I do not. Those with PV again I’d argue are more likely to use more energy which helps justify the investment in solar.
Many Sense users also have EVs, we have a single PHEV with a comparably much smaller battery. I admit the usage of our car probably did drive me to be more interested to buy Sense, I’m sure if we had a true EV that driver to want Sense would have only been greater.
So in summary, yes, pay the most attention to #1, but I still think the Sense user base is going to skew naturally toward more energy-intense houses, and if you are not in an energy-intense house, then #2 & #3 are more likely to show that you’re doing much better than you might actually be doing.
You’re right about the skewing, perhaps especially for those on the forum!
I think Sense is as close as it gets regardless of the lack of meaningful/trustworthy whole-house localized usage comparison. My rationale:
Sense’s device technology is local (high quality CTs on a household panel) and of sufficient local processing power and bandwidth (along with the cloud-based backend) that a utility smart-meter does not currently compare.
Comparing house-to-house, even in a highly localized situation, is always going to trend toward the devices vs the house (especially in America in the 21st Century). e.g. Imagine an idealized Sense situation where everybody in a 400-unit New York apartment building has a Sense and they all have the same heat/water/AC supply and major appliances (which they can’t change)! Eventually, with that level of detail over each unit’s “whole house” view, occupants will be comparing how they use (and) what they use … which might lead to gadget reduction. When you consider single-family housing HVAC and PVs and EVs and pools and so on, the “gadgets” are scaled up by orders of magnitude in both positive and negative energy impact … but they are still what actually needs comparison, vs. the whole-house.
Sense will no-doubt mature to make comparisons between similar households with similar devices.
Most people are going to become more-and-more concerned with the privacy issues around energy monitoring and feel more comfortable having control over that data and keeping it separate from their Utility. Prediction: Smart meter hacks used for elections and so on.
Amory Lovins and Vaclav Smil need more consideration. To quote Lovins from 1976, “Energy Strategy: The Road Not Taken?”
This by way of pondering: If I live in a new NetZero house and you live in a 19/20thC energy hog, what are we comparing in terms of energy “consumption”?
I recently went to Wadamanna Power Station, which was built before Kaplan’s turbines made it south. The original two Pelton wheel turbines (huge heavy beasts) generated a total of 7 MW and they were fed by a wood stave pipeline. Electric space heaters weren’t common at that point, one would guess!
As Smil points out though in Growth, the average US passenger vehicle (including EVs!) has increased (linearly) in average power over last 100 years. Why?: weight. Not the least reason being the addition of an average of 30 small electric motors for seat control, power steering and so on that amount to several tens of kg added weight. Cars, like single-family houses have become: huge heavy beasts.
Really? What about for homes that have thermal (or I should say, non-electric in addition to electric) energy use, for which Sense captures exactly zero percent of such usage?
What I’m looking for, is something that can benchmark my house across ALL energy types!
So with moving to put the tech directly in a utility-grade meter, at what point exactly should we worry about Sense selling our data out to our utilities also?
“Benchmark” would imply that you also need to have a way of calibrating the house. I’ve toyed with things like Sefaira and quickly realized that for anything other than near-idealized structures (read: new, tight construction), an expert in field can probably make a better guess. And that’s the *first stage of a multi-layer benchmark.
I’m more in the school of using Sense as a kind of “poor man’s blower door test”. You get a macro view into consumption and then, with luck, can drill down, rather than seeing it the other way around.
Now! But Sense is not selling their data as such, though I think it’s a very fine line when the development of the RNN is dataset dependent.
Trick…ier, I’d say, but not impossible. And yes, I’m definitely implying the need to calibrate the house, or more rather normalize, because that’s exactly what makes such comparisons valuable.
I assume that the Sense stats are already normalizing to some extent:
Similar homes in your area - how do they define “similar”? I assumed they are using the rather limited location and macro characterization information that Sense users self-report in the Home Details settings section… e.g. using the sqft, # occupants, type, age, etc. information to group you with houses of similar construction, occupant heat loads, etc. Now, it doesn’t say similar Sense homes in your area, but to give any remotely accurate comparison, they need to be using some kind of data whether it be energy use intensity (EUI) at even the most basic level. “In your area” normalizes weather variances, etc.
Sense homes in “your State” - Filters would be of course sense user and location.
All Sense users - only filter is sense users.
Of the three, #1should therefore have the most background characteristic information used to normalize.
It shouldn’t be any harder that some energy, in a whole house-all energy type scenario, is non-electric, but yes, you’d have to supply the required information to a model which would then normalize it. This is how energy benchmarking works, and why, for example EnergyStar can certify buildings on their scale of 1-100 and hang a plaque on your building lobby for everyone to see that, for example, if you score a 95 per their algorithms, means your building performs more efficiently than 95% of comparable properties… adjusted for the key energy-sensitive parameters that allow them to define the normalized and anticipated energy use for your building comparable to others of a similar (that key word again) utility.
So, again, if EnergyStar can do it for commercial buildings, then it can be done for residential buildings also. I just don’t know of any comparable effort on the house side.
Not impossible, and in fact potentially trivial(ish), for new and isolated construction of certain types but next to impossible, or at least much worse than +/-5%, for existing structures.
And by way of highlighting the problem (and the reason I quoted Lovins above): Consider the heat gain/loss from building and landscape shadows and compare this to expected PV solar output.
“Passive” solar gain/loss is difficult to measure if buildings aren’t specifically engineered around those gains and losses. “My place cooks on a winters day because I have tons of windows facing south but if it’s cloudy it’s freezing!”
“Active” (PV or thermal) solar gain is easy to measure over time but hard to predict in a truly built environment and is highly dependent on the state of flux in the landscape – “Hey, you just built a building and shaded my PV!”. Meaning: Ever tried a solar calculator for a city location that doesn’t incorporate building shadows? Fairly useless.
There are of course methods to measure old construction and other factors but I don’t see that happening on the necessary scale, and with sufficient accuracy, to be applicable to, say, better household comparisons in Sense. However, you can infer that data from Sense measurement in the same way you can infer the weather from a PV panel’s output. e.g. Look at the Energy Star (normalized) rating of enough hot water tanks vs. actual usage (and usage patterns!) in enough households, and you can probably infer how many occupants are in a particular household … along with how long they shower. “Compared to other Sense users you are taking really long showers”.
The comparison feature as a whole is really a one dimensional comparison that doesn’t provide much actionable information. We constantly rank as high usage in all 3 categories but that doesn’t mean much if anything. ~95% of our electrical usage is off-peak which keeps our bills low, I doubt that our neighbors have their electrical consumption this dialed in. In addition many of our neighbors have gas water heaters and gas fireplaces, our home is all electric which pushes us into the high use category. While I like the idea of the comparison feature it would be more useful if you could customize it or if it could take other measures into consideration. Pure kWh only tells a piece of a larger story.
I live in Kentucky and my comparison also shows me very high versus others. I don’t have any close neighbors with Sense but when we talk about our electric bills, mine is much lower than theirs and homes are comparable.
I don’t even look at the “Compare” tab anymore because, like you, it makes me feel terrible.
Absolutely agree with the response. You are doing the right thing by owning one of these. You are being responsible (to a point). So what if you are using more Kw than a neighbor.
You are at least trying to make a difference, whether it be in your bill or or the environment.
OK @ixu I want to differentiate between energy modeling and energy benchmarking. You’re right that those variables are important to anticipating energy use, but across a wide variety of environments, the variation will naturally give us a bell curve with which to normalize. Yes, some houses will have greater heat gain, others less. Benchmarking is less about anticipating usage and more about simply studying the energy use typical within a very large population to draw conclusions, and normalizing to tell you where you fit within the observed population of data. The EnergyStar Portfolio Manager benchmarking tool for example compares you to a characteristic population and doesn’t bother with the nitty gritty details of modeling too much because those impacts are understood to affect the whole population.
This is a good briefer on the math behind this particular benchmarking analysis:
Totally with you on the modeling vs benchmarking concepts and I have no arguments against the practical, if cautious, application of such indexes.
I’ll use an interesting and recent enough real example of where terse “analysis” vs common sense can get you though:
On the local level, based on “Peers”, to put it in the context of Energy Star and EUI, this couple probably did the calculus and figured that spending $720,000 to renovate a $1.4million Brooklyn brownstone was a no-brainer. They don’t say how much of that cost was converting a decidedly un-NetZero house into an energy efficient one but I’m guessing it was a big chunk of money. The question I’m guessing they may not have asked is "If we do some basic, cheaper, efficiency renovations and sink the big chunk into solar (potentially community or grid solar), OK we’re throwing energy out the windows of a non-NetZero house but actually we’re generating surplus energy!
Thus back to the huge energy monster: The money spent to renovate a house to make it “passive” can surely often be better spent on creating an (my term, maybe) active house. In an active house you acknowledge and weight the realworld cost (including of course embodied energy) of construction or modification of the house itself along with the cost of energy generation both locally (say roof solar) and remotely (community/grid). Sense plugs right in to that philosophy in so far as you can correlate your absolute consumption with your own habits, in many ways regardless of how leaky your windows are and what your EUI might be. Eventually I can see Sense saying something like “You should have shorter showers” because you are way over on the shower bell curve rather than “You should move out” (although that of course is coming!).
Off topic, but I wanted to thank you for the info on beestat. I’ve been an ecobee user for several years, and somehow that service never made it on my radar. Already created an account and syncing my data now. Looks like everything I’ve been missing in ecobee’s data portal.
