My 30 day view on Sense, the company

Getting Started
1

I installed my Sense a little over a month ago. I love the idea and I love the product, but I think there’s a few places the company needs to up their game in order to make it to the Series C / Series D type company.

For context, I come from a tech company finance / VC background, so my comments are focused from that angle.

Sense has positioned itself as the “high end” energy monitor because of the machine learning aspect. This is absolutely the right call based on the technology. However, there are several aspects of the customer experience that don’t align with the high-end expectations that come from a $350 energy monitor. I think this is reflected in some of the frustrations with device detection both in this forum and on Amazon.

Part of the challenge is that customers form their opinion of a product (particularly high-end products) within the first 24-48 hours of owning it. Improving Amazon reviews likely means improving the customer experience withing the first few days of owning the device.

First off, what are Sense customers looking for? They want to understand and control their electric bill, usually focused on the following three areas:

  1. Understanding and changing usage of the major energy users in the house (HVAC, hot water, etc.)
  2. Eliminating the “Always-On” usage
  3. Getting an immediate read on some new big shiny purchase (electric car, pool, heat pump, etc).

Sense does a really good job of identifying the “Always-On” usage. I think you nailed it with that category. It is easy to understand what the always-on load is within 48 hours, and take some actionable items to reduce it.

Identifying the other big usage loads is a challenge. A month in, I still don’t have a good detection on my AC (easily 30%+ of my usage in the summer) and only a partial detection on my dryer (I would guess 5-10% of my usage). While I’d love for machine learning to pick these up, I just care about getting them detected.

Recommendation

-Set a corporate goal / KPI that customers should be able to detect X% of their energy usage within the first 48 hours of owning their Sense. This is key to the positive first impression.

How to achieve this
-Provide a harness with multiple sets of amp clamps. Maybe 4ish. These can be used for Solar, multi-panel setups, or those big 240v breakers dedicated to individual devices.
-Package the Sense with a few Kasa smart plugs to handle those devices you acknowledge are hard to detect. Kasa will probably agree to provide these nearly free if you can show them how much money people spend on smart plugs after they purchase their first one.

Other Recommendation
-I think the B2C model has been a great place to start for you guys. The Sense for utilities also looks interesting. However, I would seriously look at whether channel-partner relationships make sense. This might be more of a problem to look at in a Series C/Series D funding type scenario, but my gut feel is that most consumers make decisions about electrical equipment only when they’re looking at big electrical upgrades to the house.

For example, most solar installers include some level of solar monitoring with the install. Why shouldn’t they include a better Sense?

I bet 90%+ of the people getting electric car chargers installed would be interested in home energy monitoring installed at the same time (as long as the car charger is picked up by the Sense immediately).

Homeowners are probably also very amenable to a home energy monitor sale when they’re replacing major HVAC components. This also assumes the new device will be detected immediately.

You should be talking to the manufacturers of these major devices and installers. You may have to do some direct integrations with those manufacturers products, but you’re getting a built-in salesforce when you do it. You’ll also be putting Sense in front of people at the moment they’re most likely to buy home-energy monitoring.

3 Likes
2

Excellent assessment. Well thought out, objective and incisive. I’ll add a few thoughts as a 4 year user Sense:

  • There already is one channel partner / investor relationship. One of the venture funders of Sense is Schneider Electric / Square D. They offer a rebranded (green) version of the Sense monitor and service under their “Wiser” brand. More is probably needed, but I wouldn’t be surprised if there were future more integrated offerings underway.

  • Sense is the survivor in the machine learning approach. When I first bought Sense, I evaluated several other products that also claimed to do machine learning for detection. Sense is the only one that seems to have had the staying power to make it work to the extent it does today - how much of a home’s load is disaggregated / detected natively by Sense varies tremendously. The flip side is that Sense has learned that straight machine learning isn’t sufficient give the wide range of power usage waveforms - they have had to veer into a hybrid approach with some device-level metering as well.

  • I like the idea a of a detection % KPI, but I believe the 48 hour window would be counter-productive for a couple reasons. Houses and device populations of homes are very different. And many of the paths to higher detection rates rely on the customer’s willingness to avail themselves of various integrations. For instance, I have 2 Senses for my house and can see 90% of the total usage broken out by device.

  • The current Sense monitor is probably not up to the task of collecting additional channels of high sample rate current data. Read the Sense Blog on Jonah and follow through to Jonah’s own blog and you will understand. I do think Sense is working behind the scenes to extend their integration model so they can pull in more “ground truth” data from additional devices. And I wouldn’t be surprised if they are also upgrading their data-pipelines to better reliably handle and manage mixed high rate data sampling (million of samples per second) and low rate sampling (once per two seconds - Kasa or even once per 5 minutes - Ecobee), because that’s what they have learned is required to increase that %.

  • I’m finding that the combination of Sense(2), Tesla(2), Nuvent(4) and Ecobee(2) integrations in Home Assistant is incredibly helpful in understanding complete home energy usage. Each smart device supplies part of the times series data, but not all of it.

2 Likes
3

I’m not really in agreement with @kevin1 as to his position on the post. I find the lack of research and public posting to be out of ordinary from all the VC’s I’ve sat in front of (not claiming you represent one, but based on your “flex” in the second paragraph). This seems like something you would pen to the C suite if you really wanted to, and not direct it towards the community. But to each their own…

This has been discussed many times and most everyone agrees that it would be nice to have.

Flat out impossible. The first 48 hours is the initial calibration. I would strongly advise against this because you would have more upset customers due to false advertising.

As @kevin1 mentioned they have partnered with, and received a round with, one of the largest Energy Tech companies in the world.

If you look in the forums here you’ll find the offers of free Kasa plugs from Sense employees.

Most of the Amazon reviews that are negative are because people are afraid to keep the device beyond the return policy or because it didn’t detect 100% of their electrical usage.

This person left one star “I had such great hopes for this. Unfortunately, after just over 2 weeks, it has only learned 14 devices” … Well, I wish I had 14 devices natively detected after a few months. I think I have 5 now!

If you are familiar with investments then you would know that, with the rounds they have raised, they have corporate goals, KPIs etc etc. We just don’t know what they are, and we don’t need to know if they care to keep that to themselves. From my experience, it’s a nightmare dealing with VC’s and their arm chair quarterbacks. But I can guarantee you with the money they raised, the 14 investors and Series B completed their growth metrics are being discussed daily.

I’m no fanboy of Sense. I think there’s a lot to be desired, but I don’t think it’s fair to publicly say they need to up their game for more money… Personally, I think their biggest problem is marketing and communication along with phone based tech support. Setting expectations and failure to followup - provide accessible support, leaves people wondering if they have a product they can count on. Telling people, as you suggest, that it WILL do something, when it might not because of the inherent nature of AI and detection algorithms, is in my opinion, problematic. When you use buzz words like “artificial intelligence” with people who don’t understand it, it eludes to a grander, better product. For some of us, we might be able to do electrical work ourselves, but for others, they would kill themselves. For some of us we might be able to train Sense, but for others, they would make a mess out of a good thing. It’s not easy for everyone to understand the technology behind AI and what Sense does. Their marketing team has a tough job and I think they’re a bit behind the ball. The Amazon reviews speak to the issue clearly. People expect this thing to be a magical gizmo that will immediately lower their bill and tell them when their coffee is done. Sense is a tool that can aid you. The people who have got the most out of Sense are the people who integrate it with HA or other platforms, the people who want to learn how Sense works. But I’m just another forum member so I try to help people along the way, with the little knowledge that I posses, and I’ll leave Sense to run their business.

I refer you to this Haiku dated from 2018. It’s all about the delivery :slight_smile:

They may have things up their sleeve that we just don’t know about yet. I would put money on it that they do.

2 Likes
4

Brian… Love the over view. I read your summary seeing you on Shark Tank sitting next to Kevin. There is a smile on my face as I say this. Every one of the episodes has a lesson to be learned either by the presenter or the panel. I too have been disappointed by the product. I had a 1 inch pipe burst out in the pasture and the well ran full blast for nearly 24 hours before it was discovered. But, my fault.

A cool feature would be to have …from Schneider… would be to just swap out a breaker in the panel that could communicate directly with Sense. 110 or 220… This is going to create a huge logistics nightmare. All the different panels and breakers.

The real question that should be asked at the end of the day… Would I buy another one or refer to a good friend to buy…

Also, my new hot water heater has its own energy monitor built it.

There are no ‘smart plugs’ here. And 6 months ago I pealed off 60-70 percent of my usage off the grid and sense… 3 cents this time… Later…Gerry

1 Like
5

Todd; I have no intention of claiming to be some brilliant investor. I worked at a VC fund over a decade ago and had a number of mid-senior and senior finance roles in the tech startup world. I was simply calling out my background, as a large percentage of the forum members seem to have a technical software or engineering background. My views just come from a different perspective.

I’m coming from a place of wanting Sense to succeed. I really want to like the product, and I want to see them succeed. The data-nerd side of me is impressed. But Gerry hit the nail on the head. Would I recommend the product to a friend? Honestly, I probably wouldn’t.

Like most comments I’ve seen, I want two things from an energy monitor:

  1. The ability to rapidly understand my large energy users (HVAC, dryer, etc).
  2. The ability to understand my vampire load.

While Sense has succeeded on the second count, they have not demonstrated success on the first count. Yet there are a number of energy monitors that can solve this problem for less than half the price (although admittedly these other monitors won’t work well on my house’s wiring setup).

I’m simply saying that to justify a premium price, a hybrid approach is required. I recommend amp clamps for two reasons:

  1. This will allow the monitoring of large 240v circuits that there is no other monitoring solution for.
  2. By directly monitoring a few big dedicated circuits, you can provide immediate insights into a large percentage of usage, giving people an immediate win when they open the box.

I also justify packaging smart plugs with the Sense for a simple reason. I would guess that most people who buy a Sense are more likely to have modern energy efficient appliances that will likely never get detected by a machine learning algorithm. Buying a $350 energy monitor then finding out that it won’t detect the majority of appliances will piss off customers. Simply packaging a few smart plugs with it will demonstrate that a solution is packaged in the box. I also expect Kasa would allow them to be included for next to nothing due to the network effect, so it’s not like this would be a big hit to Sense’s margins.

In my case, I’ve pretty much figured out that Sense will likely never detect my dishwasher, washing machine or refrigerator. I’m also likely to get a heat-pump water heater that will probably never be detected. But I’ll know to the penny how much my ice-dispenser is costing me. Is that interesting? Sure. It’s just not that helpful.

2 Likes
6

Look Brian,

I’m not here to attack you, that would be against the rules of the community and frankly I’ll save the raging for the gaming later tonight. It was the first part of the opening paragraph that kinda made me think “why is that necessary”. Second was the fact that most of what you are asking for / suggesting has been discussed in detail throughout the community. We all miss things, don’t search well enough etc, but the expectation is that you ask questions before making statements. I’m guilty of this as well and I’m sure I’ve been called out, rightfully so.

To an extent I agree with you. I would recommend it to someone like myself, or most of the Techy people on this board. I probably wouldn’t recommend it to my 92yo grandmother or even my parents for that matter.

This is what I was talking about when I mentioned the technology behind the product and getting the message to people so that they can understand why. Sense is using AI. There’s no magical way to diagnose a vampire load, EVEN if you have clamps on every hot in your panel (unless you have a single breaker for every singular item - not realistic). The reality is that no monitor on the market can do this with just CT clamps. Now when you mix in the AI part that means that you bring in the numbers game. You have to have electrical signatures that match what you are asking Sense to detect using the models and predictive algorithms. There’s an assumption, which again I feel is a marketing issue, that Sense will magically detect what it can’t.

As I mentioned above this will not cause everything to be detected. I would argue that the only time you would need clamps is on dedicated breakers, which Sense actually does a decent job of figuring out already. Nevertheless, this has been suggested by myself and many others to turn the middle port into a hub for more CTs. Where this would provide help would be on things like EVs which Sense has issues with.

One of the biggest threads of this community is the pain associated with these plugs and Sense. It’s not Sense’s fault, from the countless hours we have poured into the diagnosis of the issue, but it’s not very easy to solve right now. These plugs are very problematic.

I’ve visited this community 43 days in a row (which is a little bit compared to some). I too want the best for Sense, not because I paid for it, but because I like what they are doing.

2 Likes
7

3 posts were split to a new topic: New to Sense, having trouble getting monitor online