Another Topic on Commercial Applications with Home Installations (Rough How-To?)

Share your Projects!
1

Here are my thought’s on the subject that I may expand in the future, but selling this to the demographic that would want it is a hard sell because you have to usually negotiate this with a home builder or electrician, and when they hear you’ve never done a whole home they won’t accept it. I am no expert in the field of Home Assistant but have work in low voltage security and network administration. This is a rough idea on how to get started with this business venture. Warning it’s a long process.

So the way that I look at this is that you’re going to have to have a marketable feature that set you apart from tried and true reliable or known brands. This I think can be two things.

  • Offline Usability
  • Machine Learning or Semi Ambient Intelligence.

One of the best combos out there right now is Josh.ai and Control4. I can’t go into pricing but these systems usually will cost around $35k and up for a whole house with everything included.

The only way to currently beat that with available open source software would be Ada and Almond. This combined with MultiSensor 6 by Aecotec in every room would give you NLP and Room presence detection.

For machine learning you’re going to want to use node-red + induxdb + tensor flow. The reason why you want to use tensorflow is because you can reduce the model to even work on mobile devices to work offline once you’ve found a model that works. You can get fancy with this and offer this as a saas for your customers to improve their suggestions. This would be for ambient intelligent portion of the system. For more information on this search ambient intelligence on scholar.google.com

Products
For product line I would use Sonoff and I would highly recommend you check to make sure which ESP2866/32 software you’ll be using before you start. For nice custom plates I would get someone locally to make something like the Lutron Alisse Series plates for custom hardcoded scenes.

For lights, and decor don’t skimp out. Try to make sure everything will last up to 10 years minimum, 2 years if they are battery operated. For eg Big Ass Fans for your fans to use or Sonoff fan control if you have a custom fan from a homebuilder.

Network
Unified Dream Machine Pro with Multiple APs


For this you’re going to have to make sure your

HomeLab
There are a lot of different variations that a homelab can bring to your customers. You’re going to want to get comfortable with setting up docker instances, saving them and cloning them. Once you’ve got something you like that works together, start integrating it into home assistant. This can include home media downloaders and meta editors, object and facial recognition, ada and almond instances. You can get some more ideas here https://yunohost.org/en/apps?q=%2Fapps

Audio/Visual
I have not done enough research to give an honest opinion but I believe if you want quality this is one of those things you’ll have to get authorized to sell. I personally would stay away from Sonos as they have outdated some products through software updates. I will update as I will be looking into this aspect more.

Intercom Sytstem
Again will have to look into it as you need to make sure it integrates with AMX.

User Roles
You’ll need to have a number of roles including Head of House, House Member, Guest, and Support and Haos. As for how to implement, I am not sure yet.

Server
I would recommend getting a server rack for something like this. With all of the networking and AMX that will be involved you’ll need one anyways.

Final Steps
Once you have a decent prototype working, (yes you’ll have to prototype this, even if you partner with a company.) Create a custom white label ui and if you want to invest, get a custom app $50k-$75k. For the custom ui take into consideration what you want your customer to see and what you would like to see. Be sure to implement this on a private git once you are done if you have not already.

Once your server is complete I would recommend making backups of each software and instance on the main and make an iso of the entire os itself. Therefore this can be used as the install iso for your installers. You do not want to update at all! This should be a set it and forget it sort of thing. If the customer wants to mess around implement the ability to do cron backups on a raid system.

Hopefully this can get lots of people on the path of thinking about installing these in homes. Sorry if I’m being a bit vague in these directions as this could be an entire wiki in itself. I am currently working on this with backend web dev guy, a homebuilder, and a business coach. I’m in the process of getting my ESC-T and ESC-N from CEDIA so I can have a better understanding of smart home design concepts planning. Another thing you’ll probably need is project management skills for managing such a project. Then get your certs with AMX products.

2 Likes
2

Honestly, as much as I like HA, if I had to do this professionally I would stay the hell away from Home Assistant. Because it would be a very hard to manage liability.

The problem with HA is that it is pretty much eternal beta software. There is no stable release channel. Every update can (and usually will) break something, especially in complex setups as you mentioned. Because there is no stable LTS/ESR channel, even security updates have the potential to break other things. You just can’t build a commercial all-in-one offer around a system that you will have to tinker constantly with to keep it running. And locking it onto a fixed version is risky due to security concerns.

One day HA will be there. But that day has not yet come.

4 Likes
3

I completely agree. As it stands it is currently not there. But I think if one was to stick to a few products and keep a strong network like with dream machine you could make a stable release on your own. For now I’m getting training in other products for now but will be trying to develop this on the side.

4

i would also add to this discussion, that even if you get a working release and stick to it, the shortly released security problem with addons and so forth could kill your stable os at a minute. i dont think its that stable to set it as commercial backend - dont get me wrong, im an entusiast and love all the stuff that this makes possible, but without tinkering nearly everyday there would not be a running HA instance in my home

5

Hi, I would tend to disagree that there is absolutely no value in offering services around home assistant right now.

I agree that going after absolute novices that expect everything to be immaculate is probably a bit risky right now, but there is a demographic that I believe can benefit from support in “getting up and running” as well as “Training” where basically I impart my “so-called” experience over the last 5 years working with HA across multiple continents to those that have the appetite for such exciting and powerful/flexible platforms whilst understanding the whole beta caveat…

You an have a look at how I deal with this in an e-commerce environment here: https://iotci.co.uk/product/home-assistant-hub

Also, there are also certain business or specific use cases where you can simply use HA as the OS and it solves a unique problem that someone / a business may have…

I am actually about to deploy the Plant monitoring platform (This is all it does) to a company…

the benefit is they can then keep their office plants alive and notifiy the maintenance company of the status of the plants…

It is all about understanding the customer, their requirements and where you should and (most importantly) should not go… it is then all down to the SLA and support that goes along with the solution and contract… (They don’t want an SLA (which they pay for) then it is the dreaded break/fix agreement which comes at a premium)

but on the whole… “we cant sell services around HA because it is beta” is all down to the individual…as with most things in HA!

Netgear have been using their customers for testing beta releases for years and years and years… :slight_smile:

1 Like
6

I believe the main reason a professional systems integrator wouldn’t employ Home Assistant is a lack of timely technical support.

Imagine your customers contact you with a problem that is due to a bug in a particular integration or possibly within Core itself. There’s no one in this volunteer-based project who is obligated to provide you with a timely resolution.

As a consequence, your customers may have to wait days, even weeks (one Issue I posted has been unaddressed for months), before someone volunteers their time to correct it. Meanwhile, your customers grow impatient and disgruntled with your apparent inability to promptly fix the problem. That’s not the path to building a successful business.

Pros choose suppliers that can support their products in a timely manner. There’s an escalation procedure available to address critical issues. The Home Assistant project isn’t currently equipped to do that.

2 Likes
7

Theres a GPIO issue that has been open for 3 years.

2 Likes
8

Like i said… there is a market for people who have an appetite to get a truly smart home experiencing what smart home actually is as opposed to the not so smart consumer offerings and multitude of interfaces there are out there today.

Every system has patching etc it is the difference between deploying opensource vs licenced software. In my experience, i get better support from good opensource projects than licensed options…

The fact is, that HA is one of the most powerful smarthome platforms available today.

As long as the risks are clearly identified and communicated, there is no reason why installers should not be doing this professionally…

9

Except Home Assistant has a built in time bomb for any unmaintained installation: The database will bloat to the point it fills the drive, or become so slow it will be considered failed. And if it is an SD card based install, by that time the database writes will have likely killed the SD anyway.

Until the database structure is fixed, it can never be used by non-enthusiasts.

10

Can you elaborate? What are the opensource vs licensed products on which you are basing your comparison?

I doubt that’s a metric that your customers even consider when choosing a systems integrator. They want someone to fulfill their requirements and to have whatever it is that you provide to work reliably.

I admire your business ethics but I doubt many people would accept the risk of acquiring a product whose repair might take days or weeks (or even never). I certainly wouldn’t buy any car or major appliance with that kind of risk.

EDIT

From a systems integrator’s perspective, every service call during the initial warranty period is a loss of time and money. Every fix/patch/correction applied is billable time that you can’t collect (during the warranty period). It’s even more expensive if you have to make a site-visit to correct the problem. Basically, you want the product to work reliably so you can use your time to grow your customer base.

Outside of the warranty period, you can obviously bill for your time. However, if your customers need your services frequently (to fix problems) they’re not likely to remain satisfied customers. Dissatisfied customers can’t be used as references and that can impact growing your customer base.

11

Then you have never used top tier corporate support options on commercial offers. I had such an experience myself not long ago, with a bug in a closed source driver used by one of our very high tier clients. The driver manufacturer (very well known graphics chip manufacturer) offers premium support options, which obviously cost a lot of money. But for some clients, it is worth it. Within less than an hour we had a Teams conference up with one of the lead developers and two senior engineers. The bug got fixed in less than 24 hours after the call.

In open source it’s mostly luck. You’re completely dependent on the whim of some (often single person) developer and his mood of the day. More often than not, you’ll just get a shrug, ‘it’s open source, just fix it yourself’. Unless you have a team of developers who have the capacity to actually go ahead and fix these bugs themselves if needed (that’s what large corporations who use OSS typically do), you can’t base a business on that.

Of course commercial software support also varies a lot and you can get pretty crappy support. But unlike OSS projects, you can then take legal action against the supplier.

12

Ok mate, well lets simply agree to disagree and move on and continue the thread for those of us that do want to make a go of it as I believe that this is what the OP created the thread for… your points have been taken and noted for which we are all grateful…

1 Like
13

I am having the exact same argument/discussion with a corporate client, not related to HA at all… however we are going after a consumer market which is something completely different… I have lived in the corporate world for 20 years with my biggest project for which I was responsible, went live on the same day in something like 36 countries for a global tier 1 telco… i have also always had a sideline serving digital products directly to the consumer market… so again, each to their own, based on their experiences and if they feel they can pull it off… i believe that this thread was intended for those of us that do want to make a go of it, so points noted and taken, for which I, as I am sure are those who see this forum, am grateful…

14

I take it that you won’t be elaborating on your experiences concerning “opensource vs licensed products”?

Anyway, whatever those experiences may have been, the fact is that if you encounter a problem with any one of Home Assistant’s official integrations (or in Core), there’s no distinction between hobbyists and systems integrators and everyone gets into the same queue for technical support. As mentioned, the reply/fix time may vary from hours, days, weeks, to never. Anyone who wants to “make a go if it” needs to take this shortcoming into consideration.

15

It all boils down to the type of business you are. Using open source as a vital part of a commercial offer is always a liability. If you are able to technically and legally manage that liability, then great ! If you’re a single person company and you use large and complex unstable OSS code, then that liability could very well kill your business if things go bad.

It’s a strategic choice you have to make and you have to live with the potential consequences of that choice. Because in the end, as cliche as it sounds, you don’t pay for the open source code. There’s no obligation for anyone to fix anything for you, even if your entire business depends on it.

16

Also, I would definitely not be using Sonoff for commercial installations.

I’ve done 2 installations for others (who already had a commercial HA system), and they knew the risks. Also, they are quite tech savvy, so they were OK with that.

But I always make sure things keep working even if HA goes down.

17

Really the more I think about HA should adopt a commercial installers/certified support model. They come up with some kind of certification program you can pay for and get certified as an official HA installer/support technician. You already have a dedicated hardware that your selling which you could resell to a certified installer for cheaper and provide a longer warranty then allow them to resell it. Also come up with a cheaper global dashboard for the certified installer to use the nabucasa platform and allow them to have access to all their clients and provide support to them through remote access (also build in some form of block for remote video access to a clients system for privacy reasons). This could help push a portion of the industry towards homeassistant and away from control4 and other more expensive platforms. Still keep the software free just give people a way to make some money off it and a way to support their clients as a certified installer and give customers that are not tech savvy a way to have a more cost effective solution. This also could bring in more revenue for homeassistant development and even come up with more hardware partners for light switches and maybe a dedicated remote control for the certified installers to sell. Just having a commercial side of this software through certified installer channel could help it grow even more. Just my thoughts I love this platform and have been tinkering with it constantly and it’s really been rock solid for me since day 1.

18

I’ve been using it since late 2018. I can tell you that when something goes wrong with the Supervisor portion of Home Assistant, response time can vary from days to never. This is anathema to a professional installer; you minimally need acknowledgement of the problem and a schedule for a fix (because your paying clients aren’t going to accept ‘Real Soon Now’ excuses). However, the current reality is some reported Supervisor issues don’t even get a reply let alone a resolution.

FWIW, some users recently reported (6 days ago) the inability to update from one patch release to another (the operation is managed by Supervisor and it’s failing) and haven’t received a reply yet (I’m in the same situation). No professional installer would consider a supplier who doesn’t have the staff to even acknowledge a problem exists. The Supervisor piece of the Home Assistant project is understaffed and would require an overhaul before any consideration of a “commercial installers/certified support model”.

19

This is why I have my setup running in a vm with hypervisor level backups as well as the supervisor backups. If the vm fails I just do a vm restore to a earlier date or spin up a new vms and restore from one of the backups the supervisor took.

20

If you restore an instance with a previous version of Supervisor, it will upgrade itself at the first possible opportunity because there’s no control over Supervisor’s automatic upgrades (a long-standing request that’s never been implemented).

BTW, those afflicted with the mentioned bug continue to have functional systems but are blocked from upgrading. It’s fine as-is unless you need to install a patch release or, if they don’t fix the bug by May, when you want to install the May release.