Any of you do HA installs as a business? Do you skip updates?

Installation Home Assistant OS
#1

Because of the covid 19 I am being forced to look for new ways to look for an income since it will be a while before tourism is back full swing in Cancun.

I see potential in getting into the smart home business. Most of what is being offered on the market (sonoff basic, sonoff mini, etc) don’t work well in Mexico because of their size, so I learned how to buy the components and make my own sonoff’s.

HA is the glue that holds everything together. I love HA, specially the price, however not sure what would be the best way/practice to charge people to do their setup. Within the last month we’ve had I think about 3 to 4 updates. One of them broke my Ngrok configuration, and one of the most recent broke the chromecast.

That is why I am wondering if anyone is doing HA setups as a business, and how do they handle it? Do you just get them going, then “freeze” them with no updates? Do you put them on a monthly plan? Do you charge to visit and do upgrades?

2 Likes
Sell Products with Home Assistant
#2

In my experience, you don’t sell the software or hardware. Anyone can go get a quote for hardware and software to do a thing. What you sell and market heavily is the experience, knowledge, and support the user will have by buying said hardware and software from you. If you market yourself and sell a top notch service, it won’t matter what hardware and software combo you sell to them.

3 Likes
#3

I’m not a professional systems integrator but I have tinkered with home automation for almost 15 years. If I were to use Home Assistant in the homes of paying clients (I wouldn’t but that’s a separate discussion) I would definitely confirm all required integrations work properly before using them in a client’s home. Then I would not upgrade anything.

As long as the system works reliably per the client’s original requirements, there’s no need to upgrade it. In fact, the only reason to upgrade it is if one of the integrations fails to work and causes the customer’s system to no longer operate to specification.

You are certainly free to sell additional services to existing clients. However, I wouldn’t push upgrades unless there was some particular new feature that would be of special interest to them. You’re better off convincing them to pay for getting new and useful automations, or integrating more devices and services, than offering software-upgrades as a paid service.

The key thing is to ensure the end-result works exactly as had been agreed upon and continues to work that way, reliably and indefinitely. Don’t forget that every service call you have make during the initial warranty period is money out of your pocket. Reliability equates to happy clients.

3 Likes
#4

I concur with Taras. I work in industrial automation and updates rarely happen and full system upgrades Happen once every 7-10 years. I frequently work with installations that are 30+ year old hardware and a control screen running on windows 2000. Because at the time of specifying the hardware and software, that was not only what was available but what the manufacturer guaranteed would work together. Everything from the hardware major/minor revision to the software major/minor revision down to the recent month patch level. If a new version or even Minor revision or patch came out, we did not change mid-project otherwise we risk unknown compatibility. Often though, the software we spec’d was dictated by the client’s IT department that was supplying the PC hardware and OS version.

Many of those older systems I work on still work and there is no need to update them. Granted they do not connect to cloud services or frequently changing APIs. I can count on one hand the number of times I have had to update software due to a new piece of software the client wanted to use but was not compatible with their current major/minor revision. I can count on a couple fingers the number of times I have actually performed an in-line upgrade of the firmware and software. This was driven by a known bug the client had experienced. I can count on even fewer fingers the number of times I upgraded a system just because a new version was available.

My point is, uptime is everything. It doesn’t matter how new or good a system is, as long as it works and is guaranteed to work. Because downtime costs money, hurts reputation, and potentially becomes a liability.

1 Like
#5

There are a couple of ways you can go about it…

  1. Complete the setup with the current HA version and then disable auto updates. The customer should be happy without anything breaking in the future (unless device API’s change / are disabled etc.) until they decide they want a new device which may not exist in the version of HA that they have installed. In this case you charge to upgrade and add the new device / component. You would then need to work out all breaking changes to their existing components that have taken place between their installed version and the new one. It could be a lot of work this one time, or maybe not, just depending on their setup (ie: I haven’t had breaking changes affect me in a while). One time system downtime.

  2. You set up some sort of plan to keep their system up to date and fix breaking changes as they come. I would still have the auto updates disabled so you can access the situation before each update. Less work at a time (as I said above, some upgrades might not have breaking changes that affect their config at all), however this will result in reoccurring downtime’s each time you update.

#6

Thank you everyone for the great comments. On one end, I was easily able to sudo fix the issue. I am FAR from an expert, so I do much copy and paste from the internet. The original automation.yaml that I used had “auto hide” as true, and seems that function is no longer used. I just commented out that line, and it is all working well again. I didn’t notice any changes with the auto hide commented out, so not sure why it was there on the first place, or if I will see some type of issue because if it down the road, but it all seems well now.

Personally, as soon as I see there is a “new update” I have the tendency to update, mostly assuming I might miss important security updates. That is the reason why I’ve noticed that after updates, often things no longer work.

For the most part I am a believer of if it ain’t broken why fix it, but what about the security issues/patches?

To start with, my target clients would probably be turned off quickly if setting up their smart home/video/alarm system would require any type of monthly payment. Glad to see industry standard does seem to be if not broken don’t fix it. Basically I shouldn’t feel like I would be doing a bad job by leaving them with a stable system with no updates.

Of course the goal would be for my clients to learn about HA themselves and understand the basics. For example, basics = make sure you have a backup, and make sure you TEST all major functionality shortly after the update to make sure all is well :slight_smile:

#7

I would include the Google Drive auto backup add-on as standard and link it to an account for them. That way if something does go bad you know there should be a relatively easy path to fix it.

#8

Warning: off topic… sorry :zipper_mouth_face:

As a side note, I was supposed to go to Cancun back in April but due to the virus and cancelled international travel I’m still in Australia… any idea if I’ll be able to head over there this coming April?

#9

Yes, no I agree. In fact I have that add on myself. I just adjusted it. I was saving only 2 backups, now I am saving 10. I did notice I didn’t get the telegram message after an update, but since we are having internet issues figured it was that. I just let time past, and when I finally figured out I has an issue, i didn’t have a backup without the issue. Thankfully it was a really easy fix

#10

You can’t disable supervisor to auto update… and this has caused some issues in the past. But it is solved in the next update, just have to wait

#11

Consider what warranty and support you will provide.

Also think about how you will provide that warranty or support. If customer calls with issue, will you have access to frontend or access to ssh or config to fix issues if needed?

#12

Kind of missing the point of a large amount of updates. I’m sure the tin foil hat theories that devs just change things for the sake of change are fun, but in order for things to work, you often need an update to fix what the other company broke.

Things that I’ve needed a new pull of HA to fix the integration with my hardware off the top of my head the last year: neeto, ring, samsung, amazon. Your clients definitely wouldn’t be happy clients if you told them that everything worked 6 months ago, so go away.

#13

You missed the point. HA uses publicly available services that change with or without notice. For that, yes you need updates to fix what is no longer compatible. But if all that is being done is using a local device integration like Sonoff with Tasmota, or some other local hardware integration, there’s no need for updates because those devices integrate directly with HA and not through some 3rd party server. And if it works with that version, it will always work with that version because again, local control. The only need for updates at that point are feature updates when something new and better comes along.

That’s what I’m dealing with in industrial automation. 100% pure local control hardware and software. When it’s local control and certified to work, there’s no updates required if it just works. Then years down the road, client wants to do something new their current system can’t do, that’s a feature upgrade. You can’t just update for the sake of updating. There has to be a reason to update and it must be certified to work with the current system setup. No tin foil hats here. I only work with food grade stainless steel process piping.

#14

My bad then, obviously I missed the point. The OP will go get clients in the industrial automation sector, who want a smarthome that will be treated like patching a nuclear power plant.

My bad, I was thinking these would be people who might buy appliances or security hardware that connects to their home network, through something like wifi.

#15

So, I guess nobody is doing professional installs of HA? Because I only see advice from people like myself, using it privately only.
A friend of mine (who is a financial advisor) said to me this: their are two kinds of possible clients:

  1. People who don’t have the REAL money, and they buy something like Philips Hue and work from that. And as soon as they want more they expect it to be almost free and with an app that they can use to setup themselves (like Philips hue). These are NOT your clients
  2. People with REAL money who are willing to spend top notch to make it all work. These are you clients. You have to question yourself if you want to work for those kind of people. They tend to complain a lot and sometimes be unreasonable about what you have to do for them. Also you should do some research as to how many people like that there are in and around Cancun. And lastly, these kinds of people usually want to hire companies with multiple employees and a SLA to have the warranty that they can call 24/7 with issues.

His final advice to me specifically was to go and build custom garden houses for pets because even nearly broke people tend to spend ridiculous amounts of money on stuff for their pets.

4 Likes
#16

Again I do appreciate all the comments and points of view.

My target audience (cilents) are those in the middle to low upper class. For example, locally I can the shelly 1 for about 400 pesos (about $20), but I can set up a poor mans version (with more options) for about 150 pesos (about $7.50)

Someone mentioned those that don’t have real money that buy something like philips hue. That would be my target audience, which would be the middle to lower class gap.

As far as those with REAL money, I get the point. I have enough know how that I could go with fancy hide away monitors, auto curtains, etc., but their expectations can be to high. I contemplated going after that market for a few days, figured I could criticize their current setup, but figured I would have better luck with volume.

For what it is worth, I’ve contacted 4 of the major CCTV companies in Cancun, and so far all 4 of the people I’ve spoke to are idiots. They don’t understand that the fact that they (or the client) logs in to an app, it means there is an outside server involved. For the most part, for residential service we don’t get a public IP in Cancun so accessing local network from the outside is not straight forward.

Also not sure if I mentioned it, but I want to gear much onto the security angle of things. Here in Mexico if you only have surveillance cams it doesn’t do much if you don’t catch them on the act. Police wouldn’t bother tracking people based on a video picture you might have. If you mix cameras, home assistant, a door sensor and telegram, you can end up with a pretty good security system. When “away” (alarm on), if door is opened, have camera take pictures and send it via telegram to not only you, but to your neighbors which can be at your property withing seconds.

1 Like
#17

All the comments so far has been about starting up and potential customers.

What about the exit?
What are you going to do when everything has calmed down and you can get employed again.
Let’s say your business is going OK, you make enough to keep the family happy but you also have to work hard for it.
You will need to sell the company to someone who can continue your work, because sending an email to the customers “Hey everyone, I quit!” is probably not a good idea.
They expect the service they got to continue.

But if you are in this for the long run then it’s all good, except I also believe in the “don’t upgrade way”, which means you will have different versions at each clients house.
The first couple of months will be ok, but in two years, going back and fixing that old version will be like finding a dinosaur.

I can’t give you much advice, just more headache.
Sorry for that, but it’s also important to think about as much as possible before starting it all.

#18

Most process control systems in a nuclear power plant were designed in the 1970s. Patching is almost unheard-of. And with information security requirements nowadays, those old stand-alone systems which don’t have network capability are just what they want.

Anyway, I can’t imagine running HA as anything more than a hobby right now.

I certainly wouldn’t pay anyone my hourly rate to implement the system I have in my house, even without the constant upgrades. Just a simple restore from backup if my SD card dies will easily kill a couple of man-days, disconnecting and digging the hardware out of the utility closet and setting it up on the bench to re-install and restore, then putting everything back, re-connecting and probably fixing things which broke in the process.

As for upgrades, my guess is that the clients would see some new IoT thing for sale want it. At that point you’ll probably have to upgrade their system to support it, or maybe HA doesn’t support it yet and you’ll have to tell them they can’t have what they want.

I see a support nightmare in this business model.

2 Likes
#19

I’ve been self employed for the last 13 years or so. The idea is to try and pull this smart home/security system business going now, and if they are both profitable continue to do both businesses.

The sending the email to my clients “I quit” I don’t think would be much of a problem. See I don’t think I’ve gotten a post from someone that is actually doing HA installs as a business. Some of you say it would be a bad idea, but is there an alternative that wouldn’t be a bad idea?

The closest thing I can think of would be security alarm companies. They charge for original set up, then a monthly fee to provide support and monitoring. The key is that there is a monthly charge, hence the continue to offer the support.

Not sure, but for example Geek Squad from Best Buy, they would charge a fee to install your new TV, configure it with your new audio system, etc. But I am pretty sure once they leave the house, the service is over. If TV breaks, you contact the TV manufacturer for warranty. If they do a firmware upgrade on their TV and netflix credentials were reset, it wouldn’t be up to geek squad to send a tech back to the house to configure the netflix credentials again for them.

Again, I do appreciate all the different comments and points of view. I am still working on the actual product/service I want to provide.

#20

I’ve run my own construction business ~10yrs now, and it is largely dependent on referals/craftsmanship. In my area at least, there isn’t a sizeable market for this sort of thing unless you have an affordable long term support plan. I couldn’t picture talking any of my clients in to learning HA if they didn’t already want to. If I offer a no-support turn key install, then it’s probably only going to sell if it has lots of remote access features. In my wheelhouse, that would mean cloud or vpn. I’m not getting in to providing or supporting cloud services myself, and the liability with setting up a vpn to their home is too great (lots of risk not even related to HA in that regard).

Bottom line, I can imagine meeting a super tech savvy client who just doesn’t have time to install and pair all their sensors, switches, etc, and getting a job just doing the physical labor part. I’m pretty sure dude wouldn’t want me setting up his vpn certs, or even configuring his HA instance. Non-techs are going to want easy to use remote access stuff… no way I’m supporting that myself.

All this said, it seems that offering ‘HA training’ classes of some sort might have a market… IDK though that’s not my lane.