While most of you have all valid points, I think the opportunity always lies in not the obvious places. Yes, it is a pain to provide support, yes it is a challenge at times etc etc. But applying the same business model as what is used typically for installations and consulting services may not work all the time. The first opportunity is in making margin in the products you sell and in the labor involved (installation, configuration etc) - Ex: Switches, Motion Sensors, Door Locks, etc… There is typically somewhere between 20 points to 60 points of margin depending on the product . So one needs to choose the product offering and devise a solution carefully to maximize profitability. Partnering with certain reliable start ups has also proven to be a good strategy, in providing new solutions while getting some level of exclusivity in solution offering. Another opportunity is if you are offering your own product to add significant value to the set up - For example we have couple of products - dotAlarm and dotBell that integrate the existing hardwired alarm system and door bell with SmartThings. The third opportunity is in offering service based models - Ex: Calibration Services, Battery Management services, etc or even provide solution for cellular based internet connectivity . Another opportunity would to offer software solutions that help solve challenges such as accessibility (Like: SmartTiles, SmartRules etc), with a reasonable monthly fee or annual subscription (ofcourse it needs to be worth the money). The service based models can also be made a annual subscription with a deductible type model for service calls, which encourages customers to try to resolve simple issues on their own.
With all this being said, the over all system reliability still depends on SmartThings…for sure. In my experience more than 40% of issues people face on this community can be solved by having a technically sound install - Ex: Placing a hub next to a WiFi router is a big mistake or having one switch that bridges two big group of switches causes what is called as an hour glass effect impacting performance on z-wave network drastically etc etc. To ensure proper installs, the design of the system using proper tools plays a big role. What about the remaining 60% of issues - I would say about half of them come from the way configurations are made - how triggers are set up, implementing some redundancies for key triggers like routines (which only kick in if original fails) and a lot of other configuration related items. Issues that come from SmartThings platform itself are usually less than 20% of the time. And they can definitely be a pain in the neck . So what’s needed is consumer education. Set expectations with the client that it is a free platform and self management with some level of service impact is inevitable. They can choose to reach out to ST directly for support in case of issues.
Another thing to consider is to not have 100% reliance on ST platform for everything. Design the set up such that individual systems can still work when ST is down. For example - Harmony, SONOS, Remote controlled shades , Stand alone Camera solution etc, still work when ST is down . The integration and solution development needs to be modular. In a perfect world ST as a platform would bring everything together, but should not be the bottle neck in terms of causing a total havoc when things go South.
So far we’ve got more than 30 clients , all on ST platform. In the initial 6 months the support we had to provide was crazy, as we were still developing and refining the business model. But since last 4-6 months, we’ve hardly received any support requests as we have a strong implementation strategy, which helps us design the system such that ST is not the only bottle neck - This is despite ST having issues so many times. We have also revised the products we use over time - In fact some of the products we use are not even officially on “works with ST” list, but we find them more reliable than others. GE Switches are the worst switches we have seen in terms of reliability. So far we must have installed over 600 switches totally. The configuration mechanism we implement on ST also has improved since we started. We also have our own documentation for troubleshooting which we provide access to our customers (in case they want to resolve things on their own), With all this and a lot of other things ongoing, I think we are confident that providing Smart Home Services can be a strong business. The business model however has to be adapted to fit the way the smart home industry is currently shaping - its very dynamic and a lot different from some previous standard solution providers.
Over all I dont think we are at a level where we can expand drastically, but we sure can do that eventually. May be some day ST will have a model to support businesses like ours (call is ST PRO 
), but may be that will never happen. Moreover, we dont think we will be limiting ourselves to just ST. Just set the right expectations with your client, educate them as much as possible and get PAID! … while thinking of ways to get recurring revenue.
As a note, more than 60% of my clients have given us repeat business by adding new solutions or expanding their existing smart home etc. Its now getting a little difficult for me to handle the workload , especially with my full time job … so i am working on finding some help.