My transition experience

All,

As many of you know, I have been a SmartThings user since 2015, having created some official SmartThings apps that are still in the Marketplace today, along with some other popular community apps like Alexa Helper, Ask Alexa, etc. In addition, I have created or taken ownership of many of the most requested DTHs like the GE Motion Switches/Dimmers. I have always admired the community (people like @JDRoberts) and many of the SmartThings staff for their dedication to the product. As such, I have always attempted to assist when folks are having issues. I recently went through the whole SmartThings integration and I wanted to share my experience here as many of you have reached out to me privately to get advice. To be clear, I am not a SmartThings employee, but like many of you, I have a LARGE install of devices and vested interest in ensuring my environment is operational.

The good news…my transition experience was overall positive. However, I ONLY believe this to be the case because I did NOT rush into the upgrade; in fact, I ignored many of the deadlines so that I could continue documenting my 5 years of automations and devices so that the transition would be smooth. I am hoping my experience helps some of you.

My first word/advice that you will have to accept…YOUR HOME FUNCTIONALITY WILL CHANGE DUE TO THE TRANSITION. There is no getting around this; be it good or bad, SmartThings has taken steps to continue to offer a FREE service, but at the expense of some of the things that made SmartThings great (like the open-source nature of the platform). For those of us that remember the ‘dark times’ in 2015-2018, things worked, but only occasionally. There were times I had failsafe upon failsafe (with tons of alerts) that would let me know when lights simply didn’t turn off when they should have, or automations that misfired at all hours of the day and night. 2019-2020 has been the MOST STABLE (for me) in the history of SmartThings. This corresponds with Samsung starting to lock down the environment bit-by-bit. Now, I am not a Samsung apologist; see the end of this document for my thoughts on the way this transition was handled. Spoiler alert…it was NOT good.

Anyway, on to how my transition went and advice:

SmartApps/DTHs

While it may surprise anyone who knows me via this forum, but I really didn’t want to have to ‘tinker’ with the SmartThings to get it to work as I wanted…However, in the early days, it was a requirement. I did end up writing some very powerful apps (Ask Alexa being the most popular) that others found useful, but as the system has grown the need to ‘hack’ it constantly was reduced.

As such, as the first part of my transition, I tried to dump most of my custom apps. I was rather successful, with only WebCore and Echo Speaks remaining today. On the device side of things, I only kept the ES devices and my own GE motion switch DTH as I used the programming API within the DTH to do some very neat things with my home automation without relying on the hub or even SmartThings directly.

So, my first advice…. SIMPLIFY YOUR ENVIRONMENT BEFORE YOU TRANSITION. While this SmartThings App transition does nothing to the smartapps right now, the deprecation of the IDE and Groovy WILL affect all of us next year. As such, you should get used to having just the SmartThings provided apps to get you through. I personally have no interest or desire to learn the new SmartThings API, so I am going to let my apps ‘die on the vine’. The ONLY exceptions are the GE DTHs. I have spoken to folks at SmartThings and I feel these are probably the most important devices they should natively handle. Whether that happens or not is to be seen. We will see….

Alexa Transition

While there was a published deadline to transition from the Alexa app on 9/8/20, I simply ignored it. Probably not the advice I would give others :), but I was busy and heard horror stories about those that did the transition. So, I spent up my time, up to the deadline, documenting every routine and operation that used the old SmartThings/Alexa skill. As such, I learned a LOT about what NOT to do in the transition. After reading about other users’ transition experience, I am pretty confident in saying that I may have just got lucky and there are probably some fundamental differences between the Shard that people are on in the SmartThings cloud; some people seemed to follow the transition and are still having issue. However, I also believe that if you attempt to cut corners during the transition (for example, not completely eliminating the old devices and skill) you will have problems. Here is the process I used…this mostly came from @nathancu advice, but I added some additional steps to ensure a clean transition.

First, I documented in a spreadsheet every Alexa routine, what it did and its dependency on SmartThings. I then deleted ALL my Alexa routines that had ANY relation to SmartThings.

Second, I deleted ALL the SmartThings devices from Alexa via the website (alexa.amazon.com). While there is advice to wait a few minutes, I waited half a day just to be sure everything behind the scenes was stable. Not sure if this is required, but it is not bad advice!

I then went into SmartThings and removed the Alexa app. Then, I went back into alexa.amazon.com and removed the SmartThing skill. Again, I waited a good amount of time before proceeding.

For devices that I knew could (and would) exist in both Amazon and SmartThings (like my August Locks or my Ring Cameras), I made sure to name the SmartThings devices UNIQUELY so when they transitioned over they would show up with the proper context (i.e. My screen door in the Alexa August Skill is simple called “Screen Door Lock”, in SmartThings it is called “Screen Door Lock (ST)”).

I then went to the new SmartThings app, enabled the new SmartThings Voice Assistant (which installed the skill) and my devices immediately showed up in Alexa. No need to discover, no need to tinker, and NO duplicates! I then recreated my Alexa routines. Everything responded as quick as before, and all was good….except triggers that came from SmartThings to Alexa…those WERE delayed or broken, but only up to 9/23. After that, items that I had set up to trigger via a virtual motion sensor or open/close sensor started working. Everything is working great as of today, no lag, and any new devices or scenes (including virtual devices) I create in SmartThings show up immediately without the need to run discovery!

So, advice for the Alexa Transition: SLOW DOWN AND REMOVE EVERY SMARTTHING DEVICE FROM ALEXA. USE THE NEW SMARTTHINGS INTEGRATION IN SMARTTHINGS TO GET THE SKILL INSTALLED.

Echo Speaks

I know everyone who still has this working is going to cringe (because I am admitting it works), but if you didn’t uninstall the app (at least in my shard) on 9/8 the app continued to work after about 2 hours of downtime. However, we all know this will not last. As such, I have begun moving my stuff to Voice Monkey or other methods to accomplish similar behavior as before (since triggering voice routines works well within Alexa). The one thing I missed (and had to tweak ES’s code to get working again) was the automatic setting of the Alexa Guard. That will be hard to get to work without ES when it is finally turned off, but as mentioned before, YOUR HOME FUNCTIONALITY WILL CHANGE DUE TO THE TRANSITION. This one thing, in the scheme of things, is minor but I hope the Lite version (if it can be released) will address. If not, I like the suggestion of one speaker telling the other to arm the Guard. The other things that is annoying is determining where Echo Speaks should talk (the last echo device to speak code no longer works) but there are ways around that as well.

Routines

I was never a big fan of SmartThings scenes even when they were introduced in the old app. However, it became apparent that SmartThings was going to orphan routines in-lieu of scenes. In addition, scenes can be activated directly from Alexa (except anything that has a door or lock in it). So, it was time to move over to scenes. Please note that when you do the transition to the new SmartThings app, a routine is broken into automations and scenes. However, there are some things that simply do not translate well from routines, so my advice for Routines is: BEFORE YOUR TRANSITION, TRANSLATE YOUR ROUTINES MANUALLY TO AUTOMATIONS AND SCENES OR WEBCORE PISTONS. You will find out you have to make a few compromises in the old functionality to the new, but doing it manually will help you learn the limitations (and some neat things) that can be done using scenes. To be clear, I used a combination of scenes, automations and WebCore to accomplish the same things as what my routines did. While this was all a PITA to do, the end result was that the SmartThings transition after pressing the button in the old app took about 2 minutes and worked flawlessly. And, since I understood how scenes and everything tied together, once my STHM was enabled on the new app, I was able to kill the WebCore pistons that were previously changing the SHM. (Yes….it is true…you can no longer use WebCore to change your security state).

Summary

As mentioned above, all of the time spent doing the transition myself was time consuming. I do appreciate that I was able to ‘simplify’ my home automation somewhat just by going through every smartapp, DTH, device, routine, etc. and really find out how it had organically grown over the past 5 years. I found items that were duplicating functionality, and stuff that simply was no longer needed. I dropped my custom DTHs down by 75%, my smartapps by 50%, and dropped or optimized my WebCore pistons by a small handful.

Again, I am not going to be a SmartThings defender here; my actual transition was rather painless because I spent about a month on it thinking and testing. However, I never even looked at another platform as every one of them comes with a larger and very manual migration process along with the need for additional compromises.

That being said, Samsung FAILED in a few ways:

  • Communication: There could have been a LOT more posts by internal SmartThings folk on here, or someone just assigned to the forum specially to help people with the transition. I did fine but only because of working within this SmartThings for 5 years straight. I think someone on here outlined it perfectly….if there had just been weekly communications from SmartThings, even if nothing had changed, that would have been better than the deafening radio silence from SmartThings.

  • Transition tools: I get it… the ‘end state’ where Samsung wants SmartThings to be may break long standing processes, or things may just work differently to accomplish the same goal. That being said, give the users more tools to migrate successfully. I mentioned many times in this posting about me documenting my system. That has ALWAYS been a weak point in SmartThings and would have been a useful tool for users wanted to understand the dependencies of all of the things in their environment. The new app continues this tradition with LESS history and dependency tracking of which devices are used in which apps/automation.

  • Tight/Not enforced deadlines: Many people pointed this out in various posts, it seems deadlines were being stated with little regard to the complexity of the task that needed to be done or the new functions not being ready for prime time. The Alexa transition was a perfect example…the skill was not ‘done’ on 9/8 (admitted by ST) but the deadline never moved. To their credit (or detriment) the 9/8 deadline came and went without the service actually being turned off. This is a no win situation….I could have respected the 9/8 deadline if my stuff simply stopped working, but to put a deadline out there with no ‘teeth’ to it is even worse. Again, I think cooler minds prevailed at SmartThing with the Alexa transition in that they KNEW it would not work 100%. But this goes back to the communication part…no one admitted they didn’t enforce the deadline or even say it was going to shift. That lost a lot of respect in my mind…Admit when you are wrong and learn from it. They just moved on. I understand some of the ‘back office’ of this…I have been in technology for 30 years and part of project teams that had to implement deadlines. Financial/career incentives and other political issues aside, nothing can forgive the perceived lack of communication around these milestones. Put reasonable deadlines in place and help people get there with adequate tools and communication available as needed.

I hope at least one person gets useful information from this post. The next big hurdle will be the API migration and the deprecation of the Groovy environment. I literally need only two things from the final end state of SmartThings: A COMPLETELY robust (dare I say, WebCore) rules engine that can allow the end users to really customize their environment without writing programs. Second, is to support the GE Motion Switches with a programming interface to set modes within the switch. Without being able to program them, I simply have a wall mounted motion detector and a switch in one. Being able to set their mode (and activate lights with motion without the involvement of the hub) is what SmartThings goal ends up being local processing, so these devices play right into that.

Good luck to anyone that has not transitioned yet. Hopefully this all helps.

19 Likes

So, I got a few “likes” on my post above, but I also got some direct messages asking me (per se): “So you are 100% happy with your set up”. Of course, the answer is no. They are about 80% there; I remember during the “old days” where there were few other options in the home automation space and I was at 50% there…so this isn’t as bad as it has been, and I see a path forward as the solutions are mostly programming and UI design.

For those at SmartThings/Samsung that are listening, here is my list:

Virtual devices: I have to believe the ONLY reason SmartThings has worked so well is the ability to have virtual switches in the environment. I have virtual switches that can even relay information from Alexa to SmartThings that my WebCore can then make different decisions with just one piston. However, my favorites are push (momentary) button switches since they reset themselves after being set. Unfortunately, these type of virtual (simulated) switches are not well supported in the new SmartThings App. I can NOT control them with the UI. Fortunately, I CAN with Alexa and that is fine, but it would be nice to be able to press a button in the SmartThings app to get these to work.

Interdependencies: Of all of the faults of the old system, it had some listing of interdependencies and actions that occurred within the system. Overall, it was usable. The new app I would classify as poor to terrible at this. While some of the action logging appears fixed as of today, there is still no way to know WHAT triggered an action…just that the action happened. In addition, if you need to figure out what switch is in which automations, good luck…you need to have a pen and paper to do that. I get it…logs take up space on hard drives and THAT seems to be the overarching theme of this whole transition….a free service can’t use a cloud service that charges by compute, storage, etc. That is why I think they should have simply had a free tier for basic users where there was little need for logs or even cloud compute, and subscriptions for users that needed extra devices, extra functionality, etc. We can talk all day about cloud economics (that is what is what I now do for a living), but the people on here are the exception and would not be able fund all of the cloud backend. As a consumer, I still like that this service is “free”; however, give me an option so I don’t have to pull my hair out to troubleshoot or figure out why an automation didn’t fire.

UI (user interface): What can I say…I don’t like the user interface much, but that is more personal preference. The UI, however, is one of the EASIEST to fix with the right UI developer and research. If they focus on usability, they will easily be able to easily improve upon the interface. Good example…Linked Services for things like IFTTT, Ring, etc…That is currently NOT where I would expect it to live.

Brand Tie ins: One of the fears that I had when Samsung took over is the “Sony” experience. Sony, in the day, was great for developing their own standards. They would then ‘tie in’ other products to their standards, only to be abandoned soon after. Samsung is slightly different in that they follow the industry instead of lead. However, they tend to be moving toward a more tie-in mentality. Example? The linked locations. When I turned this on there was a little note saying I couldn’t use this without a Galaxy device. Ok…no biggie since I use Life360 for my precense and I don’t own a Galaxy phone (I do own an Android phone, however). I am curious if the iPhone shows the same message? I get this ‘tie in’ for TVs and their own products to have SmartThings CONTROL them, but don’t exclude functionality in the SmartThings app just because of the mobile device I am using. Shame on you Sony Samsung.

While Samsung promised (for the most part) in 2015 they would not do that…

You really can’t hold a company to its promises 5 years into the future. That was their intention then, and they kept it (for the most part) for the past 5 years. Just disheartening that the industry has always rejected tie ins and yet Samsung hasn’t learned.

Bugs/Automations ‘strangeness’

Simply put, there are still some strange bugs within the new SmartThings app. I actually uncovered one yesterday that I haven’t seen in years. Basically, in an automation, if you have a wall plug-in dimmer that was previously set to 100%, and was turned off in that state, the next time it is turned on (even with a scene stating it should be 5%) it comes on at the previous level. Little items like that are easily worked around (i.e. Smart Lighting app instead of Automations), but it shows that the folks designing the new SmartThings application (Samsung) are not familiar with some of the aspects of home automation that the original programmers of SmartThings were. They will get there, but I think this goes to the whole deadline thing…many of these aspects were not vetted with the community as beta tests as they should have been. However, having been a programmer in the past, I also feel there is some ‘wonkiness’ in how the automations still work. For those that understand how SmartThings used to be, in a SmartApp you would ‘subscribe’ to something that you wanted to happen (i.e. A change in a light, illumination, etc). This would set off an action and that is how things worked. I have no idea how the new stuff works, but I had an automation (that I moved to Smart Lighting) that would fail 75% of the time under the same conditions (illumination change). I am convinced that the device “DTH” (if there is such a thing now) and the automations still aren’t ready for prime time. I hope they get there, and that they DON’T deprecate Smart Lighting (or WebCore) until there is an 100% reliable solution.

That is what I have seen. Again, these can ALL be fixed with time, effort, and FOCUS on HOW the users are using the system.

3 Likes

I believe @orangebucket has a virtual momentary device that works with the new V3 app. :sunglasses:

1 Like

Routines in the Classic App were an excellent way to have a “button” and not get hit for a device or clutter your device list.

Scenes come in as buttons on the Alexa App but are not accessible in webCoRE. Kinda sux.

But you can have an automation that activates the scene and use webcore to turn on a switch that triggers the automation. Clunky, but it works.

Good write-up.

iOS indeed shows the sames message for Linked Places :frowning:

1 Like

Excellent write up!!! Many, if not most, should read this in it’s entirety and take good notes.

Your story regarding migration is/was very close to my experience. No spreadsheet but I used screenshots…A lot of screenshots!!! The transition also gave me an opportunity to clean up a lot of things that I no longer needed. Physical Devices, Automations, Virtual Switches, Contacts, Motions, etc.

The New App Automation is just not dependable. Mobile Presence is not dependable and now SmartLighting, which has worked flawlessly in the past, is no longer dependable. I’ve been moving everything I can, which is most, to webCoRE.

I don’t understand how someone, bless your heart Andy, can make a rule engine that’s so darn reliable and a whole team at ST can’t match or beat it.

Your Apps @MichaelS were very very good and one of the reasons why I was able to do some of the awesome stuff I’ve accomplished. Thank you for all of your hard work.

Without question, @JDRoberts IS THE BOMB!!! Whenever I am searching for something, I put what I’m searching for then @JDRoberts behind it.

In the end, I’m sticking with ST because as you stated, 2019-early 2020 was heaven! I’ll gladly pay a sub for more advanced features that I’ve come to rely on and love.

3 Likes

Yeah but still is a device hit. Hopefully Scenes will become discoverable in webCoRE then that’s that. The work around for me, without adding a Virtual device, is to just trigger the scene with Alexa.

1 Like

You can run a scene from webCoRE with a web request if you particularly want to avoid extra devices. You just have to get an access token and look up the scene ID.

2 Likes

Good to know! Even better to know how. :rofl:

1 Like

A write up on this would be great for the community as Scenes are not readily seen by WebCore (as you know).

2 Likes

How do I get the access token and scene ID?

4 Likes

I just realized I need to keep my virtual switches to activate scenes because ActionTiles does not support the new scenes directly yet.

They have beta support, I believe.

@625alex

2 Likes

Screenshot_20200929-223826_Chrome

2 Likes

Wow. Guess I need to pay more attention. Thanks!

That kind of goes against my philosophy of trying to use stock stuff…so yeah…it ends up being a catch-22 :frowning:

1 Like

I feel your pain… :disappointed_relieved:

1 Like

ActionTiles has Beta support. It may or may not work for you because the ST API is not stable, yet.