IFTTT and SmartThings (UK accounts)

Brick and mortar sales in the USA (Best Buy and Sears) are just starting now too. That’s got to reach a somewhat different consumer segment than online, especially depending on the training and quality of the sales reps.

I should play “mystery shopper” and report back.

Follow and contribute to this somewhat related Topic if you wish…

This is exactly the problem. SmartThings has come to UK market with a clear set of statements about what the product will do. At the moment it can’t do it. It says it on the website, on the POS at the store, in the email campaigns. As of right now, it’s false advertising. It’s not a home automation hub. My RF Yale alarm system from 1999 is more intelligent than my hub at the moment.

I’m not going to say which tech publication I work for, but I’ll certainly be talking to people in the industry about it. And I’ll be getting in touch with Samsung for an official response.


The timeline advisements from Support are not being made in a vacuum and I believe @JDRoberts ‘visible bug/hidden architecture’ meme was fairly appropriate:

There are people much smarter (and mostly better looking) than me working on these externalities. I’m trying to track down more updated information because the changes are taking longer than we originally (re-)estimated.

Kindly keep in mind the support team is public facing and bearing the brunt of your frustration. The UK support team are some of the biggest SmartThings fanboys we have. I know it’s super easy to get angry with them when ‘days’ turn into ‘weeks’… But send them stupid gifs because they are relaying the feedback with a heavy hand (and a British accent).

1 Like

@Aaron I thought the support staff would at least have a humongous Samsung TV on the wall with live updates on issues like this? :wink:

I’ve stayed quiet so far but it is rather puzzling that this is still not fixed and ST was released with this not working. What confuses me even more is who did you think would be your major target audience in the UK? total IoT/HA newbies, existing HA (e.g. Vera) users or people in between?

Personally, I’ve not dabbled with HA much before ST but have been spoiled with stuff like harmony, nest, netatmo, wemos, and lifx, and I’m guessing a lot of UK users have invested good money in one or many of these and other things.

ST’s advertising with promises of tieing all these together with a similar (Apple like) user experience certainly influenced my choice in ST…and I am rather disappointed it hasn’t worked out like that so far.

However, being an IT techie, geek/tinkerer and generally too busy, this hasn’t annoyed me too much yet but I can clearly see why it would anger many others!

Also, Apple’s homekit is almost upon us, Google has Weave (I’m sure OnHub is not just a router!) and existing HA wanting in on IoT (e.g. openhab v2), you guys should really be making better use of your time atm pushing out positive vibes for ST and not the opposite…

Not only do we not have a giant TV on the wall with updates, we don’t have a TV on the wall at all. Feel free to send one - minimum of 60" - ATTN Aaron and I can arrange for a broadcast platform updates (and/or EPL, NFL, RWC…)

@Aaron I wish I had a 60" tv myself!! And anyway, I’ve spent way too much money already on additional ST stuff e.g. z-wave relays, to compensate for all my existing stuff not quite working with ST yet! :wink:

Anyway, justify the cost by saying you need to further test ST integration with Samsung’s brand new 4K “nano crystal colour” tv sets! :wink:

Well, with respect Aaron - oAuth works in the US - Samsung changed the architecture. Samsung broke the architecture. Samsung didn’t apply any QA to the solution. Samsung it seems now can’t fix it in a timely fashion. Samsung could temporarily revert to the US architecture but haven’t. Samsung aren’t communicating well on this problem. Samsung I now believe launched knowing of the issues. Samsung stated that it could do things that it couldn’t. Samsung still continue to pump out new marketing saying it does these things when they know it doesn’t. . Samsung are treating their customers with absolute contempt. Not the sort of situation that a leading global company like Samsung should ever be seen in.

Seems I’m knocking Samsung at every opportunity - let me see if I can think of something positive…

… still thinking …

One good thing about @JDRoberts meme is that in the current darkness hopefully there’s light at the end of the tunnel.

… more thinking …

Ah yes…It’s potentially a nice forum with Samsung people contributing and Aaron posts amusing GIF’s & UK support are ‘supportive’ but only in the sense of palliative care without drugs. It’s very apparent that they are struggling with communication within Samsung too.

The one thing that keeps me here is that when this does get fixed, and the V2 development environment gets enhanced, reducing cloud dependency, then it’s potentially a very good platform opening up a volume home automation base.

Don’t buy this for Xmas fun though… it’s not a stocking filler - more landfill at the moment.



Further, it doesn’t work at all as far as I can tell

I don’t have any technical details (I’m just another customer here, but I used to work at IBM and have personally worked on many multi region projects, including the Olympics). However, my guess is that there was never an option to run the US architecture for the EU.

There are two reasons. Privacy and reporting laws are very different, making account structures distinct. And, in part because of reason one, the cloud support services run by AWS (Amazon Web Services) divide the world into both regions (geographic) and “availability zones” ( assigned to specific equipment) , with web traffic coming from different areas kept separate and handled somewhat differently.

AWS provides support services to many of the world’s biggest companies, nothing unusual in that. They are experts in handling massive traffic flows.

Oauth is an optional method

OAuth is not part of the AWS service itself. It is an authentication method for individual accounts that companies running their websites on the AWS service might choose to use. In that regard, no different than a currency converter they might choose to use, or a horoscope generator, or any other action they apply to their individual customers’ accounts.

Based on what information has been given on developer calls and other customer communications, it appears that SmartThings was unaware of how oauth would work once multiple availability zones were involved. Had they simply stayed with US only accounts, they would still have run into the issue eventually when they had enough customer accounts that some were assigned to a second availability zone within the US.

My own guess is that the fact that there are multiple regions involved as well is what prevents simply moving the relatively small number of current UK accounts to the current single US availability zone being used. The account structures, reporting and privacy requirements, and certifications are different.

Again, I don’t have any inside information, and I might be completely off the mark here. It just fits what I know about working with multinational web projects.

The architecture issue would affect the things that currently do work as well

If I am correct, it explains why the problem is so much bigger than it appeared at first. The entire method of account authentication, which is used every time A device is turned on in the home, may have to be rethought, even for the things that are currently working. It also explains why there would be no easy fix, and why UK accounts couldn’t simply be assigned to the existing US architecture while some technical issue was worked out.

My personal lens may be flawed

I do understand why this might take a long time to fix. What I don’t understand is why the major symptom, that a UK account can’t use IFTTT, wasn’t known before launch. It’s not subtle.

However, as I’ve mentioned, my initial use of IFTTT of SmartThings in the United States was for voice control. So I used that integration all day, every day. It would be the very first thing I would look at in considering any new system, and the first use case I would set up to test. So it may be that I am blinded by my own priority list into thinking this could not have been missed.

One of the very best things about SmartThings when it is functioning as intended is that it is an open platform, and very flexible from the customer’s point of view. Different people can and will use it very differently. As someone who is typically “an edge case” ( I use a wheelchair and have limited hand control), I appreciate this aspect very much. I do think customers often surprise SmartThings with what they can get the system to do. That’s quite a good thing overall. However, not all surprises are pleasant, obviously.


But surely when Samsung approached Smarthings to buy them Smarthings must have sold the product somehow outlining that they could deliver a multi user (family) country system - otherwise why would Samsung buy a company that could only deliver a product that could work in the USA.

Plus as Samsung is a multi-country company that have system that allow uses to login to there regions servers, why didn’t Smartthings just implement a system that used a users Samsung account as the login processes which was the used to authenticate with OAuth? I can see at least two ways of resolving this as a programmer/system administrator - Smartthings have messed up. They tried to do too much all at once, new V2 hardware, backend improvements and a UK rollout. This should have been one in stages with the backend first (making sure that it was stable and scaleable) and patch V1 hardware to work with that. Then roles out V2 hardware and a short time after that roll out in other countries in stages, giving them time to make sure a muliticountry authentication could work.

This is really bad management, you just need to watch the video on YouTube where they admit they did too much too quickly without the infrastructure in the background to support all the advancement.

If they are reported to the ASA in the UK, and they are found to have misold this product in - they will be fined, damage Smartthings as a company and put people off buying this product.

They have to stop this we don’t know when, but soon approach. And actually give the community a proper answer, if this had happened in the USA someone would have taken the to court.

Samsung/Smartthings support/management need to be honest and tell us when they will resolve what some are saying as a minor issue, but in the UK unlike the USA we don’t have a full functioning system that doesn’t fulfill what they say it should.

1 Like

Sorry for the rant, just feed up of lack of answers for piece of plastic that cost money and can tell me when I open a window or door and maybe turn a table lamp on, that’s not home automation - just a wast of money.

I wish I had waited now for HomeKit or Googles offering which would have worked out off the box first time.

“Ignorance of the laws of cloud computing is no excuse.” – to paraphrase something.

  • Was there or was there not an Alpha and Beta test period in the UK?
  • Did these tests reveal the OAuth flaw?
  • If not, why is the test bed incomplete?
  • If so, then why was the product released (or why was it released without a substantially visible warning of the loss of key features)?

Stability issues of the Platform are ongoing and, well, it is consistently unstable (i.e., some fixed things stay fixed, some get broken again, and some new things get broken, and some things never get fixed … ie., consistently).

But if the UK release is clear evidence of poor QA / testing practices … time for some new management at ST, because the poor QA practice is obviously affecting USA Customers as well, as exemplified by the rash of new issues appearing after last week’s Platform upgrade release.

I am pretty sure @JDRoberts has hit the nail on the head with his description of where the problem lies. From everything I have seen and tried out myself this theory fits. I have no idea why the IFTTT issue wasn’t discovered before release and can only guess that all the effort went into testing with the ST hardware devices and UK supported devices (which is a much smaller list than the US partly because a lot of devices are not available over here). My guess is that IFTTT and oAuth was assumed to work or was believed to be an easy fix, until they scratched the surface and discovered a major amount of work involved would be involved.

Now the point I have I is this, I understand IFTTT is very important but lets put that aside for the moment. Stability wise does the ST system crash? Do the sensors work 99% of the time? do the other sensors work to roughly the same level? When people say stability of the system is poor could the issues be RF interference, Low signal strength, high bandwidth usage on their internet connection, adding devices that are not supported etc. How do you define stability? A software bug that prevents you from triggering a sequence but doesn’t stop every other device from working? I personally have only had a few minor issues and those have mainly been user error and one Hue bridge error.

As I said oAuth is super important and I really hope that we get a real update on expected resolution soon. I know they will fix it as there is way to much money invested in launching outside the USA. Its just a matter of time.

Also just for transparency sake, I work for Samsung UK but I am not part of the ST team, although they are in my building. I participate in this community as a user of SmartThings and my opinions and statements etc are my own and NOT that of Samsung or any part thereof. That said I have made sure that the guys here are aware of the huge amount of frustration (including my own) and pointed them to this very thread in an effort to drive home how important it is to get oAuth working ASAP.


SmartTiles is also affected.

As SmartTiles is the #1 most installed Community built SmartApp, the odds are very high that problems with it would have been discovered if there had been a live Customer-Beta of sufficient duration.

SmartThings’s Betas are run under NDAs (non-disclosure agreements). Why? Perhaps to avoid spoilers for the actual launch; but also to avoid the spread of unfair instability news. The latter point is reasonable.

Unfortunately, the consequence is that I have no idea if there was any Customer driven Beta testing in the UK, nor how long or how in-depth it was.

I am aware that US Beta testing was much shorter than I hoped, though I also do not know the exact details (duration and number of Beta customers).

@tgauchat Smarttiles is a third party app and as such wouldn’t necessarily come under the scope of testing unless it was specifically tasked. That said SmartTiles is my No.1 want as soon as oAuth gets fixed.
Beta’s and NDA’s go hand in hand as manufacturers don’t want information getting out before release that may spoil a competitive edge or give a competitor advance warning of a killer feature. Its got nothing to do with spreading unfair instability news. Beta’s are expected to have problems that’s the nature of Beta’s and the reason they exist.
I also have no idea if there was a customer beta in the UK, I would hope there was… but I if so what indeed was the scope and duration?? I guess we will never know.

1 Like

That’s exactly the reason companies like SmartThings hold (or should hold…) “in-the-wild” Customer Beta test periods: To discover the items that are missed from the internal myopic regression tests and also to scale up testing.

Whether or not it was reasonable for SmartThings to miss IFTTT and SmartTiles (#1 SmartApp – not boasting, but explaining that it should have been on the regression testing radar because of the high install count!) during internal testing; the odds of discovery during a well executed Customer Beta are high – i.e., if the Beta was appropriately long and large. Instead, the UK release was rushed to market in order to meet the September 3rd IFA Berlin Samsung keynote presentation.

As a former network engineer, I’m pretty good at isolating and identifying issues of local interference and all that.

I have now spent a year with SmartThings in the US and I can say that most of the time reliability is about 85%. Good enough for me for most convenience use cases. Personally I wouldn’t put anything on it that would be expected to manage a life-and-death emergency. (I have a completely separate security system for that.)

Cloudy with a chance of Hurricanes

The problem is that quite regularly, I would say about every six weeks, there is an unexpected change so that things that worked on Monday stop working on Tuesday and take several weeks to fix. It is never a local problem. It has to do with the cloud account or the back end platform. This was the reason that existing customers were so excited at the announcement that the V2 hub would have “local processing.” We wanted to be isolated from that instability.

Unfortunately, the amount of local processing actually delivered with the V2 hub was minimal at best.

I have a good morning routine which does exactly one thing. At 7 AM, it changes the mode to “home.” This morning it failed.

Last week in the US we had a scheduled maintenance on Tuesday. Since that time, a number of US customers have been reporting poltergeist problems. Lights randomly going on and off, mode failing to change, devices of all protocols not being recognized. Schedules failing. There are at least seven threads in the forum about this, some of them with hundreds of messages.

Support has acknowledged some “hiccups” and some hot fixes have been applied, but the problems continue.

A few days ago my housemate was walking out of the kitchen carrying a beer and a bowl of popcorn. The television was turned on to an important sports event. The lights went off, and the television began randomly changing to other channels.

So for the last week, reliability has been extremely low. Nothing to do with our local installation.

For me, this has been the pattern over the last year. A typical week has about 85% reliability. A bad week has many random events, so much so that people start turning things off of the automations. The platform issues which caused the unreliability are eventually found and corrected. And then we have a period of relative reliability for another month or two.

You will find much discussion in the forums of reliability and stability, so just look around.

In the US, there are some competitors in the same price range that have much better stability and are aimed at small businesses. However to get that stability they threw out pretty much any feature that might cause instability. So they do not have geopresence, they do not have IFTTT, they have minimal or no voice control, they do not allow for custom programming, they typically have an extremely small selection of compatible devices and they may even require professional installation. They don’t work for me, because I need voice control. But they do run a simple timed schedule day in and day out.

Every system has pros and cons, and there is no one clear winner in the DIY sector, as different households have different priorities.

My own expectation for the US, as I have often mentioned, is that by the summer of 2016 there will be at least four and perhaps more essentially plug-and-play reliable home automation systems with some voice control that will meet my particular needs for features and budget. I expect HomeKit/Insteon to be one of those. I am expecting SmartThings to be another potential candidate. However, they are not there yet. But then neither is anybody else.

I’m not sure whether the UK will be on the same time schedule, but the market push from Apple’s HomeKit will make a difference. So I do expect there to eventually be similar choices in the UK market.

For right now, however, I believe most well-informed people choose SmartThings for its versatility, not its reliability.

1 Like

@tgauchat I completely agree that if it wasn’t tested it should have been. The community made apps and drivers are a huge part of the ST ecosystem and should in my view be included in any testing .
@JDRoberts interesting 85% is lower than I would have expected and isn’t my experience but I’ve only got a small install compared to yours.

1 Like

I’m also an edge case because I spend most my time in one of two rooms, both automated with SmartThings, and I use it all the time.

Additionally, when it does fail, it’s a big deal for me. If the door doesn’t unlock from the inside automatically, it takes me about 10 minutes to get the door open. If the light doesn’t come on when it should, many people will just hit the switch and be mildly annoyed but may not even remember an hour later that it didn’t come on. I do have a second option for everything but again it could take me several minutes to get that light on.

So I do notice the relatively small reliability issues more than many people who live with the system. I know I’m not a typical customer. :sunglasses:

The CTO of SmartThings posted to the forum last spring during a period of particular instability and said that he wanted SmartThings to be the last thing the customer suspected when the lights didn’t work. I don’t think even customers who consider it “quite reliable” are there yet.

I’d just like to see it get to levels of appliance reliability. Similar to the dishwasher or washing machine. I don’t expect it to be perfect, there’s a lot of complexity. However, I would like to feel surprised by a failure.

1 Like

This is why I fear the future (not the future of SmartThings specifically) – I fear the future of every “thing”.

I have a high quality dishwasher (Bosch). It has various programmable options, but the most complex feature I ever use is to set the timer delay. It has never failed; neither software, nor hardware, nor mechanical issues.

The arrival of IoT hype means that the next model of this dishwasher (or perhaps the one after that) will be “smart” and “connected”. Sure … that’s got some benefits – avoid running during peak/dirty/high-price electricity periods, for example; alert me when cycle is completed, remind me to clean the filter, automatically ship detergent based on usage, upgrade the cycle to be more efficient … well – let’s think about that last one – “software/firmware updates”:

The concern: The vendor will now be able to push “updates” to the dishwasher whenever they wish. This significantly increases the risk that what worked yesterday will not work today. Literally, it goes from under 1% risk of failure (hmmm… I’ve had the dishwasher for over 2 years, no failures; great MTBF!), to, well… software updates bring a lot of risk. They are not always properly tested. Corruption occurs, etc., etc. … I fear a failure rate of as high as 5% – or worse if we take SmartThings as the example.

OK … 5% is paranoid, and 1% may even be high, yet you can’t disagree that there is a jump in magnitude of the risk level due to unrequested, undesired, and unnecessary firmware updates to my “it-ain’t-broke-don’t-fix-it” dishwasher.

Now repeat the above two paragraphs for my Tesla (I wish!) … slightly smart Toyota Cheaprolla, and my fridge, washer and dryer, television, etc., etc…

In other words: “Appliance level reliability” ain’t gonna be like the old days.