SmartThings will kill legacy features to make a smarter home

To hear more about this, check out my podcast with Mark Benson of SmartThings. 

Samsung SmartThings will undergo several big changes in the next year or so that will affect developers, smart device makers, and customers as it tries to move everyone off its legacy software platform and onto its new app. The company is making the changes as a way to give it more control over the types of experiences it can offer and as a way to boost security. While all of the details aren’t yet clear, users and developers should be prepared for some legacy features and apps to disappear.

Mark Benson, head of engineering at SmartThings divided the changes into two categories: hardware and software. On the hardware side, SmartThings is planning to open up and offer its software to other companies, so we can expect to see other manufacturers make SmartThings hubs and compatible devices. This isn’t all that new, as SmartThings is available via USB stick on Nvidia Shield TV devices and has been announced for gateway devices supplied to ISPs by Calix.

I asked if we could expect older SmartThings devices to stop working as part of this transition, and Benson said, “Over time, there is always a point when legacy devices and products will get retired. We’ve had a long history of maintaining devices in the market for a long period of time. But as there are new technology products that come out, some older generations of products will be retired.” I suppose I can say goodbye to my seven-year-old SmartThings sensors and outlets that came with my first hub.

Benson did stress that when SmartThings retires a device it will let people know ahead of time and try to do it as smoothly as possible. Hopefully, this won’t look like Automatic’s rather abrupt shutdown, where customers were given roughly 30-days notice that their device would stop working. But the biggest changes are coming to the software side.

In the second half of this year, SmartThings is going to start a three-step process that will help clean up its back end. First, it plans to push the remaining users of its legacy Classic SmartThings app on to the modern SmartThings app that launched in 2018. This will mostly affect people who are still using their original hubs since most of the users of new hubs will have already switched as part of the upgrade.

Second, it will stop letting users build their own device handlers in Groovy, the programming language that the SmartThings platform uses. Instead, developers and users will have to use an API to access features, devices, and controls. And third, it will stop letting developers use its IDE (integrated development environment) to build custom device handlers. Instead, developers will use a new development environment that’s still unspecified. Benson notes this last shift won’t happen until some time in 2021 and SmartThings will provide more details as the time approaches.

[Get my smart home and IoT analysis in your inbox. Subscribe to my weekly newsletter.]

But those last two moves are going to frustrate many of SmartThings oldest and most technically savvy customers. When the platform launched, the idea that you could pick up any sensor or connected product and spend a few hours writing a device handler for it so it could work on the SmartThing platform, was a huge plus. Developers would write their own handlers and share them among the community. Benson says that the transition won’t abandon the existing world of wild and wacky device handlers because that’s a competitive differentiator for the platform.

“Those kinds of things are and have been a staple of SmartThings. And those things are not going away,” he said. “We have [device handlers] today on the platform and we are making certain changes to the way that those types of devices get integrated to improve their ability to run locally, like on the hub, which can reduce latency and increase the reliability of those types of devices. But we plan to continue maintaining that ability to support those types of one-off devices like that on the platform.”

That’s going to be welcome news for people, but the switch from being able to write their own code to accessing an API is still going to give some pause. Benson points out that this will enable developers to use more programming languages including those that they might feel more comfortable with, which is true. It also helps SmartThings place tighter controls on the platform because it can close off certain features in the API or refuse access to the API to developers who might be abusing the program.

We’ve seen these moves before, most recently when Google transitioned from its Works with Nest program that allowed all of Google’s Nest devices to communicate with other smart home devices in the Works with Nest program. Late last year it said it would change the way its devices worked, requiring most of the access and decision-making to go through a Google Assistant compatible hub. Some users were upset, almost all of the users were confused, and I would argue that the transition from Works with Nest to Google Assistant did not go smoothly.

Presumably, SmartThings will learn from Google’s mistake. It’s approaching this over a very long time frame and seems to understand what it users value about the program. Benson couldn’t give me a lot of details today (I imagine some of them are still being worked out) but he did pledge to communicate with users. Starting the conversation early will give them time to respond.

SmartThings is pushing these changes because the smart home is finally getting to a point where users are comfortable buying devices and automating some things. But to take it to the next level, it has to get much, much easier. Benson’s lofty talk of “experiences,” covers a lot of ground, but the basic gist is that when someone buys a connected product in the future and brings it home to a SmartThings ecosystem, they should expect it to automatically connect and then to get suggestions about how that device might fit into their current apps or experiences running on the platform. Apple announced some similar “experiences” this week at WWDC related to lighting.

So if a user has a security experience running, and they bring home a lightbulb for their porch, they won’t have to program the automation that turns the light on when motion is detected on the doorbell camera. That will happen automatically when it’s connected (or likely will happen after the user gets a prompt asking them if that’s what they’d like to do.) To get to this world, we’ll need more than a rejiggered back end on SmartThings. We need a standard, such as Connected Home over IP that SmartThings, Apple, Amazon, and Google are all trying to create.

We also need use cases that make sense to a wide variety of users so the developers can build these generic experiences and make them customizable to those who want to play around with their smart home. It’s a lofty goal, but it’s where the industry is heading. SmartThings is just doing what it needs to, so it can make sure it gets there.

To hear more about this, check out my podcast with Mark Benson of SmartThings.