CES is here, and that means that every man and his dog is on show, trying to flog smart home devices at a show that is now very conflicted on whether it specializes in consumer electronics or cars. There have been some notable launches, with some genuine advances in the technologies available, but we are still a long way away from the vision of the smart home we were so emphatically promised by the collective industry just a few years ago.
As expected, the show is a proxy-war between Amazon and Google, as the two technology titans continue to duke it out. Apple is, unsurprisingly, not at CES, but nor does it look in any real hurry to try and catch up with its rivals. HomeKit is getting better, but it is leagues behind in terms of the device ecosystem.
Google announced that Assistant was now installed on over a billion devices, but as smartphones are included there, this didn’t provide much clarity for the smart home. Meanwhile, Amazon said it had sold ‘tens of million’ of Echo devices in 2018, and that there are over 100mn Alexa-enabled devices in the wild. For the online retail monolith, the success of the third-party ecosystems will be music to its ears, as it can leverage its powerful sales channel any time it likes to boost its own Echo range.
For most smart home developers, an Alexa integration has been one of the primary objectives for launch, thanks to Amazon’s head start in the smart home race against chief rival Google. Amazon is still ahead on the number of third-party integrations, but Google’s tally is no slouch.
However, a little bit of napkin-maths suggests that there are around three third-party devices for each Echo device. That sort of ratio does not a smart home make, and even if the typical smart home was using a handful of other devices on top of an Echo and three, or the Google Home equivalent, this still paints a picture of the average smart home having no more than a dozen devices.
Of course, compared to where we were five years ago, a dozen devices can be considered a success. However, the Jetsons-like vision of the future is still years away from being realized to any great degree. A useful metric that Riot has used when evaluating smart home projections is the number of devices per room, and the corresponding minimum threshold at which a whole house could be considered smart – rather than say a single application within the home, such as its HVAC or security system.
For instance, a conventional three-bedroom house might only need one WiFi access point, which in time would likely serve as a smart home radio hub too. In addition, a couple of security cameras could provide the security elements, as well two door and eight window sensors. A connected thermostat would handle HVAC duties, and if each room in the house had one connected light, our tally is probably up to around 20 devices. Add in a couple of digital assistant hubs and smoke alarms, and feasibly, the threshold gets to about 25 – which by today’s standards, is a pretty comprehensive smart home. That vision is well within reach.
But cast your mind back to the promotional videos of yore, from the smart home startups decked out with fresh VC funds, and the smart home was meant to be teeming with devices. Damp sensors, air quality sensors, occupancy sensors, temperature sensors, sound monitoring, light quality monitoring; all feeding data back to some sort of intelligent assistant that would act on these data sources and optimize a home. Back then, the smart home was actually smart. Now, a smart home is essentially a voice-acted and somewhat automated home.
In the startup vision of yesteryear, the threshold for ‘smart’ looked more like an easy ten devices per room in a home, perhaps scaling as high as twenty once you added in a bit of luxury, such as connected blinds or curtains. Now, the industry appears to have settled for something more like three per room.
In our view, a home can’t be ‘smart’ until it is generating enough data for a controlling system to be able to use to make ‘smart’ decisions about how to run things. This includes external mechanisms too, such as knowing the routines of its occupants, local news and weather that might impact such routines, and likely a great deal of integration with the user accounts that provide much of the in-home services – such as TV, broadband, Netflix, Amazon Prime, security providers, utilities, etc.
To this end, a ‘smart’ home has to be filled with sensors, plugged into the digital lives of its occupants, and then able to make informed decisions using that heap of data, in order to actually be ‘smart.’ We don’t expect each home to house a unique general intelligence, acting as a personalized butler, as cloud-based platforms will be providing the bulk of the workload, but there does need to be some semblance of intelligence in there, even if it is purely a question of statistical approaches and deferring to humans when needed, rather than actual Artificial Intelligence.
As for new developments at CES, it is promising to see that a wider range of devices are now being connected, and also encouraging to see that the likes of Ikea are showing that you can slash the cost of some applications too. Its connected blinds are around three times cheaper than most rivals here, and get into the ballpark of what people would pay for an unconnected version of that same product.
This capability, to suddenly throw connected options into the lap of a consumer considering their next purchase, means that smart home device shipments will rise much more quickly. Light bulbs fast approaching that point, as are white goods, and in the next few years, the routine upgrade or replacement cycle that most consumers practice will see a home become much more connected than it ever was. And yet, we are still leagues away from the 100-device vision. Perhaps that sun has set?