Roomba hits damage control in smart home data sale snafu

It was meant to be a glorious investor call, fresh off the news of bumper profits and the successful $141m acquisition of its largest EU distributor, but the maker of the Roomba, iRobot, has had to hit the brakes – after a misunderstanding about its plans to sell indoor home location data sparked online outrage. The spectacle hints at the Smart Home as a Service (SHaaS) we should all prepare for.

It’s a theme mentioned in the enterprise IoT survey that Canonical carried out (see separate article), where the margins are crushed out of hardware and businesses have to turn to online platforms and services as a way to increase their revenues. A smart home supplier might sell a device at a loss, if they know they can recoup the loss through a small monthly service fee that might add up to multiples of the upfront cost over time.

As such, iRobot is looking for ways to increase revenue on the back of its deployed base of Roomba vacuum cleaners. While the devices themselves are not exactly cheap, ranging from $375 to $900, the ability to add another few dollars per deployed vacuum quickly adds up for iRobot. The company has sold over 10m of the things, so even $5 per Roomba would greatly boost its margins, for very little effort or cost on iRobot’s part.

Of course, this relies on not damaging the brand value and potential future sales, which the company looked like it was running the risk of. The public misunderstanding stemmed from CEO Colin Angle’s interview with Reuters, in which it seemed that the company was planning on selling the mapping data that its Roombas collect to the likes of Amazon or Google – saying that the company could reach such a deal in the next couple of years.

What followed was a clarification statement to ZDNet, in which the message shifted in tone, moving from “not formed any plans to sell the data” to “iRobot will never sell your data.” Angle noted that customers have control over the data generated by their Roombas, and that “information that is shared needs to be controlled by the customer, and not as a data asset of a corporation to exploit.”

But Angle adds that “you may also want your robot to work with other connected devices in your home. For this to work, we will also require your permission, and we will always ensure secure means of communication between devices.”

There are some decent reasons for using the Roomba’s mapping information with other smart home devices. Discussions of using it to better optimize audio systems to a room are perhaps quite niche, but being able to tell something like an alarm or connected camera system to expect an incoming Roomba might help avoid some false positives or erroneous alerts.

On a wider scale, one Roomba’s mistakes might help all other Roomba’s to work better, providing a better customer experience on the back of the crowd-sourced operational data. On the face of it, this sounds like the data-sharing that many consumers are concerned about it, but if this data is sufficiently anonymized, then surely the trade-off of improved performance would be worth it to the consumer.

That is, of course, assuming that the data is in fact anonymous – but there have been a number of studies that prove how easy it is to find needles among apparent haystacks of apparently anonymous data. However, the floorplan of a home is less sensitive than credit card numbers or web-browsing histories – so again, if only that data is shared, then the improved performance should be ‘worth’ the sharing.

But the Roomba’s do have cameras, which use images to help with navigation. These images are apparently never sent to the cloud applications, but while such images would still be part of the floorplan and navigation function, they would be much more sensitive to the consumer.

In our view, the sharing boils down to being able to provide something in return for the share, and ensuring that the sharing is non-intrusive and in some way relevant. Roomba collecting information on WiFi connected devices and then trying to sell that dataset to someone like Amazon or Google, to improve their marketing, would be seen as exploitative, even if it was buried inside the terms of service. However, collecting and sharing floorplan information to improve the Roomba’s function should fall into the ‘acceptable’ definition of this sharing.

Speaking to TechCrunch, after the ZDNet clarification, Angle has said that “we’ve had initial conversations around rooms and spatial context, but it’s relatively early. I don’t want to overstate the depth of the conversations that we’ve been having, but we’re certainly on a collision course with others in the home because there’s an increasingly recognized need for spatial context.

But such a deal could legitimately improve a smart home service for a consumer, and it is the kind of improvement that is hard to attach a direct value to. There is a danger of these SHaaS devices being a little overwhelming. Consumers are not going to tolerate a dozen separate $5 monthly fees to keep their smart home devices running, and over time there’s significant opportunity for consolidation in this respect – perhaps culminating in companies that prove the Smart Kitchen as a Service, or a garden, bathroom, or living room.

Whole-home offerings would, of course, be on the cards too, and service providers like ISPs, MNOs, or TV providers are well-placed to embrace them. There is also the possibility of products shifting away from upfront fees for products, and turning to a continual monthly fee for a thing, but that mechanic is suited better for more expensive appliances or products, where a business could justify the expense of sending a bailiff to collect the thing if a customer stops paying.

Rental properties might also look to revenue-share agreements with B2B smart home providers, as a way to bundle new connected appliances or systems as part of a new tenancy agreement. Likely included as part of the rent, this is another area that would require changes in the typical behaviors of both landlords and tenants, and is probably more suited to larger landlords of apartment blocks – rather than small private landlords that might manage only a handful of properties.