Internet Of Things, (IoT), can run from a sensor to keep track of watering your lawn to your car to who knows what. Pressing a button or just giving a voice command to have your vehicle pull up in front of the restaurant sounds like a great idea. So does having your car call for help if you hit something with it and you’re non-responsive.
Still, we’re not quite there in terms of all the smarts with smart devices just yet. It’s called the Internet of Stupid. Here’s a fun summary of some idiot savant servants, (perhaps one day our masters), have been up to over the past few years!
January 9th, 2017
San Diego’s XETV-TDT aired a story of a 6-year-old girl buying a dollhouse by asking her family’s Alexa-enabled Amazon Echo, “Can you play dollhouse with me and get me a dollhouse?” Other Alexa devices that heard the news report went ahead and placed their own dollhouse orders! Oops. The Story: TV Report on Accidental Amazon Orders Triggers Attempted Amazon Orders Across San Diego
December 26, 2016
While an Arkansas mother napped on the couch, her 6-year-old daughter managed to get her to thumbprint unlock her phone, at which point she ordered $250 worth of Pokémon stuff. Oops. The Story: Girl uses sleeping mom’s thumbprint to buy $250 in Pokemon toys
January 13, 2016
A Nest Thermostat update made some devices start going haywire a couple of weeks after the update. It seems this might be cause for just some discomfort. But if you think about being on vacation while your northern home in winter was set at temperature to avoid pipes freezing, then it’s a bit more of a potentially serious issue. What if you set it up to make sure your elderly parent – who really should be moving to a home anyway – is comfortable? Oops. The Story: Nest Thermostat Glitch Leaves Users in the Cold
June 23, 2015
The Wink Hub for home automation offers the benefit of coordinating multiple smart home devices. Until it got an update that basically bricked the unit. Oops. The Story: IoT Security: Gone in a Wink (Edit: FYI… to “brick” something technical like this is to basically take something and turn it into a useless brick. Apologies for any misunderstandings and thanks for the feedback.)
September 5, 2013
Trend net’s SecurView cameras for security and baby monitoring could let anyone who could sniff out a camera’s IP address take a look – or maybe even a listen – though the unit. Oops. The Story: Webcam Maker Takes FTC’s Heat for Internet-of-Things Security Failure
Lessons for Product Management
- IoT Best Practices are somewhere between infancy and adolescence. Not being at least on the smarter side puts you, your product, possibly your whole company and possibly others at various levels of risk. How much risk depends on what your device is up to and it’s capacity to cause problems.
- Data from IoT devices isn’t just a source of information. It’s potentially a data island with info now trapped in a standalone device. Beware of this and have means to deal with this failure mode.
- Mashups have been tying multiple sources of data and tools together for years. Now we have webs of devices into the mix as well. Just as framework on top of framework can begin to obscure root causes of issues, the complexity of “systems of systems” also creates complexity, multiple points of failure, and more risk. Make sure you have good maps for whatever physical and logical topologies with which you are working. Make sure your monitoring tools exist in the first place and are working at all potential failure points insofar as possible.
- Security has been shown to be lax in a variety of places within IoT. Make sure your crypto plan is protecting communications all along whatever paths you’ve got data flowing.
- Beware of adolescent IoT startups. Agile development, Lean concepts, Minimum Viable Product… all are great. All work wonderfully. But… not always. A Minimum Viable Product in this area has to include appropriate security. And a Minimum Viable Partner is one that’s got funding and obvious legs to make it into the future. Otherwise, you could end up with a bunch of unsupported devices. Consider contracts that account for potential business failure of partners.
- Hardware is so very different than software for reasons both obvious and not obvious. Plan for longer timeframes that you’re used to in software development. Plan for capital expenditures for line items you’ve never even heard of before and then triple your estimates. A Product Manager who leads the development and marketing of a plastic toilet brush probably knows more about production, molds/tooling, cost of goods, and so much more then a super sophisticated top software product person. Even with a great 3D printer in the office, working breadboard electronics and lots of iterations on prototypes, that’s not the same as lining up competent manufacturing in China, (where you’re probably going to have it built), quality control, shipping, customs and more. Not to mention cash flow issues for upfront tooling costs and initial product run.
- Even if you can update firmware, if you forgot to design a reset button in and you need one later, you are deeply, deeply… well… you fill in the blank. If you “brick” a device with an upgrade and a customer can’t recover it, what are you going to do?
There’s more. Sadly, it’s likely we’ll be able to add more very public failures to the short list above. It’s one top list you don’t want to join.