With the annual Consumer Electronics Show starting next week, the news media is filled with marketing hype about the elusive Smart Home market, which is nearly as misdirected today as it was in 1957. Watch this video of the RCA-Whirlpool® Miracle Kitchen and ask how many of those products you can buy today.
Whirlpool was wrong 58 years ago when it introduced its Miracle Kitchen, and I think it’s wrong about home automation today too. All the company did was offer a Smart Home vision that never crossed the chasm to mass-market adoption. The problem wasn’t a lack of sensors, actuators, electronics, or network standards, but marketing and the inability to understand consumer needs and wants, or the ability to show a value proposition or deliver a solution with ease of use and affordability.
I still have seen no company that “gets it” or knows how to deliver the smart home vision — not Whirlpool and not Apple or Microsoft or Google or Philips or BestBuy. A lot of innovation has occurred in the last 50-some years that is getting us closer to that vision, including smartphones and the Internet of Things, but there are still many marketing barriers that these companies don’t seem to understand.
Whirlpool’s Vision Update from CES 2016
I find it interesting to contrast the 1957 Whirlpool vision above with the video shown here rom CES 2016. At least Whirlpool still “trying” to innovate, but they still don’t get the fallacy of integrating so much digital technology into appliances that otherwise would last 10-20 years, causing them to go obsolete much faster than they should. If that’s by design, just know that what seems good for profits is not always good for consumers.
My perspective & bias
Toward the end of my 30-year IBM career (in the early 1990s), I introduced the company to the Smart Home market, represented its interests in several industry standards initiatives, and later retired to start CAZITech, a connected home consulting firm. Two years ago I wrote with skepticism about the Smart Refrigerator, and the video above reminded me of that. The problem with smart appliances is that rapid advancements in computer, networking and storage technology enable new features at lower cost each year, making slightly older products obsolete more quickly than ever before. But do you really need the latest features when appliances work fine for a decade or more? Do you really need to replace your perfectly good refrigerator with a new $9,000 model because it features a built-in, color touch-screen and wireless Internet access? What does it do differently or better to justify that cost? And wouldn’t you rather use a touch screen at the kitchen table than standing at the fridge? In my article I describe how to make an even smarter fridge with your existing iPad and less than $40 in parts. The lessons presented apply to medical devices, cars, and other smart home products too.
LG & Samsung continue to add unnecessary electronics
So what will it take?
BestBuy has the “potential” of becoming a big player in connected home products, but there’s still more work to do. On the plus side, having a large inventory and a website to help people choose is an advantage, as is having a dedicated section of the store. But still, appliances are in a different section, and consumer electronics are somewhere else, so it’s hard for consumers to actually experience the whole value proposition. I doubt that store sales staff will know much about home control systems, meaning consumers will be mostly left to explore on their own, so self-service point-of-sale education is key. I do like that their website includes buyer ratings and could evolve with instructional videos to show what automation can be done with little effort, or a lot. Those videos could even be made available in store kiosks.
A bigger problem for BestBuy, however, or anyone else trying to fulfill the Smart Home promise, will come when if-then-else automation evolves into smart agents with artificial intelligence that learn, because it will then be even harder to show the value of learned behavior that evolves with time.
Smart Homes versus Automated Homes
Consumers like the benefits of home automation but not the cost and effort to automate. No wonder this industry remains stalled, even after some 58 years. It just takes too much expertise to program the automation; there are too many different types of things to automate; and we rely too much on rules-based programs (e.g. “If this happens, then do that” or “Do this at this time”). Even the most skilled technicians can’t anticipate future needs, so adding new devices or making minor lifestyle changes can mean adding or changing the rules. While that ensures continued employment for technicians, it limits market expansion.
In contrast, a truly “SMART” home would learn on its own, discover new devices, notice changes in activity, and adjust without programming. That would need a sense of what’s going on, such as with human-like sensors that listen, see, feel, and smell. It may also need the ability to question occupants and carry on a conversation.
“I notice that you regularly do this,” the home may ask. “Do you want me to do that for you automatically?” Or, “I heard a strange noise (like a crash or glass break). Are you OK? Do you want me to call 911?”
This was exactly the vision of Merasoft, a small Utah-based company that ended up going out of business before completing its system. I was on their advisory board before retiring from IBM in 1999, and I saw their system in action. It worked well, but they had to invent technologies that were outside of their core competency. They needed help from big companies to fulfill their dream, which was based on neural networks and learning agents.
To listen, Merasoft invented room sensors with microphone arrays and digital signal processing code to distinguish between different sounds, such as door bell or glass break, and to better understand human speech — close up, far away, or around the corner. They also modified IBM’s ViaVoice speech recognition technology to be multi-user so people in different rooms could interact simultaneously.
To see, the company developed room nodes with CMOS image sensors that could recognize light or darkness, faces and even gestures. The house, for example, could notice that the husband seemed to sit on the sofa when he got home from work and turn on the TV to CNN. It could then ask if he wanted that task automated when he sat there at that time of day.
Merasoft also believed smart home sensors should include the ability to feel (temperature & humidity) and smell (smoke & CO2). And with the ability to sense surroundings and learn, there was no need to program new rules. The Merasoft project was nearly 20 years ago, and I’ve seen nothing since that even comes close.
Merasoft’s competency was in neural networks and learning agents, and most of their engineering talent was from Brigham Young University. IBM liked their use of CMOS image sensors for surveillance apps, gesture recognition, and facial recognition and had already done much work in those areas. IBM also liked their DSP-based microphone array, which could distinguish between different voice or various sounds and eliminate noise, echo and reverb to improve speech recognition. That way, voice commands could be clearly understood from a distance or distinguished from others in a party. Other sounds the microphone could recognize included glass break (a burglary?) and fall detection. And by knowing who else was at home, the system could act autonomously if the occupant didn’t answer voice inquiries. There was no longer a need to say, “Help. I’ve fallen and can’t get up.”
Merasoft became an IBM business partner and made extensive changes to IBM’s ViaVoice software to support multiple simultaneous conversations and to improve the text-to-speech voice quality so it didn’t sound like a robot. With just 30-minutes of capturing a specific voice pattern, the system could be made to sound like anyone, including me, and I’ve still not heard anything as good since.
The Challenge of Marketing Smart Homes
Think of the challenge of selling smart home technology, where potential customers need to “experience” the value proposition in order to understand the benefits. That can’t be done in retail stores like BestBuy. Consumers need to see it in a real home.
Merasoft had a brilliant marketing plan and choice of marketing channels, but I won’t discuss more here. Just think, however, about what it will take to move the Home Automation industry across the chasm separating early adopters and early majority customers and then toward the vision of a truly Smart Home. Who (what companies) might lead such a move? What partnerships would they need? How could they build a “camp” of partners supporting their technology? How could they make sure that their technology became standardized?
Understanding Market Requirements
Rather than develop products and then find markets for them, isn’t it better to first understand market needs? Now that doesn’t mean asking consumers what they want, because they may not even know. They almost certainly won’t know what’s technically possible or feasible. Instead, marketers should study what they do, enjoy doing, would like to do, hate doing, and wish they didn’t have to do.
A reason I joined the HomeRF Working Group as their marketing chairman back in the mid-1990s (before Wi-Fi) was because I was so impressed with Intel’s Ethnographic Market Research. They would send a team of (1) psychologist, (2) anthropologist, and (3) sociologist into homes to study their activities and values. They did NOT send technicians! And their research fed into a Market Requirements document that then fed into a Technical Requirements document and guided standards development and the HomeRF specification.
Another example of where NOT to get market requirements comes from IBM’s PC business. IBM had little success in consumer markets because it would ask retailers what features should be included in the next PCs, as if they knew. This meant that IBM got the same answers as all other PC manufacturers, from retailers that didn’t really know either.
So why did IBM have such great success with its ThinkPad line of laptop PCs? The ThinkPad team looked to IBM Research for product ideas (e.g. TrackPoint mouse replacement, Butterfly keyboard, and a hard disk with accelerometer to detect a fall and retract R/W heads before hitting the ground). ThinkPad market research all started with understanding user behavior and needs, followed by testing prototypes with customers to see if those needs were met.
Home Control in the Year 2020?
In the following video (from CES 2017?), Mitchell Klein, Executive Director of the Z-Wave Alliance, speaks to members of the Custom Electronic Design & Installation Association (CEDIA) about trends affecting their industry. He makes a compelling case that embedded smarts in the Internet of Things (IoT) is changing the landscape and industry leadership.
My issue with Klein’s presentation, like so many others I’ve seen, is that what’s trending and technically possible often is not yet a “must have” desire, and that’s what it will take to cross the chasm from market niche to mainstream adoption. Klein gives insight into what will be possible in 2020, but predicting market adoption still seems elusive.
I hope you were able to watch this with a bit of skepticism, and maybe you noticed that he cited market research from several different companies, cherry-picking the most positive projections. He did, however, touch on many of the drivers of market adoption, but he failed to address the inhibitors. That can be critical, I learned, after writing a major market research report on Home Controls in 2004 for Parks Associates. The chart below shows what I saw as the opportunity if only the Home Control market could Cross the Chasm separating market niche in high-end new homes (just 68K in 2002) and mainstream adoption in the 100+ million occupied U.S. households. Needless to say, we still have not crossed that chasm, and I’m not holding my breath.