Selling Home Automation Systems that Learn and Adapt

Moving forward, look at your control vendors through this lens and ask them questions based on the assumption they're heading in this direction. If they look at you funny, that's not a good sign.
Author:
Publish date:

Imagine you're back in elementary school. Today's the first day of school and your teacher shows you how to do some basic math problems. How exciting! A few days go by and teacher decides she wants to test your retention. Pop quiz! What's 2x2? You know the answer, don't you? Sure! You paid attention in class and learned how to solve the problem. Instead of giving you a fish, your teacher taught you how to catch them yourself.

Let’s pretend our home technology solutions are in class with you. You earned an A on the quiz. How did the fancy home automation system you installed do? A lousy F. Why? It doesn't make any sense. After all, the fancy home automation system cost $200k to install. An F? Here's the difference; at the end of the day, you have the capacity to learn and adapt from the environment around you. Our fancy home automation systems? Not so much.

Let's review what our solutions can do before we start throwing the baby out with the bath water. Current home technology permits events based on occupancy, sunset/sunrise, geo-location (limited), and anything we directly program ourselves. Imagine if TiVo or Pandora worked this way. Why do these companies succeed where others don't? Because they learn from their customers and adapt accordingly. TiVo monitors what you watch and makes recommendations based on activities you're already engaged in (viewing). Pandora asks you what you're in the mood for (artist, song, genre, etc.) and gives you the opportunity to thumbs up or thumbs down (though not required) to improve your stream. If learning, adaptation, and personalization are really the keys to home technology solutions moving forward, we're off to a slow start and for the most part our industry seems oblivious to this very obvious trend.

Image placeholder title

ThinkStock


Data are the common threads to any personalization engine. Let's also not forget to include a way to parse and analyze those data intelligently (a really good algorithm). Moving forward, look at your control vendors through this lens and ask them questions based on the assumption they're heading in this direction. If they look at you funny, that's not a good sign. Examples of companies heading in this direction are outfits like Alphabet/Google (Nest), Apple (HomeKit), and Alarm.com. They gather all the data and are slowly building algorithmic behaviors around basic activities like HVAC and lighting control. That's a really good sign. Knowing the home control market is projected to balloon to $60 billion over the next few years probably has something to do with it as well.

Consider the commercial sector a leading indicator here. Companies like Johnson Controls and Simplex Grinnell tout building intelligence. Do you think a facilities manager is interested in paying an integrator by the hour to write code to approximate intelligence? The winner in this segment wrote the best algorithm. Think of the algorithm as secret sauce. Who in the CEDIA world currently has any secret sauce? Every one of our systems is programmed differently (even within the same manufacturer) and left in many cases to the tastes and whims of our high-level technicians who often configure customers’ solutions to match their own tastes. Because of this, we're left to our own devices when it comes to supporting these systems. How many times have you been to a manufacturer training session and told, "We don't support programming"? It's been gratifying to see that approach change in recent years, but needless to say there's still a ton of potential for finger pointing when things don't work as advertised.

Services like If This, Then That (IFTTT) and their ilk seem to be offering the best opportunities to fake the secret sauce while we wait for manufacturers to catch up. Then again, do we really want to get into supporting our customers using a cloud service cobbling together disparate devices together? Wouldn't it be better to have one platform where all devices connected together with one app... wait a minute. I feel a Moebius strip coming on. The very same promises we started making to our customers many years ago are the same ones that the new Internet Of Things (IoT) manufacturers are spinning.

Does this sound familiar? "Imagine you wake up naturally in the morning because your lights came on gradually timed with your sleep rhythm. A pathway illuminates from the bedroom to the kitchen and the coffee is already brewing. Your favorite news show is automatically cued up on the TV and favorite songs play during your morning shower. You leave for work (your house tells you to wear a jacket since it's cold outside) and your home security system sends you an alert that you forgot to arm it."

I first heard that one in 2001. That's right. It was possible to do all these things in 2001 (and well before). The big difference is that the 2001 system had to be programmed and told what to do every step of the way. Programming and scheduling. The scheduling part came the closest to automation, but never learned or adapted. Fast forward to today. Typically, most large-scale home control systems still need the same degree of programming to deliver the sizzle described above. Nipping at the heels of the big boys are algorithm-driven solutions from the likes of Google, Apple, and Alarm.com.

We never learned to program our VCRs but loved it when TiVo started suggesting shows we might like. The VCR analogy is universal. Consider the huge changes in retail over the past few years. Has Amazon sent you any recommendations lately? Did it amaze you how accurate those recommendations were? If my home technology system sees me doing the same thing at the same time everyday, shouldn't I expect a recommendation? Something like: "Hey! We noticed you turn the porch light on for five minutes each evening and then off again 30 minutes later. Would you like us to do that automatically for you everyday?" You tap, "Yes" and keep moving.

While learning and personalization features may not be ubiquitous in home technology now, they sure will be in the not-so-distant future. The best way to figure out how they fit into your product mix is to read everything you can find on trends and test products that look promising. Very few will make the grade for our customers, and that's OK. Our customers need to know what's available, why it is (or most likely isn't for them) and what our long-range plan is to keep an eye on emerging trends. The worst conversation you can have with a customer is one where you didn't know about the product first, or you're using negativity to sell.

We're slowly approaching the point where true intelligence will make its way into our homes, bringing with it adaptation and personalization. Ray Kurzweil (author of The Singularity Is Near) says a computer will be as smart as a human by 2023 (and a lab rat by 2017). What part will you play in the personalization game?

Stay frosty and see you in the field.

Related