So What? Why Smart Systems Are Not Selling.
Our world is full of sensors and vendors fencing various analytics tools that can take that data and render quite good, actionable, advice. As I mentioned in my previous post, there is an inverse correlation between the volume of Thought Leadership talks on a subject, and the amount of products being sold in that subject area.
So why are we still hearing people huff and puff on stage doing Thought Leadership talks about “smart” systems?
Why aren’t people actually selling “smart” systems? The reason, I would posit, is reaction time.
“Smart” products are pitched on the promise that they will dramatically enhance the responsiveness of the systems to which they are connected. That promise ranges from responding to external factors (such as changes in supply or demand, in the case of a “smart” factory), to responding to internal conditions (such as scheduling maintenance based upon a vibration appearing in a beam).
It is an entirely reasonable pitch: the technology is there, the math has been around for over five decades, and the only new sliver being added to the “smart” definition that is causing all this hubbub is the ability to act upon feeds drawn from dispersed equipment.
The reality, however, has been that most of the products in the field today haven’t changed response times all that much.
Take, for example, “smart” water. Any Johnny-Bag-O-Donuts developer could write a program that summed the flows on building meters and compared them to the flow rate at the closest source valve to see if there was a leak in the system. So why is that not the norm? You might say “money”, but the answer I have heard time and again from people in the water business is “why bother? If it is serious, someone will call 311.”
The utilities I have spoken to all estimate a customer will dial 311 within 2 hours of a pipe breaking. So the program Johnny was trying to sell them means they can cut their reaction time by up to 2 hours.
That sounds brilliant until you remember they still have to put a team in a truck, drive to the site, wade through the chaos that had unfolded since they started driving...
...contact the police, cordon off the road, open the manhole cover, close the valve, rip up the road, and fix the pipe.
Placed in that context, 2 hours is a pittance.
However, offer the utility the ability to shut off the valve remotely, and they start to pay attention. How much less damage to the street would their crew find itself dealing with if the valve was closed within seconds of the breach starting?
There are 3 missing pieces to the “smart” puzzle, as it is currently conceived:
The first is a durable, efficient way to get data to and from equipment.
The second is a way to make that data exchange secure.
The third is a composite AI structure where compute positioned close to equipment can react to time-sensitive changes of state, while a centralized compute environment can process broad trends and work towards system-wide optimality.
The impediment to getting smart systems to sell are parts 1 and 2. The third is icing on the cake.