The Internet of Things (IoT) has been on my mind a lot lately. There are compelling arguments that giving “everything” some small amount of processing capacity, appropriate sensors, and a network connection – making everything “smart” – will usher in a multitude of new possibilities: improved asset use; better process optimization; easier and cheaper preventive maintenance; fewer device failures; and — by no means least – personalization of services and perhaps even products. With everything contributing some (or a lot) of data on performance, real-time analysis of the mass of information being generated becomes at least theoretically possible.
From this analysis can come new insights as we see patterns emerging that we can adjust proactively to address. We are already seeing a number of new approaches to physical logistics as a result of the ability to analyze buying patterns by geography, for example. So you get what you buy online delivered faster and at lower shipping costs because the supplier knows where best to locate distribution centers.
No advance in technology is without potential unintended consequences, however. Services like Uber (I am a reasonably frequent user and I really like what they offer) or Lyft not only make life more convenient, they also act as information gathering points for my (and other customers’) movement patterns (from where, to where, when, how often and so on). Mine this data in the aggregate and you can develop insights into transportation logistics and economics. But you can also build behavioral profiles at the individual level, and that’s where things get a little scary.
Suppose everything you do (shop; eat out; travel; walk your dog; use your phone; adjust the temperature of your home or office; rent a movie to watch online; order pizza….) generated a data trail (which of course it already does in most of these examples) and then suppose that (because everything talks to everything else) that all these individual data trails get combined into a “life profile” for each person, updated continuously as you interact with all the smart elements of the IoT and available as input to the predictive analytics algorithms of the world’s service and product providers. It’s likely that this will be both possible and common within the next decade.
Is this a good thing?
For the IT infrastructure industry it looks like it is. Lots of new capacity will be required: at the device level (sensors, low-power processors and network interfaces); at the network level (lots of bandwidth to handle the billions of mostly wireless data links); at the aggregation level (lots of servers for all those devices to connect to and lots of storage for all the telemetry); and at the analytics level (lots more servers and in-memory databases to run real-time analytics). This is an example of what always seems to be a paradox (it’s called “Jevons paradox,” after the 19th Century economist who first described the effect): as things (generally technologies) get more efficient we should need fewer of them (e.g. servers), but instead we find new things to use them for which actually creates more demand.
So we have at least one winner.
For the “innovation” industry (the idea and capital flows that create things like Netflix, the Nest thermostat, Twitter and Uber), there will also be rich opportunities: smart homes that ordinary people can actually interact with in useful ways; smart buildings that use less energy and provide better environmental management; smart cities that can manage traffic flows and services better than they do today; smart grid….a long list of things that have already been thought of and an even longer list that no one has thought of yet. Not every idea will work out, but I would hope for some big winners here too.
I am concerned that the big loser will be individual privacy and, to some extent, choice. If every part of everyone’s life is in a profile somewhere, it’s going to be hard to (a) hide and (b) beat the behavioral prediction algorithms. Maybe that’s a good thing – less opportunity for crime, fraud, misrepresentation, accidents and so on – but criminals (not all of whom are outside of the social fabric) are smart and well-funded too, and so far seem to be better at exploiting new technologies than the majority of “ordinary” people. As IoT technologies become more pervasive, the opportunities for exploitation will increase. It would be no fun to be locked out of your smart home because someone has hijacked the automation – and wants payment to unlock the door or fire up the HVAC.
Even routine commerce becomes an opportunity for an “exploit” such as personalized pricing – what I pay for something might be calculated in real time and based on a dynamic assessment of what I can afford rather than a standard price that everyone pays or a discount for loyalty or some other more familiar basis. Good for the merchant perhaps; not so good for me.
It should be possible to design and write software that makes the bad side of the IoT less probable – but this has always been possible and seldom happened. There really aren’t that many great software engineers in the world – and the IoT implies that we’ll need more great software, not less. So I’m not holding my breath that there won’t be problems, security flaws and other issues.
Ten years is a long time into the future and anything this far off has many uncertainties built into predicted outcomes. So we have time to get this right – or at least to debate the range of outcomes we might find acceptable. Better get started.
John Parkinson is an affiliate partner at Waterstone Management Group in Chicago. He has been a global business and technology executive and a strategist for more than 35 years.