Machine learning enables computers to scan huge datasets and make predictive analyses. This is becoming increasingly common in Internet of Things (IoT) enabled devices. It will likely become a household phenomenon in the very near future. Privacy concerns aside, questions loom about how this new phenomenon will impact the consumer as well as the product supply line.
Much of the predictive data will come from the local network of IoT devices based on user experiences and patterns of behavior. However, the integration of WiFi/internet connectivity means that data can come from anywhere: say, a weather service. For example, an IoT-enabled window may learn that rain is in the forecast because it’s getting a signal from a nearby weather service. Alternatively, if the user sets a pattern of opening the windows when the temperature reaches 65 degrees, the windows may learn this and start to open automatically. These actions can serve as triggers for other IoT devices on the local network, such as an air purifier that turns on when the windows are prompted to close.
The question becomes, how much of our lives can predictive analytics automate? There will obviously never be a point when machines can anticipate ‘impulse’ buys, unless they are triggered by some measurable phenomenon. But as seen with the Amazon Dash Replenishment Service, consumers can restock their household items automatically. The Internet of Things will play a huge role in realizing this vision, as it can measure consumer needs in real-time.
Predictive Analytics to Calculate the Rate at Which a Product Is Consumed
Anticipating customers’ needs is one of the most important roles predictive analytics will have in business. Let’s use a hypothetical IoT-enabled paper towel rack to explore this process. Suppose we can load our paper towel roll into the rack and press a button that says “full”. The rack can sense the weight of the towels, a weight which then becomes the nominal starting point for calculating how much of the roll is left. Once the roll is diminished to a certain extent (say, 25% is left), the algorithm in the rack can check inventory. To do this, it can subtract the number of “full” button presses it has counted since the last time it ordered paper towels. And here’s the cool part: it can calculate the rate at which paper towels are being used, and predict a time to place a new order that will get new paper towels to your door shortly before the last roll runs out.
This sort of calculation can be applied to a variety of consumer products. Soon, devices will proactively order everything from coffee to trash bags. Want to build an entire month’s worth of meals, but not buy a month’s worth of food at the same time? Just load your recipes into the refrigerator and it will monitor supplies of the necessities and when you will likely need them. Oops, it’s almost July, time to order fireworks from the local internet-based fireworks store!
Predictive Analytics Makes the Supply Chain More Efficient
On the supply side, analyzing how many orders come through and when can help minimize stocking concerns. The ability for software to predict things could impact every aspect of the supply chain for the consumer packaging goods industry. From the impacts of a major storm on shipping times to increases in labor forces for holidays or other events, this type of analysis may help streamline everyday logistics as well as crisis management.
But where does it end? Taking this to its logical conclusion, one can imagine a person that can go through their day, never interacting with an appliance. All the needs of the consumer can be met preemptively, or at least very quickly.
Image courtesy of pixabay.com
