In 1982, a group of grad students in Carnegie Mellon’s Computer Science Department – desperately in need of caffeine and fed up with the erratic volunteer restocking schedule – had a fit of innovation. They connected the vending machine to the internet so they could check from any computer to see whether or not Coca Cola was available and if it would be sufficiently cold. The design, born out of frustration, pragmatism, and pure geekery, is considered one of the first Internet of Things (IoT) devices. By 2020, Gartner estimates that 13.5 billion “things” will be connected to the internet.
Your iPhone can already “talk” to the IoT-enabled lights in your house and your car can communicate with your IoT-enabled thermostat, but what if these connected products could also communicate fresh and valuable data to manufacturers, brands, and potential consumers in real-time? IoT devices have the potential to accelerate R&D and time to market for new products, but whether they actually do so depends on the data that they collect from user interaction. We’ve barely scratched the surface when it comes to applying the huge amount of data these devices generate.
Harvesting actionable data from IoT devices could help brands build better products that cater to their customers’ needs. For instance, a “smart” teakettle or refrigerator could share data around how often consumers drink tea or run out of milk. Deeper knowledge about these usage and purchasing patterns could make a huge difference in how grocery retailers stock their shelves. Similarly, if drivers disliked their car’s GPS performance in certain weather conditions, they could share real-time feedback and reviews that brands could use to improve the product (and potential consumers could use to avoid bad purchases).
What’s holding us back from having this kind of real-time feedback and brand enhancements today? Unfortunately, we only need to look in the mirror for the answer. Consumers have been reluctant to share this much real-time data with brands. Understandably so, since many of today’s IoT devices live in the home – a traditionally protected and private space. To truly tap into the IoT devices’ data, companies need to ensure they protect and value the data and clearly communicate their efforts in order to make customers feel safe.
While IoT devices collect a ton of data now, it’s virtually useless without context and structure. Without context and structure, data cannot be categorized in a standard way and provide additional use and meaning. Instead of the existing free-text data dumps, cleaned and structured IoT product data can be easily gathered, stored, queried, and analyzed.
The need for accurate and thorough product categorization, matching, and coverage will continue to intensify as the number of connected devices continues to skyrocket. You can collect all the data you want, but it’s a useless waste of computing power unless it’s actionable. Cleaning, structuring, and acquiring the untapped wealth of product information generated by IoT isn’t possible without AI and machine learning. For example, when a consumer shares some data about a coffee machine, all integrated systems need to recognize it as that particular coffee machine (same brand, model number, etc.).
As data from the Internet of Things continues to get better and consumers feel safer sharing it, it will ultimately benefit consumers and brands alike. And we’ll stay stocked with caffeine, just like they did at CMU 34 years ago.