KM and the Internet of Things
November’s KM World Conference is like a homecoming for many of us. Similarly, the Internet of Things (IoT) can be a sort of homecoming for knowledge practitioners. As with the homecoming to our KMWorld family, with IoT, we’ve found many familiar rituals and some unfamiliar. With the Internet of Things, we may now breathe new life into disciplines of information processing, organizing, sense-making and collaborating. It is indeed a special moment in history where the old and the new converge to create the old-new world of IoT.
In Sam Ransbotham’s article “Ready or Not, Here IoT Comes”, he recalls the convergence of capabilities that seeded the Internet. Once again, with IoT, technical, sociological and leadership capabilities are converging. He writes: “First, the cost and physical size of sensor technology have dropped such that they can be incorporated into most items. Second, widespread communications infrastructure is in place to allow these distributed components to communicate and coordinate. Third, once again, savvy innovators are showing the rest of us the possibilities from the data they collect. With these in place, the smoldering potential of IoT may be ready to catch.”
We get it. It’s not just about “technology innovation.” Technology is tremendously important, as it was during the Internet boom of the late-1990s. However, it’s the information and the people, not the technology, where players will win or lose over the long term. The numbers are staggering. For example, McKinsey researchers Stefan Heck and Matt Rogers forecast that the electric grid will produce 5 billion data points per day (Resource Revolution: How to Capture the Biggest Business Opportunity in a Century, Melcher Media, 2014).
At the Columbia Information and Knowledge Strategy Master’s Program, we say that it’s about two disciplines—information management (sense-making, organizing, interpreting and acting on data) and collaboration (a sort of radical interdependence across the ecosystem). Information management and collaboration are the yin and the yang of KM. We knowledge practitioners have learned (and relearned) this for two decades. Can IoT learn this?
Comparing the knowledge and IoT data lifecycle
In practical terms, information management focuses on the information lifecycle, while collaboration focuses on people sharing insights and action. Like knowledge platforms, IoT platforms have the potential to activate insight from data to improve processes, use resources more efficiently and improve revenue. And, like knowledge platforms, failure to collaborate across the ecosystem can result in many an investor’s red ink. (Download chart 1, also on page 8, KMWorld, April 2016, Volume 25, Issue 4).
In a typical industrial scale IoT model, such as that which Janakiram MSV’s writes about in“AWS IoT: Amazon’s Knock Out Punch to the Competition”, pundits describe IoT as “components” such as sense, route, snapshot and analyze IoT data.
Michael Porter and Jim Heppelman add a business perspective to each component in “How Smart, Connected Products are Transforming the Competition”.
Combining those two models we compare the data lifecycle in IoT and KM. In both cases, genuine insights emerge, and business or societal impacts ensue.
Across the seven elements—design and strategy, sense, route, snapshot, protect, interpret, collaborate—you will see that there are many similarities between IoT and knowledge practice. At the same time, there are real differences, and they could be deal-breakers. At first glance, the knowledge practitioner has focused on eclecticism and flexibility, tweaking search algorithms and vocabularies, adding data sources and refreshing dashboards to correspond to changes in opinion or market pressure. Meanwhile, IoT’s hardware presents significant costs to switch new data sources, applications and regulations. The hallmarks of IoT are reach, volume, variety and velocity. We think the differences are more textured than that, and that one could leverage KM’s 20-year history to transform IoT.
1. Design and strategy:
Knowledge practitioners have long been good at designing with the end in mind. (“How can we decide with confidence?”) We’ve taken a broad “systems” view of a problem and pulled together dissimilar inputs into a solution. We’ve been weighting and co-mingling factors for years in our search algorithms. Whereas knowledge practitioners could track a larger variety of entities (e.g., stock price, news, social feeds, M&A rates), IoT can track data on thousands of nodes (e.g., electric grid). The winning formula has been for knowledge practitioners to get the right mix of factors, with tthe right weight, and tweak long into our product or service development lifecycle. The IoT practitioner, on the other hand, has to work upfront, planning sensor placement or updates infrequently, thoroughly weighing hazards, such as heat, vibration, water and chemistry. Reprogrammable devices are becoming more common in industrial IoT, but for consumer IoT, that is limited. In each, the barriers to entry and exit are higher than with the knowledge cycle.
Knowledge systems “sense” by collecting from the news, social media, operational systems, experts or e-mails. By contrast, IoT sensors spend a lot more time analyzing, checking tolerances, manipulating and integrating data from analog processes before they collect. With modern high-volume social media or processing tools, detailed processing in KM is more likely to occur downstream. In other words, KM’s intelligent processing is mostly at the end or near-end of the pipe. Meanwhile, IoT’s intelligent processing is in the pipe (interpreting state changes, adding other data and then pushing out as an alert, e.g., to a phone, saying, “Your defect rate is too high”).
Knowledge-based systems transmit data to an aggregation hub or analytic hub (e.g., a dashboard or a business analyst). But often data reach their destination and cannot be understood. For example, a deck of PowerPoint slides might arrive without context, and the assumptions may feel like a black box to the receiver. We all know stories of where the system failed to transmit metadata, confusing or failing to “sell” to content users. Routing could include multiple people collecting and transmitting content as a team. In KM, it can be a creative game of telephone tag. Interpretation issues could ensue based on timing, relevancy, recency and consistency of assumptions across the different sources. Comparatively, routing is much cleaner in IoT (except where transmission is interrupted). For IoT, transmission to the aggregation or analytic hub is similar, but the task of augmenting the data with metadata is generally automated. For example, temperature readings are transmitted with date, device ID, location, etc.