KM and the Internet of Things

4. Snapshot:

When we talk about snapshot, we refer to two things: the snapshot of the data in transit (or just before transit) and the snapshot of the history that can be used as a point of comparison. In KM, the snapshot is the content backed up at the source. The history or the source of record could be institutional memory, such as standard operating procedures, best-known methods or annual reports. In IoT, historical data are perpetually similarly summarized, such as trend-derived tolerances on an instrument. The more sophisticated snapshot is called a “state device.” In Amazon AWT, for example, that is a “shadow service,” which stores the “last known good configuration” of the device, and it will persist even if the device is down. Snapshots can be used for predictive analysis to forecast the consequences of events, e.g., future failures. That intelligence in the pipe is less pre-planned in the KM example. For example, an investment idea may undergo several transformations from the reporter, to the analyst, to the consultant.

5. Interpret:

Modern knowledge practitioners visualize, combine, summarize and reach for valid conclusions—a story, even deep learning. We use heuristics, interpretations, inference and comparisons to discern the relevance, context or impact. The target is unequivocally the decision-maker. With IoT, the target might not always be human. Heck and Rogers quote Ericsson as saying that, by 2020, 80 percent of the 50 billion devices connected to the Internet will be talking to each other. KM’s greatest flexibility is in the “interpret” step: The sum total of information and analysis isn’t fixed. We can bring in more sources, convene experts, create unusual combinations. For example, we might be gathering best practices and benchmark data on the employee hiring process and decide to add sentiment analysis from corporate communications. You might say that KM can pick up hitchhikers almost with impunity. IoT is more rigid. Added devices are like new passengers on a train: They need IDs checked, carry-ons X-rayed and tickets scanned.

6. Protect:

The three main tools of IoT’s security and privacy are user authentication (securing connected devices), encrypting data or metadata in transit, and securing layers of the application. Security and privacy are similarly obsessions of knowledge practitioners. KM’s encryption, authentication and access control are sometimes so strong as to discourage people from engaging in the info lifecycle at all. We find ourselves in the messy business of policy and training (for example, preventing trade secret leakage). The real difference is that KM data may be actionable by an intruder without much processing. In the IoT case (where sensors number in the millions), considerable processing may be required to interpret leaked data. Damage can be done, and KM lessons are transferable.

7. Collaborate:

KM defined as “information management and collaboration” means that we don’t shy away from the difficult questions of identity, power and language. Knowledge practitioners engage throughout the knowledge cycle, not just at a single point in time. For example, a cost center owner may need to negotiate the meaning of “product” with the marketing manager. In a similar vein, data definitions and aggregations at the sensors must conform to the analytic models that route or trigger downstream events. However, humans are largely absent. The data definitions need to be laid out well before the moment of truth when the temperature goes out of bounds, such as the temperature in the fateful Challenger disaster in 1986. On the other hand, modern machines can “learn” and increase systemswide intelligence. Ultimately, the ecosystem of IoT entities—sensing, routing data, applying rules, triggering actions—can be far bigger than we humans can comprehend. (Download chart 2, also on page 9, KMWorld, April 2016, Volume 25, Issue 4).

Collaboration is a core discipline we knowledge practitioners can bring to IoT. In IoT, boundaries of competition are only vaguely demarcated. With the creation of smart ecosystems—such as thermostats, lighting systems, air conditioners, weather models and energy planning software—now there is a delicate new balance of competition and collaboration. Porter and Heppelmann point out in “Smart Connected Products” that there will be an unpredictable ecosystem, all players vying for the dominant platform. That can cause rivalry or force collaboration on standards.

Ransbotham adeptly evokes the unsettling power shifts of IoT players, such as consumers’ commanding a premium for their data:

• Smart products may be leased—think: ZipCar—with upfront spend and low switching costs.
• Device owners or aggregators may sell data, such as concertgoer data to casino construction firms.
• Consumers may be paid (or pay a discount) to provide their data, such as commuting and carpooling data to employers and municipal governments in the Luum model.
• Being the standard can yield revenue, such as Google Android being used by an automotive consortium of GM, Hyundai, Honda and Audi.

In some industries, margins may be higher for the data owner. In other industries, ubiquitous sensors mean high margins for the hardware providers. Finally, in other industries, adaptive, deep-learning algorithms coupled with big bets on things like commodities could yield outsized returns. And, all that occurs in a delicate balance, where today’s competitor could also be tomorrow’s partner.

This is just the type of ambiguity in which we knowledge practitioners thrive. For the last 20 years, we’ve been exhorting community members to think about the collective over the individual. The true value of our knowledge legacy is our ability to convene and adapt the network. We have led differences in affiliation, capacity, skills, mental models and wallets. We’ve built innovation capacity by being adept facilitators, translators and tools providers, even negotiators of intellectual property. We’ve approached collaboration with a sense of spaciousness and curiosity.

Convergence

We learned by looking at the seven degrees that KM’s knowledge lifecycle is more flexible and more iterative, while IoT’s lifecycle has a larger analytic data volume and a larger geographic footprint. Both disciplines require critical attention to privacy and security for data and analysis in transit. IoT beckons: It can bring safety, comfort, quality and conservation to the lives of thousands of people and to the planet.

We would argue that IoT is a game knowledge practitioners should be playing. Knowledge practitioners’ critical offerings are our deep knowledge of the structure and social life of information and the importance of intentional collaboration. It’s a big deal to move from the digital to the physical. Much of our profession has been taming knowledge workers’ knowledge-sharing and consumption habits. Yet, few of us have managed fleets of “dumb” sensors that can’t think on their own. We believe that the big move for us knowledge practitioners is to the pipe, both figuratively and literally. In the pipe, IoT draws heterogeneous players, the physical systems and large fixed costs. But the pipe also brings rich, varied analysis. This should be welcome fuel for many a knowledge practitioner. It will put our convening and analytical skills to the test.