Speaker: Dave Snowden, Chief Scientific Officer, Cognitive Edge
Session Description: Crisis management has moved from planning to a day-to-day reality. However organizations are ill equipped to manage a situation where we are dealing with unknown unknowables or have to deal with multiple Black Elephants (something that changes everything!) competing for resources and attention. What is the role of knowledge and information in a crisis? How do we gain attention to weak signals where anticipatory actions would reduce downstream risk and increase overall resilience. Shifting from Just-in-time. Just-in-case sounds like a good idea but it is far from simple and in a resource starved environment may simply not be possible. For the last few decades we have based practice in industry and government on an engineering metaphor, focusing on efficiency. This approach is, to quote Lincoln, Inadequate to the stormy present. Are there better approaches that we can adopt by treating the organization and society as a complex ecology? Would such a metaphor shift allow us to do more with less? Last year’s conference ended with a rousing discussion of creating resilience in organizations and society. They discussed transforming and revolutionizing the way we do business as we move into an uncertain future, how we satisfy our clients in an ever-changing technological age, and how, in our complex societies, we provide value, exchange knowledge, innovate, grow and support our world. Our popular, and sometimes controversial, speaker Dave Snowden has again assembled a group of experienced thinkers and doers who are capable of reimagining a future based on uncertainty.
[These are my notes from the KMWorld Connect 2020 Conference. Since I’m publishing them as soon as possible after the end of a session, they may contain the occasional typographical or grammatical error. Please excuse those. To the extent I’ve made any editorial comments, I’ve shown those in brackets.]
NOTES: [This is a long read but it contains a lot of food for thought.]
This talk explains how effective knowledge management can be a vital aid in a crisis. Snowden’s approach draws on his earlier work, especially Complex Acts of Knowing. This article was one of the first articles to focus on (1) levels of abstraction and (2) the role of informal networks as “a highly energy-efficient form of knowledge transmission”.
- He is working on a European Union handbook on how to manage in a crisis. It includes a five-step process for getting out of a crisis and how to use distributed networks and your own employees to do that.
- They are also working on post-conflict reconciliation. Given the current political climate around the world, they believe this will be necessary to create a stable market.
What’s Wrong with KM? (Part 1)
- KM’s Core False Assumption: if we just surface the information (by asking them to write down what they know, contribute to a shared repository, generate lessons learned, participate in a community of practice, etc.), then magically knowledge will flow throughout the organization.
- Knowledge management professionals have been trying this for the last 30 years but it doesn’t work.
- Why doesn’t it work?
- They assume information flows automatically between people without thinking first about the nature of the information itself and how it works.
- They are ignoring the impact of levels of abstraction.
Levels of Abstraction
- The highest level of abstraction happens when you have a conversation with yourself. There is lots you understand and do not need to specifically explain to yourself because you share your own education and experience. So you can effectively communicate in shorthand. There is little cost of codification. Any notes you write do not require elaboration because you know what they mean.
- The lowest level of abstraction is triggered when you want everyone to know what you know. The cost of codification becomes infinite becomes you have to provide to everyone the same education and experience. To achieve this, you must communicate your knowledge in the simplest, most concrete and comprehensible way.
- In any information flow, you must first determine the upper and lower levels of acceptable abstraction.
- The higher the level of abstraction, the richer the conversation but the fewer the number of people who can participate.
- The lower the level of abstraction, the thinner the conversation, the greater the cost of codification and maintenance, but the more people who can participate.
Maps and Taxi Drivers
- The following section relates to work Snowden did with Max Boisot.
- He highly recommends Boisot’s book, Knowledge Assets
- Snowden and Boisot did some work together based on the work of Michael Polanyi. Snowden extends Polanyi’s observation: “We know more than we can say, and we say more than we can write down.”
- This contrasts two extremes of knowledge: tacit and explicit. (He doesn’t like these terms and prefers not to use them.)
- Boisot observed that highly abstract but highly codified knowledge will diffuse to large populations fairly quickly. Examples: a map versus a taxi driver.
- A map. It contains highly abstract symbols (e.g., symbol for type of church), which he has learned over time and is able to use to navigate easily.
- A London taxi driver’s “Knowledge.” They have to know all the possible routes by memory, including all major landmarks along each route. The qualifying exam is rigorous and has only a 40% pass rate. People who pass tend to be highly adaptive (and, apparently, highly ethical). Interestingly, their training also enlarges their hippocampus to enable them to hold the additional new spatial mapping. (It takes about 2 years for this enlargement to occur.) This is very low abstraction, very low codification, and very low diffusion.
- Both types of knowledge are valuable. However, in a competition between a map user and a taxi driver, the taxi driver will win every time. This is because using the OODA loop (Observe, Orient, Decide, Act) to plan the route is highly explicit and slow for the map user but intuitive and very quick for the taxi driver. And, if something goes wrong, the taxi driver can adapt to changes in the terrain more efficiently. (Maps fall out of date and they contain assumptions that may not be explicit. Example, the map may show a route but it likely won’t tell you if it is safe at night.)
- NOTE: Most KM databases are highly abstract and highly codified (like maps) and make assumptions about what other people know. If those assumptions change, then the database is less useful.
- So when you are thinking about the kind of knowledge you have and how it should be shared and used, first ask if you need a taxi driver or a map. Don’t automatically assume you need a database (i.e., a map).
- The taxi driver takes time to train but then becomes highly adaptive and resilient. The map user takes no time to train, but is not nearly as adaptive or resilient. Both are useful, but in a crisis you need taxi drivers. However, because you don’t have time to train them in the crisis, you must invest in training them before the crisis begins.
- Micro-Narrative or Narrative-based knowledge: humans historically have used stories to share knowledge. These stories are not highly planned and polished, they are more spontaneous natural. They are “wild anecdotes rather than tame stories.”
- These stories surface weak signals, they surface outliers (e.g., people who are thinking differently).
- These stories are a way of surfacing attitude: attitude to safety is a leading indicator while compliance is a lagging indicator
- Side note: don’t run a workshop to ask people what they know. Instead, assess how they know things. The best way to do this is by eliciting their stories. The stories that tell you what is really going on are stories of failures not success.
- The stories people value are the stories of failure. It is these stories that teach us the most.
- “The brain registers failure faster than success because the avoidance of failure is a more successful strategy than the imitation of success.”
What’s wrong with KM? (Part 2)
- We have stories, taxi drivers, and maps. And we need all of them in combination and in the right balance. However, most KM programs focus too much on maps (e.g., structured, explicit knowledge). If they do include narrative, it tends to be highly structured narrative, which is almost as bad as maps.
- One of the principle components of a modern KM system is the effective management of informal networks.
- Done right, informal networks sustain the formal systems
- When he was working at IBM in the Institute of Knowledge Management with Larry Prusak and others, the ratio of formal to informal networks was 1:60 — and that was counting only the people using specific technology.
- Informal networks are an efficient way of spontaneously determining the level of abstraction necessary for knowledge diffusion without central planning or control.
- Informal networks are composed of people who have chosen to participate.
- Over time, they built a community of trust. Because of this trust, they were willing to admit their failures to each other. This ramped up the collective learning of the informal network.
- NOTE: We share failures only with people we trust
- When IBM saw the value of the informal networks and tried to formalize them, most of the useful informal network activity moved into an external collaboration environment beyond IBM’s reach.
- Larry Prusak: If you have $1 to invest in KM, invest 1 cent in information and 99 cents in connecting people.
- Human connectivity creates trust.
- Dense connectivity between people enables knowledge to flow at the right level of abstraction for the context.
- Direct human interaction is a low energy cost solution for knowledge management.
Stimulate Social Networks
- One useful technique for increasing direct human interaction is to stimulate social networks
- Allow people to self-assemble into teams.
- When people are allowed to choose their teammates, they tend to have higher commitment to each other than when they are assigned to teams.
- Provide guidelines, a set of heuristics or enabling constraints, that improve team potential by ensuring that you work with people you haven’t worked with before (e.g., a new employee, people who do not report to the same manager, someone who has a degree in anthropology or philosophy, etc.)
- Give them a series of intractable problems to solve and offer an irresistible reward such as a three-month sabbatical
- Allow people to self-assemble into teams.
- If you ran this exercise every six months, then within 18 months you have a widespread network of people who are within two degrees of separation based on having worked together in a trusted environment.
- This is a much better investment than spending 18 months building a knowledge base or AI-based search system because you have a dense human network that can assimilate new information quickly and diffuse it rapidly at the right level of abstraction at low cost.
- They have extended this technique to address mental health concerns.
- They expect a mental health crisis in early 2021 in response to the Covid-19 pandemic, triggered by the realization that this situation will not be going away quickly. However, the official systems will not be able to cope with a mental health crisis of this magnitude.
- In response, they are trying to rapidly build peer-to-peer support networks. For example, they created a series of trios in Scotland composed of a student, their parent, and their teacher. These trios overlap and support each other.
- Next they created additional trios composed of teachers, social workers, and police.
- This is called “entanglement around points of coherence”:
- The coherent points are the formal roles that have access to the formal systems.
- Then you interconnect them in multiple three-way combinations that create a dense overlapping network that contains a narrative learning system that enables a peer-to-peer flow of micro-narratives and the ability to have conversations.
KM for Decision Support
- If you create this healthy ecosystem of overlapping networks then good things will happen even when you don’t control it directly.
- “I don’t know what I know but I know that I will know it when I need to know it.”
- This addresses the biggest organizational challenge of the “unknown knowns” (i.e., the thing the organization knows but the decision makers don’t know)
- Informal networks that are tightly connected can feed into the formal systems
- Distributed Decision Support
- There are two functions of knowledge management: improve decision making and support innovation.
KM for Innovation
- Use KM to create the conditions for Innovation
- Inattentional blindness = when people are asked to focus on one thing and do not see something else that is right in front of them (e.g., the gorilla).
- This is not something you can train against because we evolved to make decisions quickly based on partial information absorption that privileges our most experience. This is called conceptual blending.
- Conceptual blending: scan the 4-5% of the available information, which triggers a series of brain and body memories, and then blend those brain and body memories to respond to the situation quickly. (We evolved this way to avoid predators.)
- [Look! A Tiger! RUN!!!]
- “We do not see what we do not expect to see.”
- During conditions of extreme change, this is even more dangerous because you are looking in vain for a world that looks like the world of 2019.
- Micro-Narrative Approach is one way of addressing both inattentional blindness and conceptual blending
- EXAMPLE: don’t send out an employee satisfaction survey. In surveys and interviews, people tend to provide the answers the think you want.
- Instead, present (or ask them to bring) a picture of what it is like to work around here. Then give them a series of triangles on which the can index their own narrative about that picture.
- For example, one of the triangles will say that in the story, the manager’s behavior was altruistic, assertive or analytical. These are three positive qualities so the respondent will have to balance the three.
- This pushes the respondent out of autonomic response and into novelty receptive processing (i.e., out of fast thinking into slow thinking), which makes them go deeper.
- Note: Most consultancy methods are context-free but the world of their clients is context-specific.
- “We live in the tails of a Pareto distribution not the center of a normal distribution.”
- Mass Sense — when an executive needs to make a decision quickly but doesn’t have the necessary information, doesn’t have time to research the issue, and doesn’t know what to do, how to proceed? Present the situation (via an infographic, a video, text, or some combination) and then ask everyone to interpret it using the same triangles. This is commonly known as “wisdom of crowds.”
- The resulting data can be plotted on a probability map, a “fitness landscape” (Stu Kaufman) that shows the various patterns in the responses. This will show you the range of thinking within a network. You can see where the consensus is and who the outliers are.
- This is real-time knowledge management for decision support
- This approach can be used in peace and reconciliation work. Start by presenting a set of data to people who are in conflict with each other and ask them to interpret it. Then go down one level to see where there are points in common (where you can bring them together) and points in conflict (where the differences really exist).
- This is “knowledge management hitting the road.” It’s not about building processes. It’s creating a dynamic, network-based, highly visualized response.
- This approach provides the “wisdom of the network” and, crucially, it helps your workforce participate and feel involved in decision support. This is critical for good mental health during a crisis.
- It enables weak signal detection
- It also enables exaptation
- Exaptation is critical for innovation. Exaptation is a concept from evolutionary biology: when something is originally adapted for one function but under conditions of stress exapts to another function. This produces an innovation.
- The history of human innovation is “radical re-purposing” or exaptation.
- In a crisis, the single-most important thing you should do is take what you do well and apply it to a novel situation. It is a form of improv.
- It may not occur naturally so use mass sense making to associate problems with existing knowledge capability at a level of abstraction.
- Art and music come before language in human evolution. They are also ways by which we become highly resilient as a species. Why? Art and music are abstract, they distance you from reality and allow you to make novel connections. Similarly, the fitness landscape maps allow you to see new connections.
- This is another example of real-time or organic knowledge management.
- Don’t try to organize knowledge in anticipation of need.
- Instead, create the mechanisms by which the knowledge can assemble in context at the moment of need.
- Aporetic Technique introduces paradox
- An Aporia is an unresolvable problem. In a crisis, you should create more of these because they force people to think differently. This is a major part of their forthcoming EU handbook.
- The handbook includes the 5 steps to get out of a crisis
- What parts of the problem do you hand back to experts
- What if you have conflicting experts? Use ritual conflict techniques.
- What if you have multiple hypotheses? Set up parallel testing.
- How to know if you have covered the necessary hypotheses? Use the mass sense techniques.
- The key thing in a crisis is to have a set of simple processes that enforce diversity.
- Knowledge management becomes even more relevant in a crisis:
- We need narratives, taxi drivers, and maps.
- “We also need the ability to rapidly connect people and things in novel contexts so that we can create new knowledge on the fly.”
- “Knowledge is a dynamic act of knowing not a static act of storage.”
Bonus: Responses in Q&A
- There is a new approach to Strategy: Apex Predator Theory
- When radical disruption occurs, the old dominant predators rarely survive because they were optimized for the old environment rather than the new one.
- What matters is having the low energy cost of fast adoption.
- Example: IBM is replaced by Microsoft, which is replaced by Apple.
- This is because they failed to recognize early enough the weak signals of approaching radical change
- When their environment changes rapidly, apex predators have two big challenges/opportunities:
- the exaptive moment: effective exaptation on the fly (i.e., quickly repurpose what you do)
- competence-induced failure point: where they fail, not because they are incompetent but because they are too competent as per Clayton Christensen. They have a very narrow window for change at this point.
- How to increase serendipitous discovery of novelty?
- Say: if I knew the answer to the problem, I would interpret it like this.
- Then, ask: Who else is interpreting it this way?
- This widens your lens and increases the chance of serendipitous discovery — particularly across domains and disciplines.
- The challenge for KM: “switch your focus from taxonomy to typology.”
- KM doesn’t get this. They think in terms of taxonomies, which gives you boundary conditions. By contrast, typologies give you multiple perspectives.
- This new focus enables trans-disciplinary work (which is different from interdisciplinary work).
- In a highly uncertain world, trans-disciplinary work means survival.
- KM has gone too far down the technology route. We would do better by increasing human connectivity.
- Narrative-enhanced doctrine:
- This work he did at Westpoint and elsewhere.
- They enriched documents with hot links to stories from a variety of people about what that document meant. These documents/stories were socially generated over time.
- Then you can search using some or all of the underlying stories to gain different perspectives.
- “Narrative enhances documents; documents enhance narrative.”
- The only thing that worked in Iraq was field commanders blogging.
- People wanted immediate real-time experience not manicured databases.
- We are past the fad cycle of AI. We are now working with the computer-human interface. KM should be part that conversation but it isn’t right now.
- Technology helps us scale knowledge. However, we need to rethink the way we use technology otherwise we will reinforce inequalities in the current system. (This is a matter of epistemic injustice.)
- Snowden: I don’t want to be a Jeremiah, but I don’t believe is the worst pandemic I will see in my lifetime and I’m 66. Covid is God’s gift to humanity, an opportunity for us to get our act sorted out and get ready for the big one.
- Without technology, we couldn’t scale. But right now, technology is an unbuffered feedback loop. Basic complexity science tells us that an unbuffered feedback loop will always be perverted. We need to introduce human buffering into that feedback loop. That is our challenge.