Doing by learning

[Note: This is the fourth in a series of posts about the Social Enterprise and the Big Shift. The first post provided an introduction and overall context; the second looked specifically at collaboration, working together; the third looked at optimising performance, enjoying work, working more effectively. This one deals with flows, how work gets done.]

[My thanks to Mike Agner for the wonderful shot of Puerto Princesa, Palawan above]


In my last post, on enjoying work, I wrote:

In explaining the Big Shift, Hagel, Seely Brown and Davison spend time describing the changing environment caused by the introduction and evolution of digital infrastructure, augmented by public policy decisions on movement and migration. They describe a world where competition is more intense, where barriers to entry are lower, where the rate of change is high, where things are more interconnected and where there is greater uncertainty as a result.

Prior to that, when looking at collaboration, I’d written:

Experience curves were about the past; they containerised historical experience and explicitness and sought to extract value by repeating that experience in military fashion; and, in consequence, marginal utility diminished over time while marginal costs increased, and a classic diminishing-returns model ensued. Collaboration curves, on the other hand, are about the future; they seek to containerise tacit knowledge, the ability to learn,  to adapt, to evolve; value is created by making the company better at learning

These two statements, taken together, form the backdrop to the rest of this post.

Introduction: Doing by learning

There’s always been a close association between the concept of learning and that of doing. It is reasonable to suppose that there’s been a similar close association between the concept of doing and that of working, even though there is occasional evidence to the contrary.

Life used to be so simple. People learnt their trades, usually as apprentices to those who’d mastered some particular skill or skills,  and then went off to apply what they’d learnt, to ply their trade. As apprentices, they learnt in a number of ways: they received instruction; they observed; they imitated; they practised; they received feedback; they improved. They learnt. And by learning they achieved mastery in a skill or set of skills.

Instruction was meaningful for things that could be explained, that could be articulated clearly. Explicit things. Observation and imitation were more relevant when it came to things that were harder to put into words or even pictures, where the learnt skill was more deeply embedded. Tacit things. And as long as it was constructive, criticism was a valuable component of the learning process. If someone could observe you while you did something, they could notice things you would find harder to notice at the time. Some sort of Heisenberg Uncertainty Principle prevailed: your attempt at observing your actions while performing them tended to affect the action.

All this described the world that was, and not the world that is, much less the world that is to be. We used to live in a world where what you’d learnt could be stored, canned, repeated at will, “scaled”. That was the world that Hagel, Seely Brown and Davison described as based on “experience curves”, where past experience could be used to control markets. But they describe the post-Big-Shift world differently, as one based on collaboration curves, where value is created by making the company better at learning.

I love Peter Drucker, and have no qualms in quoting him repeatedly. And one of my favourite Drucker quotes is this: the purpose of business is to create a customer. In similar vein, it is reasonable to assert that work is about creating value. So, if value is created by learning, then learning is work. And work is learning.

How work takes place

If work is learning, then work takes place when a person learns. As inferred earlier, people learn in a multitude of ways:

  • by receiving instruction
  • by observing
  • by imitating
  • by doing, under supervision
  • by doing while being observed, so that feedback is available
  • by assimilating and responding to feedback

Esther Dyson, someone I regard as a mentor, tends to sign off her messages with the phrase “always make new mistakes”. For learning to have taken place, something must be new. Something must have changed. And, quite possibly, something must have been unlearnt, discarded.

Learning is about flows, not stocks. We live in times when change and speed are abundances while certainty and predictability are scarcities. And we need to adapt to those times. To be successful one needs to allow for the new abundances and the new scarcities. The firm that does-by-learning will prosper, but only for a short time: competitive intensity is high, barriers to entry are low. So in order to sustain that prosperity, firms will have to learn how to keep learning, and how to do that at speed. Continuous and quick learning.

These statements are all little more than soundbites — PowerPoint fodder, nothing more — unless we can really understand what they mean within the enterprise. So what do they mean?

My assertion is that to understand what they really mean, we have to understand in a more granular way how all this takes place. Which leads me nicely on to one of my pet subjects.

Flows. Information flows. Conversation flows. Which is what the next part of this post is all about.

An introduction to flows

Last week I wrote:

Flows are part of networks, not hierarchies. Places where network effects can be obtained, where increasing-returns models can be seen to apply. The core of the Social Enterprise is in the network, the connectivity, the connections. Connections that extend beyond the enterprise, into the supply chain, through the distribution networks, all the way to the customers and the products. Networks across which conversations flow, cutting across the silos of the organisation, straddling the boundaries, allowing the tacit knowledge at the edge to be exposed. Here are some of the characteristics of flows:

  • Flows are not transactions. They can and do include transactions, but they represent far more than that. Transactions are just one type of object that can be embedded within the flows.
  • Flows are conversational. Start and end points are imprecise, sometimes absent. There is no simple linearity to flows.
  • Flows transcend “channels”, a concept born of hierarchies and control. A conversation may start in one medium, stall, restart in a second and different medium. Bilateral conversations may morph into multilateral ones and vice versa.
  • A flow represents a continuum from past to present to future; they involve transactions (the past), activity streams (the present) and intention signals (the future). But these are all just objects embedded within the flow.
  • These embedded objects are valuable in themselves, but gain their prominence from network effects: the power of inspection; recommendations and votes; the application of cognitive surpluses; the opportunity to “save” and “replay” activities in detail, and to have “freeze-frames”.

Today I want to spend a little more time looking at enterprise flows in the context of the Social Enterprise. I’m going to share what I think they are, and hope that, with your help and comments, I can improve my understanding as well as yours.

Social Enterprise flows

The concept of flows has been around for a very long time. Scientific management and assembly lines and work flows and process flows and all that jazz. Industrial age stuff. And if what you embed those flows into the Social Enterprise, you probably deserve what you end up with. A footnote in history, somewhat earlier than you’d anticipated. Michael Hammer, when talking about re-engineering the corporation, stressed the importance of avoiding “paving the cowpaths”. The whole point of the Social Enterprise, as articulated within the Big Shift, is that the flows are different. In fact prior flows were staccato, fragmented, fossilised in comparison; they were daisy-chains of stocks, lacking fluidity, lacking adaptiveness, lacking evolutionary capability, lacking life.

Strong words? Perhaps. And if I’ve offended you, my apologies. My intention was not to shock, but to express the seriousness of this change. Historical flows were processes that dealt with the creation, passage and transfer of explicit information. They were therefore themselves easy to codify, to standardise, to repeat. And scale was obtained as a result of this codification of process.

As against this, Social Enterprise flows are about surfacing tacit knowledge. Information that is hard to codify, standardise or share. How is tacit knowledge surfaced? Through the sharing of experiences, as Michael Polanyi pointed out all those years ago. Shared experiences are not that easy to come by: they tend to require “synchronous” participation. Which limits the ability to scale.

To combat this, Social Enterprise flows are persisted: they’re archived, searchable, retrievable. The transactions, activities and intentions that make up the flows are recorded. For them to be valuable, they need to be replayed at will. And that’s easier said than done, since the persisted flows resemble firehoses. Which limits their usefulness.

As a result, Social Enterprise flows contain rich metadata: they’re auto-date-and-time-stamped, they’re geo-located, they contain information about the identities of the people involved; they’re enriched using folksonomies that avoid the traditional limiting constraints of topic classification trees. But surely all we’re doing is paving the cowpaths, playing at semantics, pretending that the stocks of yesterday are replaced by the flows of today, while really continuing to do the very same things? These flows represent a point in time, they’re recorded, archived, fossilised. So what’s changed?

The change happens because Social Enterprise flows are participative: people can copy, amend, re-use, correct, augment, enrich them. They’re flows. They move. They have life. As Doc Searls used to say for open source, the NEA principle holds: nobody owns them, everyone can use them, anyone can improve them. But even that is not enough, not unless people can learn about the changes, and learn quickly about the changes.

Which is why Social Enterprise flows are about publish-subscribe: the learning has to be shareable; and it must be beneficiary-led. Broadcast and hierarchy do not scale, the learning withers and dies.

What happens in the flows

People ask questions, and share answers. They share learning. They observe, and share observations, provide feedback. They list and rank and rate, and share their valuations. They inspect and correct and share the corrections. They represent different points of view, they challenge, they debate. And they share their reasoning.

People move information around. And share their perceptions and views and valuations and ratings about that information.

People learn, and continue to learn. They do this at speed, adapting to internal as well as external stimuli.


People talk to people. People relate to people. People learn from people, people teach people. People buy from people, people sell to people. In the end it’s all about people. In the past, we didn’t have the ability to connect everyone up affordably and efficiently; we couldn’t record and replay transactions, activities, intentions; we couldn’t review, rate or provide feedback; we couldn’t correct, repair, enhance, enrich.

We couldn’t scale our ability to share our tacit knowledge, and to keep sharing our tacit knowledge.

We couldn’t scale our ability to learn, and to keep learning.

A coda. When we converse with each other, we tend to embed our conversations with  social objects, the rolling stones that gather the moss of learning. My next post will look more closely at the role of social objects within the Social Enterprise.


9 thoughts on “Doing by learning”

  1. Thank you, JP, I think your post just put me into a state of flow…let me try to paraphrase your thinking around stocks and flows.

    The old infrastructure was built and optimized for stocks, not flows. It’s hierarchical, it’s driven by command and control and its purpose is to transport and deliver explicit information from sender to recipient. What are the best channels to do this? Well, it’s good old email, staff meetings and the phone.

    Contrast this with the new infrastructure which is being build for flows. It’s networked, it’s powered by individuals within a collective and it’s purpose is to transport and deliver social objects.

    And this is where I see a profound shift in our typical workday happening: moving away from a day that is dictated by emails, meetings and phone conversations (all remniscent of efficiently communicating stocks of knowledge) towards a day that is build around the creation of our own social objects (call them stories), the consumption of new social objects (call them feeds) and the weaving of new connections to increase the flow (our participation in and expansion of new networks).

    This would also explain why the shift in how we work is so difficult to accomplish – it’s not so much the behavioral changes that are necessary (the tools are getting way to easy to use), it’s because we are still in the mode of moving heavy stocks around, and those are best “transported” through our old infrastructure. (ps. here’s the chart I had created

  2. Hi JP, to augment your comments about the supply chain… They are evolving from ‘Production push’ to ‘Demand pull’, low levels of distributed inventory with shared ownership and conversation and context, instead of ERP process enforcement… “The times, they are a changin”

  3. Awesome. Another great post! I especially like your emphasis on tacit knowledge and the conditions for enhancing “flow” around tacit knowledge which, in the words of my colleague, John Seely Brown, is actually quite sticky.

    More broadly, you are highlighting a huge untapped opportunity for companies. What if companies took talent development seriously and understood that the most powerful and effective form of talent development and learning does not occur in training rooms but rather day-to-day on the job. Given that, how would we redesign our work environments to amplify flows to accelerate learning? I started to explore this theme in a blog post here

    Can’t wait to read your next posts in this series! I feel that a book may be in the offing here . . .

  4. Hi JP

    I described the transitions in learning from moving from the model of 
    Accrual of Experience (apprenticeship, medieval age) – learning through observation and correction under the tutelage of an acknowledged master in the art. Everything was art, even science was an art: maestros defined the current state of knowledge based on their own discoveries and analysis. With population grown and education however, it turned out to be too impractical for a master to teach too many people. So they started to formalize and codify the explicit reproducible knowledge, which led to..
    Accrual of Knowledge (formalized education) – the birth of the university and formal advanced education so students have a strong basis in “what is known”. Thins remained in place for hundreds of years as the rate of growth of how much there is to know did not expand beyond the capabilities of the state of what is needed or applicable in business or production environments. However for all the reasons that Hagel, JSB and Davison described, the rate of change of what is relevant is outpacing what we know now, leading to the new challenge…
    Accrual of Awareness – the ability to process information in the flow. It is taking what we learn as a basis and then add new skills to, as Toffler said, “to learn, unlearn and relearn”. These are skills for us to understand how to listen to raw information, understand how to categorize and perform cursory analysis and perhaps create the metadata to organize it as needed. 
    The real challenge is that these skills of managing flow of knowledge are not taught in any practical sense. Unfortunately, the emphasis is still heavily on using proven knowledge solely and devaluing tacit, context specific knowledge. We have seen that tacit, even experimental knowledge can outweigh the base of proven knowledge. For example, Univ of Washington FoldIt project that solved through crowdsourcing how to find the right protein that led to a breakthrough in HIV drug research in a few weeks where previously ranks of medical researchers had been looking for almost 15years. ( Incidentally, I am reminded of that example while Sitting and listening to a workshop session by Dion Hinchcliffe and Peter Kim at Enterprise2.0 )

    My point is that we need examples of the atructures and methods of what goes into managing flows. I do agree with John, Joi Ito and you that one place to look for them is in how squads or groups handle complex flow in multiplayer games. I was hoping to hear more about that this past SXSW but didn’t really hear that. I am not sure this has become a science of its own but I do think it should be a priority. That’s the next book we really need.

Let me know what you think