« management is rubbish | Main | on (completely new) software »



I think you should adopt an evolutionary view on this... assume people *start* with principles, trust, and good intentions and end up with confining and specific rules. And that this outcome is selected for in an evolutionary process.

How might that work?

Well... about transparency: If an act is transparent in a forest, but no one is there to see it... did the act really happen? Where does the need to monitor arise? It arises with the separation of ownership and management (principle/agent).

Why contract an agent? To do more than you can do on your own. To leverage your capital. To employ their knowledge.

Over time (or maybe from the start), knowledge accumulates and is concentrated in management as opposed to ownership.

Knowledge accumulation anc concentration leads to power.

Power combined with self-interest leads to a principal/agent problem (agents acting in their own interest instead of the principal's).

Agents (with knowledge) can "hold hostage" the principals.

As a condition of keeping the capital employed, therefore, principals require monitoring (they know this game and therefore insist on it up front).

And you end up with a perverse outcome in which the monitoring structures are the most valued part of the system for the principals. It is a lemon principle in action (Akerloff's lemon principle is that all used cars are discounted because of the existence of "lemons" in the market with no cheap way to determine whether or not you're being hosed. A rational agent therefore discounts all offerings in the market): Without a clearly "accountable system" in place, the firm's cost of capital will go up. Securities reg's are a feeble attempt to counteract this ("see we've got the toughest rules going, you can trust these firms").

There aren't many ways to avoid this outcome... but here are a few ideas...

1) Avoid principal/agent conflicts by taking specific measures to align them. ESOP, profit share, bonus your butt off. Do it innovatively to avoid short term/long term tradeoffs.

2) Avoid principal/agent problems by staying small. You can stay small many ways. I'm most interested in a "cellular" model whereby work units are kept small, and have the capacity to be redeployed as necessary, and are incented to work together. And, like WL Gore, force large units to restructure into smaller units as soon as I get large. This, I think, simplifies the knowledge management and accumulation problem by "storing" relatively unified chunks of knowledge separate from others. We avoid huge costs of complexity, but incur unknown opportunity costs of lost collaboration.

3) Build knowledge management into the organization from the ground up. You've somehow got to shortcircuit gains from knowledge-hoarding. Transparency would be achieved in this system.

As you might guess, I find this stuff fascinating. The challenge is see is this: to build an accountable system. Accountable systems have common characteristics:
- they are able to make a commitment (many organizations, especially large ones, avoid making commitments... not necessarily on purpose, they just can't).
- they do what they can to achieve that commitment
- they measure their achievement (or not) with brutal honesty
- based on the results of their measurement, they confront reality and change what's necessary
- they report to their "principals" regularly and honestly, in so doing, they re-establish their commitments
- they are "ready to report" at the drop of a hat. It is known throughout the organization "where the buck stops" for any given initiative or activity (there is "point accountability"). I see this as "transparency on demand" as opposed to "always on" transparency.

FWIW... :)


Rockster, excellent! I agree in most aspects - but let me hook onto a few 'concepts/terms':

You mention "units" - as you may glean form some of my other posts I would suggest abandoning all notion of a "unit", even boundaries of corporations. I would suggest basing the Business Model (how to use the resources etc.) on a holistic view of the flow started by a customer.

That would, maybe, make aligning (as in your 1. above) and breaking down of information-based power structures easier.

"Knowledge managment" in its ultimate form is complete transparency - when everybody knows everything...or what?

Ahh, and yes, "monitoring" and "accountability"! For my ear (and it may be only me) that still is hierarchy and opaque structures linked terms. If truly transparent they would be moot - maybe only for one simple human reason:

"If I know that you can at any time see and hear whatever I'm doing I will at all times do my best". Like standing on a podium. And you do not even have to look over my shoulder, the mere knowledge should be enough.
If not I should be fired :-)

Add that my peers and subordinates can see all, then the pressure increases - and the opportunity to 'show off' as well! Having peers approaching me in the cafeteria telling me that "they have seen what great stuff I was doing this morning" is after all a billion times stronger than a bimonthly talk with my boss I think :-)

Perhaps the clue is a single, simple thing - open up everything, complete transparency, end of story. That will p*** off quite a few, but hey, suspect those are the ones you'd like to get rid of any way.
With complete transparency hierarchies will crumble (ouch need a different set of practical process guiding mechanisms then!), with time the customer could be included in the whole shebang and principles based guidance could replace detailed and stifling rules-sets. Issue solved :-D

Thanks again for the input, as you can see, I too think these things are damned fun!


Riverbeds also change direction more easily, and without requiring dismantling or parallel, redundant constructions (like pipelines). Not too easily - but over time, rivers find their way around and over changes in the topography.
It's not a bad analogy, Sig (but no babbling brooks, please!). I agree with Rockster about staying or becoming small - it aids transparency no end when you all work in the same room. I don't think that collaboration is necessarily impeded by having separate, smaller 'units' (I'll use that term because I'm thinking of groupings of function and knowledge that don't necessarily have any organisational ties at all) - a flow of information and activity can connect nodes in a value network that would otherwise be disconnected (and may 'disconnect' again when the function is complete). This is I guess the 'virtual corporation' idea, or more pithily, Jon Husband's ( http://blog.wirearchy.com/ ) "conherderation" of collaborators working on a temporarily-common goal, subsequently breaking up and re-forming in infinitely diverse ways.


I guess my main point is this: as soon as you separate ownership of assets from management of those assets, you end up with a hierarchy - principal/agent.

Your conceptions require that the value-creating activity can be done by one person in a serial process. As soon as the activity is done by a team, you've got a team of agents and a problem trying to determine who contributes what value.

So... I say treat the outcome (hierarchy/rules) as the result of an evolutionary process (evolution is the key, it removes any notion of "intelligent design") and assume each generation was a better solution to the current environment. The tweak, then, has to be in that evolutionary process and probably the environment's demands.


Guys, as you say "evolution is the key", and "over time rivers find their way.." - I can only say hear, hear!

For that we need two things: A force and the means.

I think the force is with us (in these Star Wars times..), there is a reason why Dilbert has such a following, why Cluetrain and Hughtrain and the many blogs says much of the same. Even managers pipes in. I believe the force for change is much stronger than the force to conserve as it were.

The means on the other hand, ouch. Blogs for me chips away on the Hierarchies (Scoble comes to mind), but still most of the enterprise software and physical space (office/plant design) blocks the natural evolution.

In essence, I think the most efficient way forward is to focus on the physical infrastructure, less on attitudes and theories. The new BMW plant comes to mind, more internal and external blogs, a new way to build enterprise software - that might be the trick.

(Of course, keep in mind, as you know, I've got practical interests in the "evolutionary enterprise software" direction :-)


Yep... build into your thinking though that evolution selects "for" attributes... so what we've got now is "best fit" for some twisted reason. Is it best fit because there was no tech to enable better fits? Maybe...

Hey... I just passed you the musical baton: http://www.rockster.ca/?p=27


Rockster, I'm slooow here, could you expand on the "for" in your comment?

And I'm really glad you said passing the baton could take some time :D


Evolutionary processes look for attributes that are better fit than others. Generally referred to as "selecting for" that attribute. Routines in organizations that are less fit are not selected.

If you believe that evolutionary processes are indeed at work in organizations, then your thinking has to wrestle with the fact that all of the things you rail against are there because they were somehow the best solution (they were "selected for" their attributes).

Scary, ain't it. See my latest post on views of risk...


Ahh, yes, think I got it. :-)

And you're of course quite right!

In this http://thingamy.typepad.com/sigs_blog/2005/01/marketing_manag.html
I suggest that the current "reality" (that I'm railing against) has evolved by natural order, and has been the best and perhaps only solution in order to make large organisations work.

Nevertheless I say the solution is wrong today - because there are better solutions to solve the "information" issue, namely IT (if only it weren't modelled on the old makeshift solutions).

In other words, the solution (hierarchies, marketing, management as we know it - all information issue solutions) would perhaps been different, maybe not existing even, if information technologies had arrived in parallel with production technologies.

I do agree with you as long as no alternative method or technology is offered to handle the information issues, we'll get nowhere, the current solution would stay put as the best.

E-mail did a little dent in the rigid information structure. Usenet added another tiny one, and now we do see blogs adding a slightly better dent as the outside world finds little direct doors into the organisations.

In a way I find the growth of corporate blogging to be a indication of the underlying forces.

And luckily, it's untouched by managment :-D

(Had a similar experience once, being the clueless manager: http://thingamy.typepad.com/sigs_blog/2005/03/early_days_netw.html )


In the environment you're hoping to thingamize, there are some things I'm skeptical can change... (they can be kept out of a startup, but it is frickin' hard work):

- fear-driven decisions (job loss, shame, etc)
- the human instinct to distrust difference
- generalized lack of patience
- the human tendency to take the easy road
- the human desire to gain power

And on that happy note... I think those are most of the drivers of negative organizational evolution. Can they be overcome?


Rockster, I agree, those will hold back most.

I still think (know rather from mails and other comm.) there are some leaders that will want to change, that are fed up enough.
At the very moment they have a practical alternative to make the information flow among their 'resources', then they will try. (Not pushing our software only here, it can be many possible alternatives... or should be)

And if only 'one' is succesful as in better bottom line, some quite useful human traits may kick in:

- fear (from being overtaken)
- greed (seeing others being succesful)
- sheepishness (going baahh when others moves on)

So, one success should be enough or what?

The comments to this entry are closed.

My Photo


  • Phone: +33 6 8887 9944
    Skype: sigurd.rinde
    iChat/AIM: sigrind52

Tweet this

Thingamy sites

  • Main site
  • Concept site

Tittin's blog


Enterprise Irregulars


Twitter Updates

    follow me on Twitter


    • Alltop, all the cool kids (and me)


    Blog powered by Typepad
    Member since 01/2005