There is an article in Harvard Business Review this month about how companies are beginning to organize knowledge workers in a new way.  The concept has been called a “flow-to-the-work organization” and it reflects a new way of thinking about how knowledge workers are held in relation to the company.

The article by Roger L. Martin, dean of the Rotman School of Management, is called “Rethinking the Decision Factory.”  Since the main effect of these knowledge workers is to make decisions, he calls the collection of knowledge workers a “decision factory” and contrasts them with regular factories (which he calls physical factories).  The article has some problems, so let me jump immediately to the conclusion, and come back to a discussion of the rationale at the end.

Martin proposes that companies should keep pools of decision makers around who are not allocated to a particular division or business unit, and are therefor free to be redistributed around the company to where the demand is highest.  Not all knowledge workers, but a sizable fraction.  This is the approach that Proctor & Gamble used after the acquisition of Gilette and there is  evidence that the ability to rapidly redeploy these knowledge workers for this sudden demand allows the merger to proceed far quicker and to save a huge sum.  Filippo Passerini organized these loose knowledge workers into a global services unit, and used the term “flow-to-the-work” to describe the project oriented nature of how they are allocated to work.

The positive point is that this allows the organization to be more agile, more able to turn quickly.  Agility is a key factor for success in the future.  These global services groups operate like consultants, coming in to work on a specific project.  When that project is complete, they go back to the bench or off to other projects.  He compares this to the movie industry which brings together highly specialized teams for specific projects, and then disbands them.

It is an interesting idea, but I feel that Martin’s supporting arguments are not entirely satisfying:

• He suggests that one of the key flaws of current organizations is that knowledge workers are allocated to fixed jobs, and that knowledge work by nature is very spurious, and this results in periods of time where knowledge workers in those fixed jobs sit idle by making up work to do.  If this is the case, why wouldn’t the knowledge workers continue to do this make-work in order to preserve their position on the project?
• He says another problem is that management has no idea which knowledge workers are really needed and which are not.  I can’t see how restructuring knowledge workers into a pool will will resolve this.  Management is likely to be just a mistaken about whether a particular knowledge worker is needed in a particular position regardless of whether that worker came from a pool or not.
• He describes these decision makers as simply consuming data and producing decisions, regardless of whether it is their data or someone else’s data. This completely ignores the element of tacit expertise.  Many knowledge workers provide value only after a significant investment in understanding a particular market or audience.  This over simplified portrayal of a knowledge worker as a interchangeable automaton seems far too convenient and ignores the real nature of (much) knowledge work.
• Given that external consulting companies exist, it seems far more likely that the company would forego the pool altogether.  If knowledge workers are as interchangeable as Martin suggests, then why not bring them in from Accenture or McKinsey when needed.  That way, when the project is over, there is no need to keep them on the payroll, and you can pull from a potentially larger pool.
• While bringing together a tiger team to solve specific big problems makes sense, it is not clear to me why those chosen to join the team could not be pulled from those fixed positions.  Is this formation of the separate pool really just a way for some parts of management to get around the lack of support from the various business units?  When a person is allocated to a tiger team, does it really matter whether they came from a pool, or from a fixed position?  Is this anything more than just an attempt to centralize control of knowledge workers?
• If a person is allocated from the pool to business unit X, wouldn’t that person have all the same incentives to assure that they continue to need him on unit X, just as if he was permanently there?  Wouldn’t unit X work to justify their need of such a strategic skill?  Once working in unit X, does it really matter whether he official works for the pool or not?
• The entire argument is for knowledge workers, but isn’t the same argument even more valid for manual laborers?  Shouldn’t machinists be reallocated to projects on demand? Or warehouse managers?  Is this about a special quality of knowledge work, or really just an expression of the desire to structure the company in a project oriented way for maximum agility?

There are many questions about the rationale given by Martin, yet I have no evidence to contradict his conclusions either.  It would seem to me that having decision generalist (like “judges”) who can be dropped in to resolve troublesome situations does sound like a good idea, but those people could not be expected to gain the domain specific expertise of the kind of knowledge worker that you need to succeed in an ultra-competitive world.  From his cookie-cutter description of a decision maker, it is not clear to me that Martin really understands the value of tacit expertise.  In my mind it is inconclusive as to whether organizations will rush to organize knowledge workers in a centralized flow-to-the-work organization, or not.