A basic assumption so central to running an organization that we never question it.  What if it is impossible to collect process knowledge?  The thought follows from Steve Denning’s excellent article “Can Knowledge Be Collected? Lessons From The Health Sector” in Forbes this month.  His examples from the health sector involve determining what to do, and how to get it done.  Let me take a stab at translating these ides to the more general case of knowledge workers and process knowledge/management.

• No correlation of spending on knowledge with results – Building a huge library of business processes may cost more than it is worth.  “Booz & Company’s annual innovation reports repeatedly state that spending on knowledge has no relation with results.” There is no reason to believe that process knowledge is special in this regard.
• Knowledge collections per se are useless – The collection of processes definitions is only valuable if they are used.  His story about a company that motivated collecting knowledge could just as well have been about process knowledge.
• Knowledge collections can slow you down – “In this constrained setting, access to knowledge just slowed the team down. Without a desire to learn, improved access to knowledge was a liability, not an asset.”  I would not use this to mean that collections of processes should necessarily be avoided, but as we see below there is a judgement call about which processes to collect, and which to avoid.
• Knowledge collections can actually make the firm less effective – a large collection of well known process definitions may discourage people from inventing a new better process.  There would be a pressure to demonstrate that the existing processes are less optimal, before you could move forward on an innovative one.  Organizations without large formalized collections of processes would be freer to move forward on innovative approaches.
• Tacit vs explicit knowledge – This is the core of the issue: many professionals use tacit knowledge to guide their work.  Some experts “argue that knowledge is inherently tacit: knowledge consists of internalized understanding such as how he or she accomplishes particular tasks.”  It is not clear that professionals (knowledge workers) necessarily understand this.  Under this theory, the concept of “explicit knowledge” may be a contradiction in terms.  It may simply not be possible to write down the process to be used.
• Three types of knowledge– we can generalize his examples from medicine
  • Precision Process Knowledge – “In essence, symptom A is known to be the result of disease D to which we can apply treatment T with the reliable outcome O. ”  This is of course the kind of knowledge that process management has focused on.  If you really can with 100% accuracy recognize the situation, and 100% know the solution, then your should be able to without difficulty model and enforce this process.  Process collections of this nature, assuming that the right process can be found, should be valuable.  The flaw is thinking that all processes fall in this category.
  • Intuitive Process – “we may have a whole array of symptoms, A, B and C, which may be connected with a number of possible diseases D, E or F which may have possible treatments T, U or V which may lead in some cases to outcomes O, P and R.”  This is the classic unpredictable scenario.  Perhaps you don’t yet understand the symptoms well enough, but you can’t wait; you have to start some form of treatment.  Later you will get more accurate assessment of the symptom and understanding of what to do will emerge.  Or possibly the symptoms are clear, but the possible treatments are simply not known by anyone.  In this case, some experimentation with treatment is justified, and possibly a discovery will be made — but only if you are innovative.
  • Behavioral change Process – I found this category interesting.  You pretty much know what to do, but getting someone to do it is a challenge.  We are not considering the case manager themselves (let’s assume they are motivated and competent) but such a case manager has to work with partners and others to achieve goals.  Case managers can not do everything themselves.  Motivating or convincing others may be haphazard and this must be monitored and the approach adjusted as you go along.  This is another source of unpredictability in the process.
• From intuitive knowledge to precision knowledge, and back – the category may change over time!  You may have perfectly tuned processes today, that start failing.  A process melt-down.  We know that BPM is all about continual monitoring and improvement of a process, but what if the situation changes from a precision process to an intuitive process?  You can not simply adjust an existing diagram, but you may need to retreat to a non-defined process until new intuition about the situation is known.
• The role of judgment in applying process collections – It is clear that a process library is neither entirely good nor entirely evil.  The trick seems to be in accurately identifying which type of process you are in.  Can you identify intuitive or behavioral processes? And then can you avoid premature formalization of these processes types?

The article is interesting because it questions a bedrock assumption that we in IT typically hold dear:  that knowledge can be collected and collecting it is always good.  If this is not true, or possibly just not always true, then we might need to question the approach to managing knowledge.  These ideas translate to process knowledge as well.  Some careful consideration is needed before we blindly commit to a course of automating all processes within an organization.