The Death of Reason: Project on the Rocks: Ain’t No Big Surprise

[Blog bot]

Источник: http://thedeathofreason.wordpress.co...-big-surprise/
==============

In the article “Blowup” contained in  Malcom Gladwell’s book "What the Dog Saw and Other Adventures”, he  discusses the failure of complex systems. His point is that we engage in a ritual of disaster (around nuclear power plants, air traffic control systems, shuttle launches) where we study all the myriad and minute pieces of evidence we compile from the disaster, we develop theories on why it occurred, then we make new rules/procedures/laws to prevent the same disaster from occurring.  The underlying principle is that we can learn from our mistakes and prevent future disasters – put another way, we identify an anomalous event or situation that we believe caused the disaster, we make certain we recognize it in the future and we go away knowing we will be permanently safe.

But what if we aren’t?  What if the systems are just so complex that these trigger mechanisms aren’t anomalous but are inherent to the systems. What if any complex system will fail, time and time again, and you just can’t prevent it? 

Keep that thought in mind.  Now think about the largest ERP, CRM or infrastructure project you are running.  Are these less complex?  You have 10’s and in some cases 100’s of people, working on individual tasks, all of which are interconnected in inputs, outputs and completion dates.  Not all of them report to you.  Not all of them are doing this full time.  Not all of them share your same motivation or primary interest.  Not all of them really care.  And not a one of them will take the blame if the project goes bad – that particular little nugget of concern is held by only three people – the sponsor (who will blame the next two guys), the P&L owner from the consulting operation (me) and the project manager (many of you).

It’s really more surprising when these projects succeed then when they fail.

Do we give up?  No, of course not.  I am paid the medium sized buck to manage exactly this kind of risk.  But I am beginning to rethink how I manage it.  Specifically, I am trying to be less hierarchal and less rigid in the command and control process under the theory that decentralized process and decision making is less brittle and, therefore, more likely to deal with the inevitable failures more quickly and more flexibly.  By doing so, the problems are dealt with while still small as opposed to letting a string of connected issues develop into a huge mess.

Here are my thoughts:

1.  Reduce command and control:  Rather than having a highly structured organization with a single CandC PM, I’m thinking about much smaller working teams with defined deadlines and little task management.  Let them figure it out provided its documented and delivered, on budget and on time.

2.  Know that you will fail:  Don’t plan on perfection, plan on failure.  You will miss deadlines, and you will miss requirements.  Admit to it, acknowledge it early, make certain everyone knows about it, then adjust to it rapidly.  Put another way, don’t fight the failure – work within it.  Or, stop the Tae Kwan Do and move to Aikido.

3.  Don’t get all hyped over SureStep, PMI, SDLC, etc:  Use very simple pieces of it, make certain it works for the business, but don’t get religious over it.  It’s a roadmap of the interstate highway, not a treasure map to the fountain of youth.  Stop working the methodology and work the business problem.

4.  Hire thinkers, not skill sets: I’ve been promoting college hires for years now and have experienced tremendous success in making such hires.  That’s because we hire thinkers and teach them skills rather than hiring skills and try to get them to think.  Skills need command and control, written policies, methodologies.  Thinkers take care of customers.

Again, these are my thoughts and, in the spirit of TheDofR, I’d love to hear you thoughts as well.


Filed under: Managing Clients Tagged: Malcom Gladwell; Bad Projects; Failure

Источник: http://thedeathofreason.wordpress.co...-big-surprise/