Thursday, November 29, 2007

[incompetence] failure analysis and human nature

This mini-series began with "There may be a reason", continued with the article on "Human Resources" and concludes now with "Failure Analysis".

The series is a reaction against the massive losses of the past few months in governmental departments and it focuses on the private sector first.

The greatest problem with failure analysis is that no one wishes to acknowledge a personal error. As Amy Edmondson and Mark D. Cannon wrote, in The Hard Work of Failure Analysis, August 22, 2005:
Most people prefer to put past mistakes behind them ... self-confidence and self-esteem [are at stake].
If we examine the mechanism in our heads, it's pretty clear – we make the error, we hope we get away with it, when we are found out we rationalize it or blame something or someone else.

So what hope does failure analysis have?

Somehow the Child Endowment details of millions were lost. Clearly the source of the problem is somewhere and has to be identified in order to prevent it happening again. For the Minister to accept responsibility would be a fine thing although everyone knows it wasn't directly him to blame.

The senior official who did take the heat – well not even he was directly to blame for the loss of the CDs. The blame lay somewhere within the system [excluding, for the moment that it was induced from above, which is not, as has been shown in the Micro-Control posts, beyond the realms of the possible].

How to find the real source of the error, with everyone scurrying around watching their butts? Who's going to be honest?

Edmonson and Cannon continue:
Not so long ago, a large European telecom company [attributed] large failures to uncontrollable events outside the organization e.g. the economy and to the intervention of outsiders.

Small failures were interpreted as natural outcomes of experimentation, or of not adhering strictly to the company's core beliefs. Similarly, consultants simply blamed the client, concluding that the client was not really committed to change, or that he was defensive or difficult.
Frequently, lower level people are reluctant or unable to identify the bigger problem. This may be because:
1. they feel it’s not their business;

2. they are less able to think conceptually about issues;

3. they believe that, even if they raise these issues, they won’t be acted on;

4. they are frightened of recrimination, especially from immediately above them;

5. they don’t want to add to their workload.
There's an awful lot of froth and bubble surrounding failure analysis when it comes to humans taking the blame. Alexander Dunn, in Root Cause Analysis, Assetivity Pty.Ltd., 2004, quoted Latino and Latino [see below] on causes of failure:
1. Physical – why the machine broke;

2. Human – who failed to do what he should have;

3. Latent – organizational problems which led to this situation.
They argue that the most effective, sustainable solutions are those that address the latent causes. Yet we often see troubleshooters focus almost exclusively on addressing physical causes, while management focuses on human causes.
And what of the people themselves who investigate the errors?
[Some say] that failure analysis is best performed by “experts” [but] these specialists quickly become the bottleneck themselves. Every individual brings his/her own knowledge and biases to the problem solving process.
Some possible ways around the impasse:

Amy Edmondson and Mark D. Cannon mention Julie Morath, Chief Operating Officer at the Minneapolis Children's Hospital, who has a "Good Catch Log" to record information that will be useful in better understanding and reducing medical errors. Any error, no matter how small, is recorded and the sum total reviewed by panels and patterns identified.

The pharmaceutical industry is a special case - about 90 percent of newly developed drugs fail in the experimental stage, and drug companies have plenty of opportunities to analyze failure. Perhaps that's where we should look for solutions.

Certainly it's worth the effort. Firms that are creative in analyzing failure benefit in four ways:
1. analyzing a failed drug sometimes reveals that the drug may have a viable alternate use – e.g. Pfizer's Viagra, Eli Lilly, Evista;

2. a deep probing analysis can sometimes save an apparently failed drug for its original purpose – e.g. Alimta;

3. failure analysis can often include customer feedback – e.g. Xerox;

4. the value of the learning itself [in the analysis] is often overlooked. For example, researchers in one of the early German polymer labs made a mistake which led to the development of nylon.
Certainly that's a method but who would carry it out? Alexander Dunn advised:
Empower the workforce to solve problems within their area of operations and encourage the use of team based problem-solving approaches for more complex problems.
Maybe but as Anon wrote:
I have encountered many people, mostly in their 50's, professionals, possessed of detailed skill sets and experience, whose position in the hierarchy, and therefore decision making abilities, has been displaced by a "team" of dysfunctional incompetents, effectively "above them", making rules and procedures that are unworkable. This team proceeds without a knowledge base, and of course major cluster f**s happen frequently.
And don't forget:
A camel is a horse designed by a committee. [Issigonis]
Surely it comes down to the composition of the teams in the first place and whether their briefs are within their range of their competence. Management always likes employees to be multi-faceted but even such people have their core competencies and other areas where they're a little fuzzier.

Then there are the triple factors of people's aspirations, their own self-assessment and their capacity for good interpersonal relations. All very well for management to ignore this and say staff must get along with each other. All very well for group leaders not to wish to hurt the feelings of members.

It needs a trained or a natural team leader, not just an elected team member for the week and each team leader needs to have a dash of competence, a dash of people skills and a sense of where the team fits into the whole. In the end, a quietly spoken, concerned team leader will win out where a brash, focused expert might not. Or a team member with core competence in that area but no people skills will not.

So, if we empower the team, based on three basic concepts:
1. direction [knowing what has to be done and how];

2. freedom [to do the job without interference];

3. support [being provided with the necessary resources]
... and give them a quality team leader, then a certain esprit de corps can develop and everyone knows the group will take the heat for any error but the individual most culpable is still not going to be able to hide. The matter can be dealt with quickly and with little fanfare.

If management is simply looking for scalps whenever there's an error, the team approach can't work. If management is simply looking for the errors to cease, then the team leader is now empowered to approach what he/she knows to be the trouble and make adjustments to the team.

This provides sufficient elasticity to save the hide of the miscreant on this occasion [who may indeed be skilful but just made an error this day] or alternatively, the team, in discussion, can come to see the systemic error – the flaw in the lines of communication, for example - and make recommendations, free from fear.

Now we come to the biggy, the granddaddy of them all – genuine support, genuine listening and the "taking on board" from top management. Dunn again:
We are conditioned from childhood to wait for direction. Effective empowerment requires a high degree of management support behaviour, particularly in organisations with previously strong hierarchical management structures and “Top-Down” decision-making styles.
It is essential that recommendations are acted upon. Esso Longford, Columbia and Challenger [and some say 911] are all examples of failure to act on advice.
It's not only management's willingness to genuinely listen, it's also the ongoing culture of analysis and the “Good Catch” procedures in place right throughout, from upstream to downstream.

The greatest problem within the team is if the worker with the highest level of experience and expertise lacks people skills and if another is then appointed over him. It's therefore all in the appointments and in Top Management keeping small leadership "pairs" or "trios" abreast of company direction, drawing them into the planning process.

It's also laying on these leader "clusters" the constant need to address even the smallest error.

High Reliability Organisations, show a preoccupation with failure. They constantly encourage the reporting of errors and treat any failure, no matter how small, as a symptom that something is wrong.

This preoccupation with error creates a culture where people gradually become less frightened of participating and eventually give their considered opinions, with no fear of recrimination.
This is the key to the whole business – genuine elimination of fear – let's face it, we are a very untrusting lot when it comes to possible blame being laid. Reason and Hobbs argue that successfully eliminating fear requires the proactive establishment of an organisational culture with three components:
1. a Just Culture in which there is agreement about what are blame-free and what are culpable acts;

2. a Reporting Culture where any reported error invokes a team response, and where the reporting worker knows he will be rewarded for the report and guaranteed that it does not go on his employment record;

3. a Learning Culture to prevent future errors and failures.
Of course, on the “snitching” in N2, it's the error which must be snitched on, not the individual, otherwise it will become a Stalinesque, oppressive work culture and all the good will be undone. There has to be a recognition also that most errors are not solvable by Root Cause analysis becasue they are an amalgamation of many factors.

W. Hess was a commenter on Dunn's article:
Accidents happen because of an amalgamation of errors, not simply because of one person's "human error". Design faults, management pressures, under-funding, poor training, long shifts, poor device interfaces, insufficient safety procedures, poor maintenance and many other issues come together to create an accident.
Root Cause Analysis is loved by management because it enables blame to be placed and makes the rest of the organization feel "safe" while latent problems still linger. It’s also cheaper.
Management is vital here and must support its own process. Robert Latino, in Supporting Root Cause Analysis: A Manager's Perspective, Reliability Center, Inc., Jan/Feb 2001, says:
One role of a manager is to be a protector of those who utilize the process. Sometimes we refer to this as providing “air cover” for ground troops. He must guarantee protection for those who act as the company would wish them to.
Gerald Blair, Senior Lecturer in VLSI Design at the Department of Electrical Engineering, The University of Edinburgh, says the manager has three major roles to play:
1. planner and visionary;

2. provider of resources;

3. protector.

The manager should be there to protect the team insofar as they are stepping outside the comfort zone in the interests of failure correction and the overall profitability of the company.
Teams are a useful concept and it's madness not to draw on your pool of expertise, even with contributions from those not specifically within their area of core competence. And as for sacking older "competents" who no longer fit the company vision, this is also madness. These people should be part of the team "leadership pair or trio" and a balanced solution emerges.

However, Chris Dillow's dream of unfettered worker empowerment and an end to Top-Down management is unrealistic.

Unmanaged teams with no established knowledge base, no matter how generally bright, would be directionless and unable to stay focused, as Anon's comment above outlines. HR is a case in point here. Further, there would be a cluster effect of self-justification and exclusion of the "genuine competent" in group managed teams, especially excluding an older, former sub-manager.

The old style sub-manager may be anachronistic to a point and yet he still has much to offer, not only in skills and knowledge but through stability.

Also, there's nothing wrong, per se, with a boss. A powerful boss with vision can create profit through his drive and expertise and profit creates expansion and that enables worker benefits and then everyone's happy, e.g. the former Hewlett Packard. Of course, he can also be a maniac.

It scarcely matters if a boss is a prickly customer, as long as he has vision, a brain and is eminently fair and rewarding of good intentions in the interests of the company. If his employees feel he does care about them, in other words and that they will be listened to, it's worth any amount of university management theory.

This was the case at Hewlett Packard with Lew Platt, an old company man who steadily increased profitability but was forced out by the new PC "wow girl" Carly Fiorina and it's history the disaster she and her successor brought to HP.

Similarly, it matters not how democratic the team is, if they lack expertise and are all out for themselves. Within your team, who's going to look out for you, protect you and reward you? I think it's a pipe dream to say that a nebulous, egalitarian group is going to do that.

A good team leader might though.

Leaders do rise in an organization [or can be moved across laterally] and more often than not it is not brown-nosing alone which will do it - there has to be a certain ability because those above will also stand or fall by the actions of those below. A good leader is a good leader.

I really do think that one natural leader is worth more than a mediocre, egalitarian group and will naturally form advisory groups at the least and ideally teams in line with modern practice. The very nature of a good leader is that he/she listens and acts on what is heard.

Dinosaurs who are past their sell by dates generally do not rise but older experts can maintain their place. There has to be a sort of give and take here, with some mental flexibility and some wisdom.

The problem is getting this sort of person into or retaining them in the public sector. Pay parity with the private sector is one thing but it should be diverted largely into team leaders, rather than to high-flying blow-ins expecting fat pay packets.

Of course, the public sector is a sort of "tragedy of the commons" with no aim, no profit motive and therefore no strategic direction. To try to make it profitable is an error but to reward increased efficiency and an expanded range of services is as good a goal as any.

There are key differences between the two sectors but effective failure analysis would appear to be a desirable goal common to both.

1 comment:

Welshcakes Limoncello said...

So much work in this, again. I don't know why you are not being "head hunted" by large corporations, James. Or perhaps you are. I see your point about the pharmaceutical industry maybe being a model and also your point about team leaders. And, as you say, there is nothing wrong with the old concept of just a "boss" - if he/she is a good one.