Thursday, July 30, 2009
Technology and its applications are so complex as to be unmanageable. Edsger Dijkstra spent much of his career trying to get that message across to the business world and the budding computer industry. Today, he is remembered more for optimized search algorithms. If that isn't the industry in a nutshell...
Many years ago I tried to tell people that technology must be controlled, not managed. My slogan was "control technology or it will control you." We are so in love with the notion of management and have so much antipathy for control, that this message, like Dijkstra's, falls on deaf ears. I am not, in any way, comparing myself to Dijkstra. I am comparing his "audience" to mine. Today, the tag line on my web site (www.m2dxtx.com) is "Leadership for change, Management for effectiveness, Governance for stability." The three are not mutually exclusive but the probability of finding all three in one person is quite small.
Control is not a bad thing--it is an essential thing. Automobile travel without control would involve so much risk that no reasonable person would attempt it. The control starts with the design and production of the machinery itself. An automobile is designed to be controlled. It is also designed to function within a larger system of controls. Awareness of this allows the designers to prioritize their efforts and focus on differentiators suggested by the system rather than on mere "performance" factors. For example, it would be a waste of time designing a vehicle for mass production that could negotiate a 90-degree turn at 80 mph. The system of controls insures that this level of performance is unnecessary.
Use of technology without controls is also fraught with risk. We require control over the design and production of technology to insure that the product is useful and usable within the larger control framework that is our business context. Control over the use of technology is needed to insure reliability and safety for all just as our traffic laws and their enforcement produce a sense of safety and predictability for those of us on public roads.
Because no one likes the idea of control, we are calling this "governance" but make no mistake, governance must be about controls or any effort is wasted. People want and need consistency. Consistency produces contentment and the role of government, according to SunTzu (The Art of War) is a contented populace.
So, if by management, you mean counting (or accounting), you won't have success applying it to personalities or to technology. If, by management, you mean coercion, you can, for a brief time, deliver the appearance of consistency and contentment with personalities or technology, but you will only be masking a growing problem. If, by management, you mean a system of controls (governance) that produces consistency, predictability and reduced risk, only then will you be able to say that your technology (and your people) are being managed.
Fortunately, the controls necessary for effective implementation and application of technology are well understood (if largely ignored). The Software Engineering Institute's CMMI and ITIL are but two specifications for a system of controls. These are thorough and consistent and understanding them will enable you to create a system tailored to your organizational needs.
The future starts when "control" is accepted and welcomed.
Tuesday, July 21, 2009
There is concern about cost that is well founded. The problem is that costs come in a variety of disguises. Now discussions of cost have assumed inordinate importance in the questions of access. Once again, healthcare is about cost and health care is about access. These are two distinct issues.
There is virtually no one who advocates the denial of medical care on the basis of inability to pay. The entire health insurance industry emerged in response to a rise in costs driven by improvements in the science and technology of medicine. These improvement demanded better education for medical practitioners (added cost) and the technology has become more complex resulting in increased cost for both development and support.
While all this was happening, we became a nation of sedentary, over-eating, narcissists who believe in the idea that modern medicine can fix whatever we do to ourselves and make us all(well, me anyway) into beautiful people. We have learned to game the system to get the plastic surgeries we want and the pills we want and the therapies we read about.
Insurance coverages have been broadened continually in response to forces too numerous to mention with the result that more premium dollars go out requiring more premium dollars to come in. The insurance companies have developed bureaucratic defenses, requiring second opinions, demanding justification based on diagnostic testing, even making a practice of denial of the initial claim to filter out those who aren't really serious. The additional staffing and data handling is paid for by premium increases and forced out to the medical providers who increase their charges. Rising costs are everywhere and no one can join the debate with clean hands. Everyone wants someone else to absorb the costs.
In the midst of all of this, we sometimes lose sight of those who simply stay away from health care because they don't have enough money to pay for the other, even more basic necessities. We always lose sight of those who try to take care of themselves by buying health insurance, which they can afford only by accepting caps and deductibles or limitations on coverages. These people looks good in the statistics but rarely show up in the doctor's office because paying the premiums has put them into the same category as the uninsured in that they have no financial resources left over to pay for the visit. Further, they now have to live within the insurance bureaucracy that demands diagnostic justification, turning what might have been a $100 office visit into a $300 one.
Access to health care for everyone should be the sole topic in Washington. Access has relatively simple solutions. Let's solve that problem first, and, by the way, we have already agreed that ability to pay will not limit access.
Costs are a completely different issue and one that will require all interested parties to make substantial changes in thinking, planning and delivery.
It is deplorable (to use a word employed by a past President in a slightly different context) that we continue to allow doctors, administrators, insurance CEOs, technology vendors, pharma, and politicians to continue to point fingers at each other while the full cost of inadequate health care is borne by me and you, the patient/consumer.
Wednesday, July 15, 2009
- need (acknowledged)
- known cause of pain
First and foremost is availability of resources to be applied to making improvements. Data and information quality diseases have much in common with human diseases in terms of diagnosis and treatment. There is much discussion today concerning the state of health care in the U.S. The discussion focuses not on diagnosis or treatment--those aspects are well understood (if imperfectly practiced)--but on paying for the diagnosis and treatment.
It seems that financial resources or the lack of financial resources is the single most important determinant of physiological well being. If we examine the whys behind this, we soon see that expectations have much to do with it. The person without financial resources learns to expect that some problems will be chronic and learns to live with them, perhaps at a lower level of function. The financially well-off person learns to expect that every problem has a cause and a cure and that time and money will produce the expected well-being.
Neither is absolutely correct and both sets of expectations produce advantages as well as disadvantages.
We can apply the lessons of health expectations to data quality. Larger or wealthier companies expect that they will be able to attack a quality issue with sufficient resources to conquer it. Smaller or less well-off organizations will not feel able to dedicate one or more people to the issue and will elect to "do the best they can" (see previous post). Small business leaders will see that everyone must be involved in the solution for it to work and that alone will cause them to turn away from a frontal attack and "make do." Large business leaders may believe that the right manager or leader with sufficient resources can bring it off.
Again, neither is absolutely correct.
A person or an organization resigned to living with pain is always going to find it difficult or impossible to improve while a person or organization immersed in full scale battle with the problem may well miss opportunities for improvement.
As it turns out, a "data quality" campaign is like a campaign against bacteria--almost meaningless. Because the scope and scale of the campaign preclude considerations of nuance, we find that we make enemies from within the ranks and everything degenerates until nothing is happening. We can make progress against a specific bacterium or a specific quality issue but we soon realize that we can't hold those gains without creating a framework within which we can establish trust, confidence and consistency. That framework has come to be called data governance. In the case of physiological disease, the framework is Medicine.
Whether you're a have or a have not, the resource issue turns out to be far less important than we might have thought. Consider expectations first.
- Can we live with or adapt to the pain?
- Have we already adapted? How?
- What limitations are imposed by the adaptation?
- We can choose to treat symptoms, cure the disease, and prevent the disease. Which is within our reach? What can we do? What should we do?
In most cases, the best choice is to treat symptoms while making lifestyle changes to prevent the disease. Sometimes we have to cure the current disease or we die before we can implement the lifestyle changes. The point is that we always have options. A specific option must consider the past, present and future. A combination of options may produce the best result. Last but not least, have and have not is not really about resources but is about expectations. Commitment is often born of desperation when we realize that we just can't tolerate the future implied by our current expectations. Now we're really ready to do something meaningful.
Thursday, July 9, 2009
Do you want your future to be composed of cans or do you want a future of shoulds?
Should is closely related to could.
If you could do what you should do, would you do it? If you should and could but don't, what kind of future do you have before you?
Is your past characterized by "might have", "could have", "would have", "should have", or as my father was fond of saying, "mighta, woulda, coulda, shoulda?"
What's the difference between could and can? It might be knowledge or it might simply be practice. For many people, the biggest difference is the realization that there is something beyond "I can." Parents fill this role as do teachers, mentors and good friends. The process of revealing the new world of could is known as coaching.
What we should do is a function of goals, history and current context. Most of us get paid to know what what should be done. Most of us also take the easy way out and do what we can rather than what we could or should. In fact, "Do what you can," has become a universally accepted surrender. When the boss says it, it means that
- they don't know what should be done
- they don't know what could be done
- they don't want to be bothered with knocking down roadblocks
- they don't really care about the outcome
When I say it ("I did what I could.") it means
- I know what should have been done
- I know that I could have done more
- I told them but they wouldn't listen
- I was not committed to a quality result
We nearly always allow ourselves to choose the familiar path. When faced with a choice between can and could, we choose to do what we have done in the past--can.
We cannot get the data quality we need unless we have the governance we need and we can have neither if we continue to do as we've always done. This is macro as well as micro advice. Governance is not committees and steering groups, though it may have need of such. Data quality is not one definition, though that may be helpful. Both are about contextual consistency and predictability. This goal could and should be achieved in whatever ways are appropriate to the context within which the consistency is desired.
Consistency is a product of process and the foundation of improvement. Once the process produces consistent output, you have freedom to classify and categorize its output in whatever ways are suitable to its customers. We are currently engaged in trying to classify, warehouse and use inconsistent products created by inconsistent processes.
What could we do? What should we do?
Friday, July 3, 2009
Article 1 of this constitution describes a representative form of governance, recognizing that the needs for deliberation and timely decision making can best be met in this way. This was particularly true in a time when travel was by foot or by horse (or other animal propulsion) or by water propelled wither by wind, oar or paddle.
Two thoughts come to my mind:
- What might this article say if written today?
- There has been no need to modify the principles set forth during the ensuing 222 years.
All of this leads to a third thought. If the goals of corporate governance are substantially the same
- more perfect Union--Every CEO wants the company to operate as a unit, with a single purpose
- establish Justice--A sense of justice is a prerequisite for people to focus on their duties and responsibilities.
- insure domestic Tranquility--Inter-personal and inter-organizational dissension is a primary cause of lost productivity.
- provide for the common defence--The company must defend its position in the marketplace and each employee is critical to that defense.
- promote the general Welfare--This goes hand-in-hand with justice. It's human nature to want things to be better.
- secure the Blessings of Liberty--Personal liberty is always subject to the other goals.
then maybe we ought to consider whether the method should be the same.
It's hard for me to consider data governance (which is where I'm coming from) in a vacuum. The goals of data governance are substantially the same goals outlined above. Defense is about defending the integrity of the data resource. Union is about consistency. Justice and welfare is about everyone living by the same rules (which produces consistency).
I don't want to make data governance sound so impossibly complex that we throw up our hands in surrender. The message I'm transmitting is that we have models to use. We do not have to reinvent governance.
One of the difficulties in any governance model is to come up with a definition or picture of "the governed". We go through life happily assuming that everyone else is "just like me" in terms of their wants and needs. Mostly that works, but every now and then, we run into someone who isn't "just like me." When that happens we have two choices. Either we try to make the other person just like me or we adapt our view of "me" so that it includes some new parameters. In corporate life, it is exceeding dangerous to assume that anyone in a role different from ours is "just like me."
Even if we restrict ourselves to data governance, we find that we have to include as "governed" many who are filling different corporate roles and are definitely not "like" us. Again, I go back to the American Colonies in the mid eighteenth century. Imposing or trying to impose a set of rules on people whose life and needs I don't understand is destined for failure. The secondary message is: either include everyone in designing the rules or (poor second choice) understand the needs of the others before designing the rules.
Everything I see and hear about data governance is from the point of view of the person whose role is management of the data resource. There isn't a single person in the marketing department who would ever conceive of the need for data governance. Of course, we can spend time in learning to talk the marketing language and becoming familiar with marketing problems, then we can show them that some kind of governance is needed and they will agree. They might even agree to invest some time on a committee. Eventually, though, they're going to wonder if this is a good use of their most precious resource--time.
Making laws (standards) is a messy process. Much of the data governance effort is about the process--identifying stakeholders, building consensus, the political side of things, while the standards and processes become a very small box on a big diagram. My thought is that we don't even know the stakeholders until we understand the processes. The political side is essential, but there is a lot of good we could be doing if we would focus on the processes and standards.
I keep saying this because, while there may be similarity in the way two corporations handle governance, I have serious doubts whether it will ever be possible to export one company's solution to others. The political implications of forcing an outsider's will on a population would cause "failed" to be stamped on the effort nearly immediately.
Bottom line: You're on a burning platform. Don't wait for someone to save you. What do you have? What can you do? Do it!
Thursday, July 2, 2009
Let's take a look at how to win the World Series and see if we can learn anything about how to implement governance.
The first thing to recognize is that it takes an entire season--it isn't done in only a couple of weeks in October.
It takes the cooperative efforts of an entire organization.
Management must find and hire the right set of talents and abilities.
Coaches must turn the collection of talents and abilities into a team.
Each person must have the desire to excel as a part of a team.
Each person must come to share the vision of Winning The World Series.
Leadership must emerge to keep the vision in front of everyone.
We must win today's game (over and over again).
I must become a baserunner (if a batter) or keep the batter from becoming a baserunner (if in the field). I recognize that I won't succeed all the time but that doesn't keep me from wanting to succeed every time. Winning today's game means winning more of these smaller contests than I lose.
In order to win the small contests, I am prepared. I practice, I consult coaches, I talk with my teammates. I cultivate the knowledge as well as the abilities required.
I choose equipment that fits my needs.
I learn to win the contests in my own environment and in foreign environments.
I cultivate personal and team consistency.
When all of these things are done consistently and well, we find ourselves with at least the opportunity to win the World Series when October finally arrives.
What jumps out at me in all of this is the need for planning, preparation, patience, desire and commitment. I'm sure that no one out there believes that a governance implementation can be launched and completed in a few weeks. How long do you think it should take? Months? Years? Decades? Since their is no finite season or schedule to constrain us, maybe the best answer is that it will take as long as it takes.
That said, it seems incumbent on us to decide how we'll know when we have completed the task we have set for ourselves. I realize this seems self-evident and trivial but as I visit with people and groups I have developed the impression that the stable state is still undefined. What that means is that we are eternally implementing when we should be improving.
In the absence of another definition of the stable state, I have offered two principles for that state:
- No [data] pollution and
- No nasty surprises
Since these represent a whole series of contests, each of which we are committed to winning, while understanding that we won't win them all, another important property of the stable state is that it embody learning and self-modification (improvement). When we have created the property and the principles, we will have "won the world series". The next step is to understand the contests that make up "today's game" and equip ourselves physically, mentally and emotionally to win those contests.
Today's problem is that we are losing contests that we don't even know we're involved in. There's an old poker adage that says "If you look around the table and don't recognize the sucker--it's you." In [data] governance terms, if we look around and don't see the loser--it's us.