Given the importance that corporations place upon “outcome oriented” decision making, it is a curious corporate phenomenon that the success of any new policy is judged not by how well it helped uphold its intended governing principle, but instead how closely the “letter” of the policy was followed.
Organizations that religiously enforce policies, rules and procedures, without regard for their underlying principles, often fail to meet the most basic of business capabilities with their data. It is rather astonishing that so late in the “Information Age” many Fortune 100 firms cannot declare with a degree of certainty how many customers they actually have.
Unconvinced? Try asking the question “So, how many customers do we have?” within a large financial conglomerate and stand back from the veritable storm of subsequent questions that arise.
Should the definition of “customer” include the party who is insured, or should it be the party paying for the insurance? Should I include counter-parties to a loan, and what about contra-parties to a trade? What about beneficiaries, guardians, conservators and the lawyers representing each? In order to approach these questions correctly, one must first carefully understand the specific principles of legal and fiduciary responsibilities the organization should have with respect to each party in order to classify them as a “customer.”
Without the use of principles, however, such simple questions cannot be easily resolved. This is particularly true in environments where different parts of the organization have policies and practices operating independently, or even at cross purposes.
For example, when IT implements a purchased solution, completes an acquisition or builds another warehouse to answer business questions, the primary focus is to accomplish the task at the lowest possible cost in the least amount of time. Unfortunately, this policy tends to inadvertently trample upon a variety of information architecture principles that serve to protect how information integrates into the organization’s existing data, including a basic business principle of “know your customer.”
For clarity, it may be best to take a step back to discuss more precisely what a principle is and how it is useful.
What is a Principle?
“A principle is ‘a fundamental, primary, or general truth, on which other truths depend.’ Thus a principle is an abstraction which subsumes a great number of concretes. It is only by means of principles that one can set one’s long-range goals and evaluate the concrete alternatives of any given moment. It is only principles that enable a man to plan his future and to achieve it.”1(Ayn Rand)
More than thirty years ago, American philosopher Ayn Rand, wrote that when principles (i.e., one’s conceptual faculties) are abandoned, we diminish our ability to project the future as individuals, and we diminish our ability to communicate as a group.
“Only fundamental principles, rationally validated, clearly understood and voluntarily accepted, can create a desirable kind of unity among men.”
“Consider: in any conflict between two men (or groups) who hold the same basic principles, it is the more consistent who wins.”2 (Ayn Rand)
Likewise, when basic principles are clearly and openly defined in a corporate environment, the decision-making process works to the advantage of the rational individuals who debate, in good faith, how best to adhere to the spirit of the principle. When principles are not clearly defined, the decision-making process works to the advantage of the irrational individuals who seek to advance private agendas and evade the spirit of the principles set forth by management. Such individuals are irrational in the sense that they seek to undermine the overall corporate welfare for the sake of advancing a private agenda, such as protecting a personal fiefdom at the expense of company profitability.
History is rife with examples of what occurs when managers blithely follow policy instead of principle. A classic example comes from the former Soviet Union where the Central Planning Authority established the principle that the output of nuts and bolts from factories should keep pace with increased industrial demand. To achieve this goal, a policy was issued that mandated that every factory produce a certain gross tonnage of nuts and bolts. Naturally, the intent of this policy was to increase the overall output of nuts and bolts. However, with no market forces to act as an objective watchdog, human ingenuity was immediately employed to use the policy as political cover to advance private agendas.
In this case, the workers of the factories created nuts and bolts that were twice the normal size, thereby meeting their gross tonnage requirements in half the time by producing half the number of nuts and bolts. The fact that the nuts and bolts of a new size could not used by anyone in industry was inconsequential to the factory workers. They had fulfilled the “letter” of the policy and thus had the necessary political cover to take the easy way out.
Before one has a good chuckle at the plight of the Central Planning Authority, we should remember that though corporations are themselves market driven, internal departments are often shielded from the realities of the marketplace, and fiefdoms and middle management fiat are commonplace. One does not need to look very far in a large corporation to find a department that is blithely cranking out the wrong size bolts to blindly fulfill the “letter” of a policy requirement.
Regardless of the particular discipline of thought, principles provide a guiding sense of what is fundamentally important, and what constitutes appropriate action and behavior. However, lists of principles are not the solution, as they may easily conflict with one another or be poorly formed. In careful contrast, it is essential that a set of principles form a unified and coherent whole.
When principles are well-formed and integrated into a cohesive system, then they represent the fundamental nature of how things exist and interact. The paradigm that is then created is what constitutes a whole and consistent “philosophy.”
When executives of organizations establish their key principles, and firmly implement them, the organization responds accordingly to test those principles in the arena of the competitive marketplace.
For example, when Ford established the principle “Quality is Job #1,” it was an important first step. However, it would have been a meaningless bromide without including the implementation of policies that empowered all employees to adhere to that principle. Policies that allowed employees to identify when daily actions were impeding quality resulted in an effective turnaround. In addition, the clear principle of “Quality is Job #1” lent the rational to individuals in the company the necessary air cover to overcome objections of the entrenched factions who wanted to keep cranking out the “wrong size nuts and bolts” in the name of some conflicting policy. In fact, in 2009, statistics from the Global Quality Research Systems Study already have demonstrated that Ford has surpassed Honda and tied Toyota in quality.3
Similarly, Japanese car manufacturers have experienced unprecedented success by empowering their assembly-line workers, giving any of them the authority to stop the assembly line if and when any individual detects a defect.
There is an expense associated with stopping the assembly line. However, the benefit of halting production to investigate the source of a defect, and its potential remedy, is now painfully clear to competing automobile manufacturers.
Once the philosophy of a corporation, or the philosophy of some discipline within the corporation, is well understood, it should be relatively easy for the various principles that comprise that philosophy to be remembered, internalized and acted upon in a consistent manner to achieve the goals of executive management.
Principles versus Policies?
Although related, policies differ from principles in several ways. The first distinction is that policies, rules, laws, procedures and guidelines are often specific applications of a principle.
The single biggest challenge that is associated with developing policies, however, is that the circumstances under which the principle are to be applied must be predicted in advance. As a result, an endless list of policies may be needed to address the various circumstances in which a principle may be applied. To illustrate this we can look at our legal system.
Unless a particular law has been violated, individuals do not face legal penalties when their actions violate a basic principle, no matter how egregious the act or how much harm results.
How frequently have we learned of an individual’s release because the law did not yet exist or the existing law was worded too vaguely to permit its proper interpretation in compliance with the U.S. Constitution?
Corporate policies and procedures are not immune to these challenges, and the policies that pertain to the treatment of data are not immune either. In all fairness, much of the harm that befalls data does not typically violate any particular corporate rule, policy or process, no matter how egregiously it might violate information architecture principles.
However, once an individual understands that they will at the very least be held accountable for an explanation on how their treatment of data has conformed to a set of published architectural principles, an internal and external dialogue begins as to what the best course of action should be.
“Principles” may be applicable to any number of things and disciplines, such as religion, law, mathematics, chemistry or physics. There are principles for personal conduct, principles for group conduct; and it does not stop there, as there are principles for every discipline and activity.
Examples of fundamental principles that we hold to be universal truths in our culture include “liberty is a birth right” and “all men are created equal.”
When well-formed and integrated into a cohesive system, principles represent a philosophy, which is the highest framework, above which further analysis and organization is not possible.
When a system of integrated principles forms a philosophy of that discipline, individuals can either agree or disagree with it, as philosophies cannot be proven, although their effects can be observed and measured.
A system of principles pertaining to information architecture should therefore articulate the highest level truths that the stakeholders of data wish to adopt.
An example of one such principle would be the principle that “data must be understandable.” As such, if data were not understandable, its resulting usefulness and value to the business would simply be non-existent or severely impaired.
Advocates of Policies
Advocates of policies would likely assert that their policies are based upon principles. Perhaps this is true in some cases, but when the associated “principles” are not clearly stated, the decision-making process becomes driven mechanically by rules. Once the process is driven by those devoted to blindly following the rules with no thought of the underlying principle, the resulting disaster is inevitable.
In contrast, those advocating a principles-based approach often think that simply stating the principles to a rarified audience is all that is required. However, the lack of policies can lead to undesirable decision making just as fast as any policy-based approach.
A robust set of principles must be well-formed and clearly formulated into a cohesive system of thought. Business processes and infrastructure must still be developed, except that all decisions should be tied to one or more specific principles, even when there may be a multitude of policies and rules involved.
The need to develop and support a principles-based approach is therefore just as important as a system that employs a strictly policy-based approach. A principles-based approach should not be an excuse to forfeit the necessary infrastructure that can provide accountability to the pertinent principles that are being applied.
Which is Better?
We have all witnessed that it is possible to fully obey “rules” and produce perverse outcomes.
We have recently seen the result in our economy of securities traders, rating agencies, regulators, compliance officers and management at all levels doing just that. And as we can painfully see, it is simply not realistic to expect that every situation can be foreseen in advance so that a corresponding rule, law or policy may be created to address the situation.
Data, for example, can be managed and manipulated in an infinite number of ways. Even if all of the rules could be thought of and documented in advance, they would be so intertwined with one another that it would be impossible to determine how to implement them effectively. Attempting a massive legal system of data management would generate such a high level of complexity that tasks could never be fully completed in a timely manner no matter how many resources could be brought to bear on the problem.
As mentioned earlier, principles are not subject to proofs, but their effects can be measured. One economic model in competition with another economic model will produce measurable effects, such as the standard of living, that illustrate how effective one model performs over another.
Metrics can be applied to the characteristics that underlie the principles to illustrate the level of maturity at which a process is performing. This holds true for data governance, as well as for any other discipline or activity.
If we start with the formation of principles, the underlying premise must be that all principles are chosen for a reason. Every principle that comprises the particular system of thought is chosen for one or more effects that occur when the principle is adhered to.
If a principle is thought to have no discernable effect if it was adhered to or ignored, by definition it would hardly have been a principle.
Knowing that principles are being implemented, and knowing that they are having the intended effect is essential to any executive. The task is to identify the reliable metrics that can illustrate the intended effect; and then of those, it is important to determine which of those metrics are more readily collectable.
A bottom-up approach to metrics would be to first identify the metrics that are in need of improvement within and across the organization, and then define the principles that would impact them favorably. This is often how policies are defined.
A top-down approach is to first determine which principles are essential to the organization. Once these are clearly understood, then the next task is to define the metrics that best illustrate the effects. Using a top-down approach, we will more readily achieve an “outcome oriented” set of measures for each principle. Although this sounds simple, some metrics can be applicable to more than one principle. To properly trace the impact of each principle, it is important to determine a set of metrics that distinctly illustrate the adherence and effect of the given principle.
Metrics that detect the effect of multiple principles are acceptable to include as well, but the ability to isolate at least one good metric related to the particular principle is critical for properly evaluating its effect.
In enterprise architecture and data governance, the outcomes that correspond to every part of the architectural framework can be measured as well. This includes the effects of standards, reference architectures, blueprints, methodologies and governance. Frequently, the more concrete and precise the artifact, the easier it is to measure its effects.
Whether more concrete or less, if the result of the architectural artifact itself cannot be measured, then the assertion would be that the architectural artifact cannot influence the outcome one way or the other.
Uses for Principles
When we consider the possibilities of principles, they are an important tool for executives to orchestrate the many parts of the organization, no matter how small or large, simple or complex the organization may be.
Just as the Ford Motor Company, which is a U.S. auto manufacturer employing union workers, surpassed the quality of Honda, a Japanese manufacturer renowned for quality, so too can other businesses.
It can depend substantially on the vision of executives and the principles that they define for their organizations to compete in the global economy.
As such, every CEO is like an admiral commanding his own naval fleet. It is the marketplace of ideas and the ability to implement those ideas that can determine the success or failure of those ideas.
However, a ship that has a captain who is either bereft of ideas, or lacks the ability to implement them, is not a ship I would eagerly enlist with to brave the high seas of the global economy.
As such, let’s begin with how we would go about defining and then measuring principles, and what the value proposition would be for successfully doing so from the perspective of an enterprise architect.
Principles Must Be Understood
If an enterprise architect cannot identify a factor that could buttress or weaken the implementation of a principle, then the principle may be poorly understood.
Measuring the efficacy of an architectural principle, or other type of principle, is generally not possible without the use of metrics. The intent of each principle must be fully understood in order to help identify the intended consequences, as well as any unintended ones.
Assuming that the enterprise architect has the appropriate breadth of experience and understands the business and IT, it will be recognized that every architectural principle, and every architectural decision, will impact the enterprise in one or more ways.
As such, it is critical to understand the downstream effects of each principle, if followed, if not followed, or if followed incorrectly. In fact, each and every architectural decision should be tracked back to the architectural principle(s) that influenced the decision.
If architectural decisions are not tracked, then it is likely that architectural principles and their associated frameworks are not being followed, as there is no way to verify that they are being followed.
Also, since there is a chance that developers may actually follow the architectural frameworks set forth, it is therefore crucial that those frameworks are thoroughly thought through.
As a result, those that develop architectural frameworks must have experience with a wide array of architectural approaches. They should have experienced many different types of results in their careers that have demonstrated a variety of characteristics, both desirable and not. But what kind of approaches do we mean?
In recent times architecture has been viewed as:
- “A skill to see the world as a complex system and understanding its interconnectedness.”4 (Sterman 2000)
- “A skill of thinking in terms of holism rather than reductionism.”5 (Ackoff 2004)
- “A method and framework for describing and understanding the interrelationships and forces that shape system behavior.”6(Senge 2006)
Viewed from the perspective of holism and interrelationships, an important prerequisite for architecture is experiential learning of different approaches and their results.
Building an Architectural Principle
Aside from simply stating a principle, how would someone communicate a principle to someone that does not share the breadth of experience as an enterprise architect?
If we were to build a principle from the ground up, a principle should consist of a “name” that readily identifies it so that it may be easily remembered. A longer “statement” that explains the principle more thoroughly should immediately follow. The next step to help convey the concept should be to explain the “rationale” for the principle in terms of what it is intended to achieve. Listing a set of implications that illustrate the various effects of the principle would also help.
Once the principle and its supporting documentation have been vetted by experienced practitioners, then there is no better way to communicate the intent of a principle than to identify meaningful metrics that would be attributable to the principle when correctly applied.
The first important consideration essential when assessing potential metrics is that there are often a wide range of metrics to choose from. One important hint, however, is that there are often metrics associated with a principle depending upon the maturity level of the enterprise and its processes.
Just as tests for each subject area will vary by grade, metrics of a given architectural discipline will vary by process maturity level.
There are five basic maturity levels of the software process maturity for CMM that, as a guideline, appear to work well for metrics-driven principles. They include:
Maturity Level 1 (Initial)
The most basic maturity level is where the environment lacks sound engineering practices, and also lacks the stability within which to define a process. In the initial level few standards exist, procedures are ad hoc by individual, and the results are driven almost entirely by individual effort.
Maturity Level 2 (Repeatable)
The next most basic maturity level is where a formal attempt is made to imitate the practices that were evaluated to be successful on previous efforts that share a lot in common. In a repeatable environment, the processes and procedures are documented so that they may be followed again in the future.
Maturity Level 3 (Defined)
At the defined level, end-to-end processes are defined and integrated into a coherent whole. These processes are then tailored to suit each project that uses them. Entrance criteria and outputs are defined for processes to help standardize the artifacts for cross-project ease of use.
Maturity Level 4 (Managed)
At this level of maturity, productivity and quality are measured for the most essential processes. This level of maturity helps to create reliable predictions of the resources required and the outcomes that will probably occur.
Maturity Level 5 (Optimized)
At the highest level of process maturity, defects are detected early and a culture of process improvement is in effect. Overall, there is little statistical variance among similar projects.
Now let’s experiment with the kind of metrics that we can apply to the principle “Data is Understandable” within the context of these maturity levels.
Metrics by Maturity Level
Nothing is purely black or white; there are always shades of grey.
The first consideration for dealing with metrics of our sample principle, “Data is Understandable,” is how to deal with the fact that there are almost always exceptions to what is being observed as the general trend. Hence, if we operate by the exception, we will accomplish absolutely nothing.
Let’s consider a potential metric, such as “there is a quantifiable percentage of fields that have or do not have a proper business definition associated with them.” There may always be exceptions.
Therefore, a valuable guideline for metrics-driven principles is that the criterion of any metric should follow the basic 80/20 rule.
This takes care of the problem that nothing is purely black or white, or typically all of one thing or another. This makes it easy to determine that there is compliance with the rule simply by determining whether the percentage is above or below 80%.
So, if we restate our metric for the “Initial Level” of maturity, our sample principle could be written as the following:
“The environment is said to be within the “Initial Level of Maturity” until the number of data fields, encompassing all files and databases have proper business definitions.”
Given the 80/20 rule, this would mean that the environment would remain in the “Initial Level of Maturity” until it was verified that at least 80% of all fields had an appropriate business definition.
Since counts of fields are relatively easy to come by, valid statistical samples could readily determine the ratio of fields having and lacking appropriate business definitions.
However, it is also natural that whatever can go up, can also go down.
As an example using a new metric for the same principle, where we fall from the “Optimized Level” of maturity down to the “Managed Level” can be:
“The environment is said to be within the ‘Optimized Level of Maturity’ as long as the number of data fields, encompassing all files and databases have been properly mapped to the reference framework, such as a conceptual data model.”
Given the 80/20 rule, this would mean that the environment would only remain in the “Optimized Level” of maturity as long as it was verified that at least 80% of all fields were mapped to the appropriate reference framework.
Sustainability of Principles
Governance of principles only becomes possible with metrics to provide transparency.
Organizations normally meet governance with a diatribe of opposition for good reason. Too frequently, governance impedes productivity, increases cost and is overly subjective. But, surprisingly, these issues can all vanish with the introduction of metrics-driven principles.
With good metrics come simple processes and controls for assessing the organization’s ability to adhere to the principles that they spotlight. Basic automated processes and simple workflow automation tools can non-intrusively track and trigger a wealth of objective measures.
The companies of the modern global economy will be driven by principles throughout their organization, where adherence to principles is automatically assessed using the automated collection of the appropriate metrics.
When any department receives work, it will enter as a service request that will be tracked, which will quietly gather metrics all along its life cycle. Business intelligence illustrations will depict how well activities comply with the applicable principles, and management will not only be metrics-driven, but rather will be principles-based metrics-driven.
In the final analysis, there is a role for both principles and policies, albeit fewer policies and more active and engaging principles. Corporations will operate based upon their principles, and governance will fade into the background as a distant memory of people telling you what you can or cannot do.
The key is a set of principles that form a unified philosophy, transparency through metrics, and non-intrusive governance that results from simple workflow automation in every department of the corporation.
Now that we have had a glimpse into how to develop and assess principles, we are ready to begin developing a working philosophy of enterprise architecture that functions like a well-oiled machine in your environment. Employing this basic approach, you too can illustrate the use of principles in decision making. As we have briefly seen, principles can apply to any area of the enterprise and any discipline to provide a unique competitive advantage in the global economy.
Metrics-driven principles can aid data governance in any organization to improve upon processes in action. More importantly, they can empower stakeholders to implement principles without necessarily prescribing how the tasks should be performed. Thus, you can measure the success of your principles without an overly burdensome set of policies, rules and procedures to implement them, and a bureaucracy of governance to oversee them.
Bureaucratic systems, policies, rules and guidelines have never led civilization to prosperity and greatness.
With these new approaches we may all avoid some of the bureaucratic overhead in our lives, while simultaneously providing a more effective way in which to live and work in a more principles-based existence. Now that’s a great way to demonstrate “Architecture Made Easy.”
Please feel free to express yourself if you enjoyed this article, and don’t hesitate to indicate which articles in the “Architecture Made Easy” series are useful to your organization. In addition, corrections, enhancements, and suggestions are always welcome and are requested. Please note that regarding this topic, there is much more to come.
[“Metrics-Driven Principles: The Philosophy of Information Architecture” is the 6th article in the Architecture Made Easy Series] (JimLuisi@OptOnline.net)
- Ayn Rand, 1986, Capitalism: The Unknown Ideal, Signet, 1986, ISBN 0-451-14795-2
- Ayn Rand, 1986, Capitalism: The Unknown Ideal, Signet, 1986, ISBN 0-451-14795-2
- J. Sterman. Business Dynamics: Systems Thinking and Modeling for a Complex World. McGraw-Hill, New York, NY, 2000.
- R. Ackoff. “Transforming the Systems Movement.” Opening Speech at 3rd International Conference on Systems Thinking in Management, May 2004. Philadelphia, PA.
- P. Senge. The Fifth Discipline. Doubleday, New York, NY, 2006.