« Business-IT Alignment: Three Levels of Change | Main | The Conceptual Basis of Configuration Management Data »

August 15, 2005

Truth-based Management and a Single Version of the Facts

I.

Performance measurement has stampeded onto the scene in even casual literature discussing management.

The force of its impact is hardly about measurement per se, however. Measurement is not new. Rather, the buzz is all about what "performance" means.

Scholars, consultants, analysts, vendors and managers alike are aggressively exploring the scope of the term, searching for standardization of both its philosophy and its application in practice.

The resulting range of perspectives, from scholars to managers, is so sweeping that it often appears we could discuss nearly anything at all under the umbrella of "performance".

Despite this, much of the discussion now centers around a handful of ideas that appear with renewed urgency in many perspectives.

The two biggest of those ideas are:
- you can't manage what you can't measure; and,
- to manage effectively you must have everyone on the same page.

What makes those two the kings of the hill is that they each address the top problem in management today: operational complexity.

But -- what makes them such frustrating goals is that they each run counter to the "natural" probabilities of how management is exercised...

II.

Complexity is simply unavoidable. Business organizes its performance expectations around capabilities that actually require complexity to exist; and then it brings those capabilities into a game that features constant change. But at that point, ironically, the two most attractive characteristics sought of operations are simplicity and repetition.

Knowing that the knot of complexity should not really be untangled, management at least needs to reduce the number of significant variables crowding its moment-to-moment thinking, and improve the signal-to-noise ratio by deciding which ones to keep and which ones can be left out.

The normal strategy for this is to divide up and distribute responsibility for what must be thought about, so that fewer things must be attended to by anyone in particular. But naturally this brings the challenge of agreeing how to divide things up, and soon afterwards, how to overcome the many silos of information that result.

To recover from that, management usually brings a serious determination based on two resolutions:
- fact-based management; and,
- a single version of the truth.
In the typical management scenario, measurement is relied upon to "establish the facts", and decision-making relies on those facts to establish the type and priority of changes that will provoke improvement of performance. Interestingly, this approach makes it clear that measurement is a form of discovery -- and as the philosophers note, we tend to see what we're looking for more than we see what is there. Analysts further point out that the discovery instrument we use predetermines what can be discovered. Consultants note even further that the instrument is only as useful as its user is skilled. And so on... What emerges is a fairly definite realization that these "facts" -- born as they are through selection and invention -- are simply ideas that find agreement amongst the group of users who are considered to matter. But meanwhile, the discipline used to eliminate arbitrariness from "facts" is nothing other than competition from other facts!

Once we see that facts are actually propositions, the need for management to rely on them must move its focus from the facts themselves to the mechanism that validates the facts. When validation stops, the occasion signifies that the group intending to "stand by the facts" considers them to be "truth" -- and only at that point does management actually use the information to make decisions. In other words, the norm is truth-based management.

III.

Managers are responsible for deriving that truth. But managers everywhere have suffered from and battled information overload and inconsistency. That very experience is what has led many of them, whether politically astute or not, to acknowledge that "truth-based management" has been the reality. And, as the management of business operations has itself been segmented into different component departments or functions, the competition amongst different resulting versions of the truth drives executives crazy.

Hence the call for "a single version" of the truth. What this is supposed to mean is a single set of facts that represent the state of any subject or object to the highest degree of undisputed accuracy. The usual expectation is that this accuracy is practically indistinct from the validation that we've just discussed. But why is this expectation incorrect?

The answer is that the call for the single truth is driven by only one thing: the need for confident interpretation. This need presumes that with "one truth" there will be no contradiction in the meanings that can be drawn from the facts. Contradiction is a strong word, but within its literal definition of "speak against", it simply means that there is more than one way to describe something. In business, that variety is usually the desired rule, not the exception.

Naturally, when there is more than one way to see something, there is an underlying reason. The main reason in management is because management itself is distributed. Having different views of the same facts is justified by the same thing that justifies distributing management in the first place: the need for multiple points-of-observation. That does not conflict with the idea that facts should be accurate -- management should pursue a single version of the facts; and yet because they will be interpreted in various ways, the emphasis should be on the objective of the interpretation, not on the results of every available metric.

IV.

And that brings us to today's circumstances. But what does it have to do with measuring performance?

Performance refers to the degree of achievement towards a predefined target pursued in a predetermined way. Due to operational complexity, it is increasingly difficult to determine the extent to which achievement levels are the result of the predetermined activity path. If achievement levels are too low, too inconsistent, or too brief, then underlying chains of cause and effect must be identified to reveal the mechanism needing adjustment.

But the next thing that management has to do in that investigation is decide what it is that must be assessed about the "performance" mechanism. The assessment problem stems from confusing "performance" with other issues such as "value", "quality", "efficiency" and "execution". The common fault is that one of these issues crowds out the necessary visibility on the others, and poses as the whole story of performance.

In managing the business, each one of these issues is a mandated interpretation of circumstances -- one that results in more and more data being cultivated in a search for more precision in their respective explanations of what's going on. The challenge becomes to select from each perspective only what is necessary to construct a logical end-to-end account of the causes of achievement levels (effects).

The second thing management must decide is what degree of precision is necessary to basing the assessment on measurements. The third thing is then to decide what to do about the results of the assessment. Differences between the apparent chain of causality and the desired chain will have to be rated for their entrenchment and criticality. If the chain is altered, management may need to follow suit...

Most businesses have operated with the assumption that precise interpretation by a manager is the higher responsibility. But complexity is forcing us to realize that the base level of effective interpretation likely lies across management perspectives, not within them. Meanwhile, the higher responsibility within a given perspective is accurate but judicious observation.

V.

In an environment where complexity is the rule, performance is a concept that must be articulated across multiple perspectives. Performance is not just an event; it is the outcome of a competency that navigates diverse opportunities to direct activity towards goals. This often means seeing a single event in multiple ways, to find the chance of making it mean the most in terms of value.

A classic example of a single-fact with multiple interpretations is from the customer service world, where telephone support is expensive by the minute yet building customer relationships is the source of long-term profit. In this situation, the average length of a support call should be determinable with high accuracy and little debate. But the impact of that call length can be assessed in several dimensions. One point of view on a 10 minute call can be that 10 minutes is a high dollar cost. But another view is that the information exchanged in the ten minutes might increase the strength of the customer's efficiency thus loyalty and likelihood of buying again. Or it might signify that the support agent has now finally mastered a range of skills sufficiently to take the full range of issues in a call and resolve it without transferring it (even more expensively) to some other agent. These different dimensions are important to acknowledge and assess, but what should not be in dispute is the fact of how long the call was. Meanwhile, support costs need to be calibrated such that they enable impacts that drive the development of business value for as long as necessary. Since there is that important relationship between time, costs, and impact, the relationship must be managed, not just the parts, if the goal is to "improve levels of achievement against target levels through that means" -- that is, to improve performance.

VI.

In managing performance through measurement, the challenge is to first have an agreed model of performance that is comprehensive enough to address the important multiple dimensions; this is the "truth" aspect, the same page. Meanwhile, within any dimension of interpretation, consistency in point of view should be a goal; that is the "fact" aspect, the measurement foundation. Therefore, a mechanism must also be established to conceptually share "facts" across dimensions.

VII.

The biggest point of complication is where the interpretation of facts is also required to generate new data that must be used as "facts" in another dimension. Experience has shown that the classic way of doing this -- namely, hierarchical rollups of reports across management levels -- must be supplanted by information translation across dimensions.

This means nothing less than that a framework must be in place to provide an architecture of development and interpretation of facts. The framework is used by all management perspectives and levels. It provides a standing logic that explains the nature of the information needed to account for performance -- thus relieving anxiety about levels of accuracy and precision without discouraging them. Without this framework, performance measurement always risks merely replacing, at no reliable benefit, one kind of complexity with another.

Posted by Malcolm Ryder at August 15, 2005 7:35 AM

Trackback Pings

TrackBack URL for this entry:
http://www.malcolmryder.com/cgi-bin/mt-tb.cgi/105

Comments

Post a comment