« About Performance Management | Main | The Performance Analysis Framework Pt. 1 »

April 18, 2005

Performance Analysis versus Business Intelligence

What Performance Managers Manage

Because managers are responsible for production and execution, they are accountable for achievement.

That achievement comes in various flavors. When not looking for "quality", we usually express accountabilty in terms of "performance".

Good managers always work to make the achievement valuable in one way or another. Used strictly, performance names the observed effect of making progress towards some value that is targeted to be generated, captured and delivered.

In that sense, performance is obviously aimed at ultimate outcomes, and the beneficiaries of performance are naturally focused on the expression of "results"...

But analyzed as something one can manage, performance per se is also about how you get to the results. So it is about the direction of ongoing execution, and about states that have been asserted to represent progress. In fact, we might say that the essential job of a manager is to "make progress".

I.

Progress might be constructed or it might be cultivated. Either way, to understand current progress, managers must determine whether current and foreseeable conditions are complying with ones that typically generate states presumed to advance operations towards the target.

To represent progress, compliance measures ensue -- comparing actual conditions versus prescribed ones. This prescription is, in effect, a theory of progress.

Keep in mind that the "prescription" is really a kind of forecast of the future that was made in the past. When the "prescription" was made, it was influenced by the mindset of its author at that time, and its bases could range -- on the one hand from assumptions to histories (i.e., hypothesis to proof?), and on the other hand from preferences to rules (i.e., options to requirements). As suggested here, this leaves us with prescribed conditions that come in various flavors... from wishes to plans.

The problem this illuminates is that in evaluations -- a comparison of prescribed versus current conditions -- the current bases might similarly range. The evaluation doesn't just bring facts, but instead it brings a current mindset that proposes what facts are interesting, and then compares those interesting ones to the prescription from the previous mindset! This may not be helpful. For example, comparing actual (current) myths to prescribed (prior) plans tells us ... what?

So "compliance" might be technically true yet still be conceptually undisciplined and therefore have seriously debatable worth.

The point is that it's one thing to determine that you got what you asked for, but it's another thing to determine that you asked for the right thing. The former identifies a "correct" achievement, while the latter identifies a "relevant" one. Arguably, the former case should only be called a "progress assessment", while the later warrants the heftier label "performance evaluation".

Meanwhile, an even greater point is that achievement takes its meaning from the context in which it is recognized. Just as activity is not necessarily achievement, achievement is not necessarily progress -- and progress is not necessarily good performance. Sometimes, progress reports reflect somebody doing a great job at something that no longer matters... Although managers must make progress, it is not sufficient.

II.

Scientific management increasingly relies on analytics that will classify and test the recurrence and correlation of actual and prescribed conditions.

To reliably feed that analysis on a business-wide scope and scale, business intelligence systems are deployed -- really, for automated surveillance of business conditions.

The essential purpose of business intelligence (BI) is to gather enough evidence of activity patterns, in operations and in transactions, for identifying the most probable patterns associated with "valuable" outcomes.

However, business intelligence starts without an idea of why things work, and it builds a picture of what is "working" or not, regardless of the plan.

Effectively, BI generates new information to consider in the context of performance, but it does not describe performance.

This is not a problem, however, because BI is just one application of analysis, not the total analytic opportunity for determining performance.

III.

More importantly, analyzing performance takes additional different steps -- ones both contiguous to BI and at a higher level than BI, yet supportable with BI.

A strategy can be safely thought of as a "theory of advantage". Leaders promote the strategy as the mandate for the organization's execution.

Typically, when managers translate a leader's mandate into actions, the step initially considered is to decide how to configure resources and methods to leverage environmental conditions that "drive" progress.

"Drivers" are the conditions that usually obtain for value to emerge from the interactions of the organization's operations, locations and relations with other parties.

Management's job is not just to orchestrate the interactions but also to cultivate those conditions and, through designing the co-operation of conditions and interactions, to prescribe when the emergence of value is most likely supported or inhibited.

This design amounts to a "theory of progress" that the decisions follow.

That theory of progress consists of selected drivers logically arranged to interact in the organization and its operating environment. As a version of the theory is tailored to the characteristics of the specific organization, the tailoring produces a "strategic plan". Since both the environment and the organization are constantly changing, the strategic plan must be continuously adjustable yet stable.

Operational effectiveness is perceived mainly in the degree to which the plan is faithfully realized. In other words, effectiveness is a result of the succesful aligning of the drivers and the organization's capacity.

The main purpose of performance analysis is to discover and prioritize the factors having the most significant current impact on alignment -- i.e., on the "effective" realization of the plan.

Performance analysis information starts with the existing idea of why things should work (i.e., generate value), and acquires more information to test that idea against reality. (Here, BI's work provides support to both forecasting and validation, complementing performance analyses.)

The necessary alignment requires that performance-related information be continuously processed to determine when and how another adjustment (iteration) of the alignment is needed. This processing is a closed-loop of management awareness disciplines that promote the sequence of:
- studies leading to decisions;
- interactions leading to effects; and,
- confirmations leading to validated perspectives

Importantly, because performance presumes value-oriented adjustments, attempted changes promoted around the cycle must include transmitting the value proposition and perspective, not just status data and task assignments. Therefore, the acquisition and circulation of the performance analysis information is more likely to depend on successful knowledge management practices than on automated event monitoring.

IV.

For performance analysis, the organization needs to comprehensively manage information that is used to accomplish six key tasks:
1 - develop a performance logic model (theory of progress)
2 - develop the supporting plan
3 - initialize and prioritize action
4 - monitor status
5 - confirm underlying or correlated events, and
6 - reconcile significant differences.

This will primarily involve information that describes circumstances through a different set of terms than does BI. The performance analysis vocabulary includes:
goals,
assumptions,
relevance,
priority,
variance, and
status.

By articulating the value-oriented aspects of the performance logic model, those six types of information as a group characterize performance analysis in contrast to BI.

Although performance analysis and BI share attention to the last two types, performance analysis has two needs to be met:
- one, for the first four types of info to be logically interrelated as directives; and
- two, for that logic to be continually applied as a major real-time influencer on business operations.

Closing the loop on analysis, the directives are considered against feedback evidenced in variance and status. This final consideration provides the picture of how follow-up activity succeeds in turning "plans" into "actuals".

Posted by Malcolm Ryder at April 18, 2005 7:12 PM

Trackback Pings

TrackBack URL for this entry:
http://www.malcolmryder.com/cgi-bin/mt-tb.cgi/22

Comments

Post a comment