Issue Two
- July 2002
1. EXPERT'S
CORNER
Must Your Scorecard Be Balanced?
By Arthur M.
Schneiderman
Note: Art will provide a
keynote presentation at MRT's Seventh Annual
Metrics Conference
Background
Conventional wisdom1 mandates that a scorecard contain a balance of:
financial and non-financial,
lagging (results or retrospective) and leading (process
or predictive),
externally (customer) and internally (processes)
focused,
and short-term and long-term metrics
It also demands representation within a prescriptive
framework; most often financial, customer, internal, and learning and growth.
But is this really necessary? Let's first look at the
origins of the "balanced" part of the scorecard.
The balanced scorecard resulted from the confluence of
three streams of late 1980's management thinking:
Total Quality Management (TQM) practitioners were
discovering that non-financial measures were much more useful in the day-to-day management
of their organizations ("you get what you measure") and were struggling with
determining the vital view metrics that they should use in steering their organization's
limited resources.
Accountants where loosing both the eyes and ears of
management to the new non-financial measures2 and were failing in their effort
to regain their past prominence by reengineering traditional product cost systems3
(Activity Based Costing) in light of the compelling criticism by both internal and
external advocates of the Theory of Constraints (TOC).
IT professionals were desperately seeking
non-transactional IT applications to expand their internal market from operations to
management in the hope that that would forestall their eventual relegation to a part of
those operations.
The first balanced scorecard was created in 1987 to
address the first of these issues. Although it was "balanced" in the current
sense, its inclusion of financial measures was for pragmatic not conceptual reasons (see
"How the Scorecard Became Balanced"). Three years later it was discovered by a
collaborating accounting professor and IT consultant, who recognized that it also provided
a solution to both of their professions' most pressing challanges.
But just like the three blind men (or six, depending on
which version you choose) first confronting an elephant, each of these three scorecard
proponents approached it from their own parochial perspective. To the Process Management
devotee, it remained a tool for identifying, communicating and tracking the vital few
"... measures of those processes whose improvement is critical to the success of the
organization." To the accountant, steeped in double-entry bookkeeping, income
statements and balance sheets, and the like, the need for balance and control appeared
essential. While to an IT consultant, the opportunity to create software systems that
extract managerially useful data from their costly data warehouses became a godsend.
Now I know that some of you are thinking that I'm being
grossly unfair to today's promoters of the balanced scorecard, but my point is that each
of us views the scorecard from our own often biased perspective. Right now, the accountant
and IT consultant perspectives are dominating the scorecard framework. As proof, just
search the internet using the keywords "balanced scorecard" or
"software." Virtually every major enterprise software company and accountancy
now offers a balanced scorecard product that they can even have certified as conforming to
some self-proclaimed "standard."
So to answer the question posed in the title, I'll take
the perspective that I've maintained from the scorecard's very beginning:
The most valuable use of a scorecard is as
a driver of a strategically focused improvement process and as such need not and usually
should not be "balanced."
Before you decide that my definition is too narrow,
keep in mind that strategy creation and deployment can be viewed as processes and in many
organizations are themselves in need of significant improvement.
Is a Scorecard for Control or
Improvement
Current balanced scorecard practice often mixes
measures of control and improvement. In a nutshell, there just isn't room on a manageable
scorecard for control measures. There are far too many of them. Furthermore, control
measures can only be managed at the process level. Every process in an organization has
the out-of-control potential to significantly damage stakeholder satisfaction, so ALL must
be effectively controlled (see Step 4 in my Process Management Model). Process Control is
part of every employees daily job activities and should not be singled out for special
attention.
A number of years ago, I had the opportunity to sit in
the cockpit simulator of what was then the next generation of commercial aircraft. Missing
was the vast array of instruments that we are use to seeing. They had been replaced by a
very high resolution color LCD display about the size of today's laptop screens. In the
simulation, what appeared on the screen was only the set of virtual instruments that were
critical to the particular activity that was currently going on. I was told that even that
was unnecessary, but it made the pilots feel more comfortable. The engineers had
designed the system to display only anomalous measures: instruments that were outside
their acceptable range under the current flight situation. And the automated system
already knew what had to be done and was taking the actions required to bring these
measures back to their nominal ranges.
The same approach is appropriate for control measures
in process management. Only a pattern of out-of control situations that can't be resolved
by existing recovery processes should be a candidates for scorecard inclusion. For
example, the number of "serious" out-of-control situations per month, or the
time to resolve an out-of-control situation are potential scorecard metrics, but only if
reducing their numbers is an identified strategic priority.
Scorecard metrics should be used to align
"non-production related" activities around the vital few improvements (change
from current in-control practice) that can impact achievement of the organizations
strategic objectives. Note that the introduction or improvement of process control (e.g.
SPC/SQC) itself, does fall into this category of activities.
Some balanced scorecard advocates make use of this
dashboard or control panel metaphor. That metaphor is only useful if, as with modern
flight decks, it excludes control measures and limits itself to measures that require
process improvements for the organization to be successful.
Let's now look at each of the identified elements of
scorecard balance.
Financial and Non-financial?
OK, I'll admit it right up front: I still don't
understand why we need financial measures on a scorecard at all. Many view this position
as unacceptable heresy. I do realize all-too-well the practical need to include them to
make the scorecard more palatable and sellable to executives who are often steeped in
traditional management by the financial numbers. And I acknowledge the perceived need to
signal to the the stockholders that their interests are still paramount. But to earn a
place on a scorecard, a metric should be directly actionable and I would argue that
financial measures simply are not.
Financial results are always dependant variables in the
mathematics of metrics. They are determined by the independent, non-financial metrics
which fall into two categories: controllable or uncontrollable. Only the independent,
controllable metrics are actionable. There is ongoing and very constructive debate
over whether any independent metric is really uncontrollable in the long run (for instance
exchange rates can be hedged, supplier prices contractually smoothed, and risk shed
through insurance).
For example, you cannot really manage long-term sales
(yes, I know that you can play lots of games with short-term sales numbers); you can only
manage (i.e. improve) the controllable processes that cause sales to happen (new product
development, marketing, sales force training, answer-getting, etc.). It's appropriate
measures of those processes that belong on a scorecard. And the same argument holds for
unit cost reduction.
As proof of this, just listen in on a typical
management conversation around an unfavorable variance in a financial measure. Inevitably,
the discussion will first move to "uncontrollable" causes (exchange rates,
economic conditions, supplier price increases, tight labor markets, etc.). Only if that
fails will explainers turn to the underlying processes and all too often revert to
finger-pointing instead of real root cause identification. Now compare this to a similar
discussion about a well-conceived non-financial metric where the process owner can have a
clear understanding of the root causes of variances from plan as well as credible
corrective actions.
I will reluctantly bow to pragmatism, but I can't
conceptually defend the mandatory inclusion of financial measures on a scorecard.
Companies that really benefit from a scorecard process will inevitably move the focus of
their attention to the non-financial scorecard metrics. And remember, if you can't make
good money after stretch improvements in your most critical business processes, than it's
probably time to reassess your strategy.
External and Internal
If you accept my premise that the scorecards highest
and best use is in strategically driven process improvement, than the next question is
whether it must have a balance of stakeholder (external) and process (internal) metrics.
Because they are linked to strategic imperatives, scorecards are usually crafted at the
top of the organization and deployed down to action agents who are the only ones that can
"make it happen."
Implicit in this cascade of scorecards is that the
higher level ones will generally have metrics that steer lower level scorecards to
required internal process improvements. For this reason, high level scorecards tend to be
unbalanced toward external or stakeholder metrics while lower level ones need to focus on
internal process metrics. The actions taken by higher level scorecard owners are
principally related to steering and diagnosing, while that of lower level scorecard owners
is process improvement focused.
The hierarchical structure of organizations implies
that the number of scorecards increases as you move down the organizations. Consequently
scorecard metrics, taken in their entirety, should be unbalanced in favor of internal or
process metrics. Another way of saying this is that for improvement of each external
scorecard metric, there are usually several internal improvements required, and each o
these have a place on someone's scorecard.
Leading and Lagging
This "requirement" goes to the very heart of
the issue that lead to the need for the creation of an instrument that would raise the
visibility of non-financial performance measurement in the first place. You can not manage
lagging indicators ... they are inherently after-the-fact measures. Certainly spectators
are interested in the results. It's interesting to know who won the World Series; but no
aficionado, no participant would be content with that meager information. To affect the
outcome, we need to focus on the leading indicators. So this issue translates into the
basic question of who is the BSC for? If it's for interested outsiders, lagging indicators
have their place. But if it is intended as a management tool, as a driver of future
success, then it must be dominated by leading indicators, principally process metrics ...
the only things that CAN be managed (see Selecting Scorecard Metrics for a more
detailed discussion of this subject).
Short- and Long-term Objectives
Long-term objectives run the high risk of inadvertently
incenting short-term inaction. A distant goal looses its ability to motivate in the press
of day-to-day business. By the time that that distant date approaches, the gap between
current and desired performance is likely to be insurmountable thus assuring failure. What
is a far better approach is to break a long-term objective into intermediate quarterly
and/or semiannual milestones (using for example the Half-Life Method). If progress toward
the th long-term objective is slower than required, more resources can be directed
toward its achievement or the long-term goal must be reassessed. If progress is faster,
resources being used for this improvement can be redeployed to other needed areas.
Bottom Line
Well, where does this leave us. If an organization were
to force its scorecard to contain a numerical balance (equal numbers) of measures in each
of the above categories than I would argue that nearly half of them don't really belong
there. My advice is to avoid altogether selecting scorecard metrics based on any
prescriptive and arbitrary framework that can clutter the scorecard with un-actionable
measures. Instead insist that metrics on high level scorecards focus on performance gaps
in areas deemed most important to strategically targeted stakeholders. Then deploy these
metrics down the organization to those processes whose improvement will contribute most to
their closure (see my e-paper on How to Build a Balanced Scorecard).
In my opinion, most scorecards should be unbalanced
toward internal, leading, short-term metrics. At the highest scorecard level, where
management diagnosis rather than process improvement is the main purpose, there is a place
for long-term, external metrics, but here the purpose is to trigger a review of the
appropriate lower level scorecards which should contain mostly internal, leading,
short-term measures. To accomplish this, those external metrics must be directly linkable
to metrics associated with internal drivers.
For example, customer satisfaction indices are poor
scorecard metrics at any level unless they are disaggregated into measures of the major
drivers of customer dissatisfaction such as poor responsiveness, quality, delivery, or
product features. Once the principal drivers of customer dissatisfaction are known, they
becom potential high-level scorecard metrics that are then linked to the appropriate
internal process drivers.
It's worth noting here that the popular strategy maps,
used in communicating the balanced scorecard story, may require the inclusion of
non-scorecard measures in order to weave a compelling logic path linking strategy to
action. These adjective-like measures, which perform a logic rather than action function,
need not meet the same test as effective scorecard metrics.
It's ironic that the first balanced scorecard, created
in 1987, was called the "Corporate Scorecard." But its subsequent renaming often
encourages dysfunctional behavior. When I use the term "balanced scorecard," I'm
simply bowing to the current vernacular. So don't be misled by its name, a Balanced
Scorecard need not be balanced.
NOTES:
1 - see for example: Translating Strategy into
Action – The Balanced Scorecard, Robert S. Kaplan and David P. Norton, Harvard
Business School Press, Boston, Massachusetts 1996, ISBN 0-87584-641-3, preface.
2 - see for example: Relevance Lost: The Rise
and Fall of Management Accounting, by H. Thomas Johnson and Robert S. Kaplan, Harvard
Business School Publishing, November 1986, ISBN: 0875841384
3 - see for example: Relevance Regained: From
Top-down Control to Bottom-up Empowerment, by H. Thomas Johnson, Simon & Schuster,
June 1992, ISBN: 0029165555
©1999-2001, Arthur M. Schneiderman All Rights
Reserved |
