Analog's commitment to TQM had a direct
impact on how its balanced scorecard evolved. From its original creation in
1987, it underwent annual review, where various management teams addressed the
question "Are we measuring the right things?" The first operational
scorecard covered Analog's FY 1988 (see Evolution of the
First Balanced Scorecard for examples of actual scorecards). Some of the measures that we put on
that scorecard had been tracked for several years (e.g. on time delivery and new
product booking ratio). Others were in the process of being defined (e.g.
cycle time and yield). But cost and employee productivity were there as
goalless "place-holders"
as we labored to find good operational definitions. The last one,
labor turnover, was the best measure that we could think of at the time of how
well we were satisfying our employees.
The major change in the 1989 scorecard
resulted from our recognition that although the results metrics: delivery,
outgoing quality and leadtime (as seen by customers) applied equally to all of
our products, the process metrics: processes defect level, cycle time and yield
differed significantly for our two major businesses of ICs and assembled
products. It made no sense to aggregate these on the corporate scorecard,
and so we tracked them separately. Although we persisted with cost and productivity place-holders,
good metrics defied our discovery. We also recognized that direct and
indirect turnover had very different business implications and so we
decided to track them separately.
The 1990 scorecard reflected our conscious
decision to start the transition from measuring delivery performance against our
promise date (FCD), to measuring it against the customer's requested delivery
date (CRD). A detailed revenue model had shown us that our existing new
product metrics were poor predictors of our future performance and so we decide
to replace these metrics with a detailed product tracking system. However,
the desire to have some metrics associated with innovation on the scorecard led
to the inclusion of two measures of how well we were on track to our 1988-1992
strategic plan. We therefore included absolute bookings of products
introduced post-1985 and the aggregate forecast of third year bookings for the
current vintage of new products.
The scorecard remained unchanged for 1991
and 1992. I left Analog at the end of 1992 and so I have no firsthand
knowledge of its subsequent evolution. However, it is still in use today
and looks very similar to its earlier ancestors. One area that has
continued to evolve is the set of metrics associated with the new product
generation process. More than a third of the metrics on the 1996 scorecard
dealt with this process. Of the eight new product metrics on that
scorecard, half are process (vs. results) metrics dealing with issues of cycle
time, WIP and rework.