The real opportunity in this area,
consequently, is not the manipulation of data in an attempt to
reduce forecast error but is the resulting discussion and activity
necessary to maintain some level of accuracy. Frequent
communication between marketing and production departments will be
required to include as many known changes as are practical. The
reluctance by marketing personnel to commit to an initial forecast
and then to change it at a later date will be more than offset by
the benefit of promotions without surprises and by the overall
improvements to customer service.
Improved relations between those who are
concerned with the demand for goods and those who produce them are
needed. A medium for this relationship can be frequent attention
to the forecast and the history of demand. One national
manufacturer has actually taken a materials manager and placed him
in the marketing department to manage its product forecasting.
Managing the forecasting system not only implies monitoring the
specific details and reacting where necessary but also eliminating
adjustments when the forecast is tracking correctly. Some simple
measure of the forecast error will allow management to focus on
the most critical products first. Forecast error will occur in two
fundamental ways: random variation and bias.
Random variation is indicated by forecast errors that cancel
each other over extended periods of time. The pattern of forecast
to actual sales will alternate between high and low, etc., and
cannot be avoided. However, attention should be given to safety
stocks for these products and to the forecasting parameters to
avoid overreaction by the forecasting system. Random variation is
measured by the mean absolute deviation of the forecast to demand
and is smoothed exponentially using the following formula:
Bias is a similar type of variation where the
errors do not cancel each other but are generally in error in the
same direction. Over time, this bias will lead to an increase in
the cumulative forecast error, which results in either over or
under total production. The bias is measured for each sales
period, which updates an "average" bias for the product
by also using exponential smoothing. The formula for this update
The two measures can be compared and reported
simultaneously by calculating the tracking signal, which gives a
measure of the forecast bias in relation to the MAD.
The tracking signal allows products of varying
sales volumes to be compared using the same measure. The tracking
signal varies from -1 to +1, which are the bias extremes in either
direction. Again, this calculation allows forecast managers to
focus their efforts on the critical problems first, leaving the
system to deal with the routine calculations. The meetings between
production and marketing managers can now be planned on a
practical basis. Working only on the most critical items and on
the ones with the highest tracking signal will provide the most
effective use of the forecast information.
In the area of process control, SPC
(Statistical Process Control) has proven to be a powerful tool in
ensuring higher quality product and improved employee relations.
This is not just the result of plotting a few numbers but, rather,
is the result of constant
attention to the process and of concentration on the most
serious problems first. The resulting cross-departmental
discussions encourage the most effective improvements. The same
holds true for forecasting product demand, or more specifically,
managing the forecast. Successful forecast management requires:
establishing an initial forecast that is used throughout the
company, a system for continual updating of the forecast as new
information is made available, signals to show where the most
attention is needed and, most important, involvement of the people
involved in getting the goods to the customer.
stay current on bullet-proofed manufacturing solutions, subscribe to
ezine, "The Business Basics and Best Practices Bulletin."
Simply fill in the below form and click on the subscribe button.
also send you our free
Special Report, "Five Change
Initiatives for Personal and Company Success."
personal information will never
be disclosed to any third party.
leaders have a responsibility to educate and train their team
members. Help for developing a self-directed, World Class
Manufacturing training program for your people is just a click
are welcomed to print and share this bulletin with your
manufacturing teams, peers, suppliers and upper management ...
better yet, have them signup for their own copy at:
the escalating spam-wars, it's also a good idea to WHITELIST
our bulletin mailing domain via your filtering software or
This will help guarantee that your bulletin is never deleted
Knowledge you’ll not find at offsite
seminars nor in the books at Amazon.com
Lean Manufacturing - Balanced Scorecard
ISO 9000:2000 - Strategic Planning - Supply Chain
Management - MRP Vs Lean Exercises - Kaizen Blitz
Lean Six Sigma - Value Stream Mapping
All at one Website: Good
Six Sigma Training Thinking
Out of the Box
Supply Chain Inventory Management Total
Quality Management Principles
Manufacturing Implementation Lean