Steel Works, Inc.

1
Steel Works, Inc.

• OM 783
• Supply Chain Strategy and Analysis

2
Background of Steel Works, Inc.

• Annual sales of 400 Million?
• Can this be possible given information in case?
• Conflicting data? What do we do about it?
• 5 Locations
• Custom division
• Develop products for single customer
• Specialty division
• Sell custom products to other companies at higher
  price

3
Inventory
4
What is the objective?

• Stay in business!
• Grow business?
• Improve customer satisfaction? How?
• Minimize inventory cost
• What are components of cost?
• How measured?
• Provide high level of customer service
• How high?
• How measured?

5
Consultant Recommendations

• Discontinue products that have low sales.
• Use statistical forecasting to predict demand.
• Reduce number of warehouses.
• Staff Reaction
• Wont work! Why?

6
What do data in case tell us?

• Exhibit 2 80 of products have high volatility?
• Maybe, but
• What about Exhibit 1?
• 5 large customers (gt 25 mill)
• 107 small customers (lt 1 mill)
• Is this a potential problem?

7
What is proper model for this scenario?

• Are we in a continuous review or a periodic
  review inventory setting?
• Why? How do we know?
• What are proper parameters for model?
• Why?
• How measure in reality?
• Can we validate our choice of model and
  assumptions?
• How?
• Look at production batch sizes given in case

8
How to analyze case?

• Model suggestions?
• Assume we use Periodic Review with r 1 month
  and L 1 week
• Calculate expected inventory and compare to Steel
  Works, Inc. actual inventory.
• What do we find?

9
Recommendations?

• Why do divisions operate separately?
• Any economies to be gained by combining?
• How can Steel Works better manage their customer
  relationship process?
• How can Steel Works make sure that they are
  continuing to use right model, right parameters,
  etc.?
• Should we adjust production parameters by
  product? How?

10
Questionable Assumptions?

• How confident are you that demand is normally
  distributed?
• How can you test?
• Are data accurate?
• How can you test?