Operational Strategy Case - PowerPoint by hrn39202


Operational Strategy Case document sample

More Info
									Operational Strategy: Balancing
    Planning and Control

            Vish V. Krishnan
     The University of Texas at Austin
Recapping from Last Class

• Coordination across functions and business units is a major
   – Creating incentive compatibilities is key.
   – Minimizing incentive incompatibilities essential for smoother
• Firms organize themselves to minimize transaction costs.
   – Markets motivate, Hierarchies coordinate.
• Double marginalization is the phenomenon under which
  profits in a vertically integrated firm is higher than the
  combined profit of unbundled firms (with monopoloy power).
• Transaction characteristics, Value chain profit patterns, and
  Channel structures should influence scope decisions.

Profits in a Vertically Integrated Company
                                                     Marginal Cost = $30,000
      100,000                                   Marginal Revenue = 100,000-1000q
                           demand                   100,000-1000 q*= $30,000
                           curve                              q* = 70
                                                           p* = $65,000
                                                  Maximum Profit = 70* (35,000)
                                                          = $2,450,000

      30,000                                              Marginal Cost = $20,000 + ps
                                                        Marginal Revenue = 100,000-1000q
                                                  q      100,000-1000 q*= $20,000 + ps
                                          200                  80,000-1000 q*= ps

       Marginal Cost = $10,000
  Marginal Revenue = 80,000-2000q                   Marginal Cost = $65,000
      80,000-2000 q*= $10,000                               q* = 35
                 q*= 35                                  p* = $82,500
              ps = $45,000                        Maximum Profit = 35* (17,500)
Newco’s Profit = 35*35,000 = $1,225,000                 = $ 612,500

     Total System Profit = SepCo’s Profit + NewCo’s Profit
                 = 1,837,500 < $2.45 million                                               3
Double Marginalization in the Microsoft Case
 Excerpts from "An Expensive Pig in the Poke: Estimating the Cost of the District Court’s Proposed
    Breakup of Microsoft”, 9/21/00. Stan Liebowitz , Professor of Economics , University of Texas at
 •  There are several reasons to believe that the price of Windows would rise. As has been noted in a
    declaration by the government’s expert Carl Shapiro, and in an Amici brief by Litan, Noll, Nordhous
    and Scherer, there is a well-known problem that economists refer to as double marginalization. The
    double marginalization problem occurs when two firms, each with market power, produce
    complementary products. Each firm attempts to charge a markup that would maximize its profits,
    taking the other firm’s markup as given, and in so doing the price for the two combined goods contains
    a higher markup than if a single firm had set a profit maximizing markup for the two goods jointly.
    Thus, if one believes that both the AppCo and the OpCo will have market power, prices after the
    breakup would be expected to increase." [at pages 3-4.] "A detailed analysis of Microsoft’s pricing
    clearly demonstrates that Microsoft’s behavior in application markets can be classified as that of a
    price cutter, or a firm following a low-price strategy. After the breakup, new leadership will exist in one
    or both companies and each will have to choose a pricing strategy. It is impossible to know in
    advance what pricing strategies will be adopted by these companies. Although it is possible that both
    the OpCo and AppCo will follow the same type of low price strategy used by Microsoft, it is likely that
    at least one, and perhaps both companies, will adopt a different, higher-price strategy, perhaps due in
    part to the double marginalization factor mentioned earlier." [at page 10.]

Roadmap for the Day
    1. Discussion of the Goal

    2. Push and Pull Control Systems

    3. Toyota Case Discussion

    4. Review of other Operational Items
       •   Robust Quality
       •   Time Pacing

Principles of The Goal: Jonah’s Wisdom

                            Maximize throughput
                            while simultaneously
                           reducing inventory and
                             operating expenses.

     Learning Objectives

     • Performance Measures From The Goal

     • Managing Bottlenecks: Setups, Lot sizes

     • Managing Uncertainties, Statistical Fluctuations

     • Revising Cost Accounting Concepts...
From Firm Level To The Plant Level...
  “The Goal of a firm is, of course, to make money”

of Processes

                                                              of Money
  The goal is translated to the plant level using the three performance
  measures (listed in the order of importance):

  1) Throughput
  2) Inventory
  3) Operating Expenses

  These objectives must not be considered in an isolated fashion -
  they must be considered jointly.
The Goal and Sub-Optimal Decision Making

  • Decisions in companies are often
     – Sales chooses to sell products that maximizes
     – Marketing/Manufacturing may be interested in
       maximizing gross margin per unit.
  • The product mix that maximizes profit may be
    different from the above.
     – Product takes more time on a bottleneck.
     – Product is more complex to design.
Conventional Cost Accounting Vs Goal

Issues raised by Goal regards Cost Accounting

              A                   B

       1 min./part          2 min./part
       10 min. setup        10 min. setup

Consider how cost-accounting evaluate the reduction
  in batch size on A:
   • Batch size, Q =100

     • Q=10
Marginal Value of Time on Various Resources

  • Marginal value of time on a non-bottleneck is ----

  • Marginal value of time on a bottleneck resource =

  • Activation and utilization are not one and the same.

  • Lot size calculated using economic order quantity idea
    overestimates value of time on non-bottlenecks.
     – economic order quantity helps determine the optimal
       batch size given setup times and inventory carrying
Bottleneck Management
 • By now, we know the enormous significance of
   bottlenecks. What are some ways to detect a
Getting More Out of a Bottleneck
  • What are various ways, identified by The Goal to
    manage bottlenecks effectively?

Impact of Uncertainty on a Production Line

   • Fluctuations accumulate on the negative side
     instead of canceling out.
      – Remember the boy-scout line
      – The Goal calls it “Dependent Events”

   • A balanced line creates a managerial nightmare.
      – Queues build up.
      – Bottleneck shifts.

Takeaways From The Goal

  1. Identify the system constraint(s)
  2. Decide how to exploit the system
  3. Subordinate everything else to the
       system constraint
  4. Elevate the system constraint
  5. Do not allow inertia to become the
       primary system constraint
Balancing Planning and
Control in Production
Stockless Production at Hewlett-Packard
 • From the HP video, you see that the process is as follows:

   Raw                  Gus         Terry             Mark            Dave
         Push(6)                   Pull(3)           Pull(1)

 Space       2 tables              2 tables          1 table

 WIP         30 units              12 units          4units       Does Pull (1)
                                                                  dominate other
Cycle        3:17 min.             1:40 min.         0:19 min.

 Rework 26 units                   10 units          3 units
Quality   Hidden                   Visible           Visible
Two basic control strategies: Push and Pull

                           When does a server work?

      PUSH                                                              PULL
 anticipating future demand                   reacting to current demand
          forecast demand and                          produce to replace the
        then produce to meet that                      number of units withdrawn
         estimated future demand                      from the downstream buffer
                                    Demand Forecast
                      A                                   B

  PULL (1)
                      A               1                   B               1
Virtues of Pull Control
     •   Tightly linked production system; Low safety nets
          – Low inventory
          – Low wastage of space and materials
     •   Quality problems low due to immediate feedback. Problems bubble up to
         the top.

     •   Cycle time lower in a Pull system

     •   What capabilities do you need to run a system like this?

     •   Embodied in Just-in-Time Production System
Just-in-Time Systems: Sailing Metaphor

     Water level represents inventory level
     When water level is high, ship can sail smoothly. But problems (rocks) are hidden.
     Only lowering the water level (reducing the buffer size) exposes the rocks.
   The Toyota Production System (TPS)
• What was the historical context in the late 1940’s when
  Toyota began working on TPS?
• The key objective of TPS: Elimination of Waste (muda)
• Elements of TPS:
• Jidoka and Kaizen:

• Just in Time

• Heijunka

How Toyota Implements/Transplants TPS?

   • At its heart, TPS is a Pull system (which is
     reactive, does not plan).

   • How does Toyota implement TPS?
      – Is there no inventory in the plant?

      – Is there no planning?

      – How long is the line shut down normally?

   • How did Toyota transplant TPS?

   • The operating word to describe Toyota is
The Seat Problem

  • What is the seat problem?
  • The key dilemma in solving the seat

  • Cost of stopping the line versus
    veering away from TPS
Quantifying the Dilemma
 • What is roughly the cost of stopping the line?
   – Lost sales?

    – Overtime - entire plant?

    – Overtime parts of the plant

 • What is the cost of not stopping the line?
   – Can it be quantified?!

Recommendations to Doug Friezen...

  • How should the seat problem be
    solved pragmatically/smartly?!

    –Immediate term

    –Medium Term

    –Long term

What Toyota Did
   •   Doug Friesen formed a team that did a 5-Why’s exercises on the seat
       problem. Discovered that:
        – Seat reordering form was illegible to supplier.
        – Supplier plant conditions needed changes (poor lighting etc.)
        – Product needed redesign for manufacturing.
   •   To resolve the problem, Toyota
        – Installed poster boards on the overflow area that displayed the
           number of cars in the area.
        – Set itself goals to gradually reduce the overflow area size to zero.
           If the size went beyond the goal, an andon-like action was
        – Sent a Toyota team to the supplier to help them make production
           changes (involve workers more deeply in the process),
           streamline Toyota’s interaction (ordering form), and redesigned
           the seat for optimum production.
   •   As a consequence of these actions, the seat problem was largely
       solved in the next few months, but product variety was continuing to
       pose challenges in the plant in other areas.
Lean Production
 •   The Toyota case exemplifies the workings of a lean enterprise - a term
     coined by MIT’s International Motor Vehicle Program (IMVP), and
     described in the book “The Machine that Changed the World”
 •   The thesis of IMVP is that production management policies are key
     determinants of plant operating performance. Plants that are lean are
     capable of simultaneously achieving quality, productivity, and variety
     (“mix complexity”) much more easily than non-lean (buffered) plants.
      – A lean enterprise is characterized by low (but not zero) inventory
         levels, quality as a part of the process, and broad spans of worker
      – Geographic location of the plant does not influence a plant’s
         performance as much as its corporate parentage and its
      – In the absence of management infrastructure, technology seems to
         have little effect on operating performance. A well-trained flexible
         work force, product designs that are easy to build, and a high
         performance supplier network are far more important.
Where the Field is Currently:

• Performance variations among global firms in the
  automotive industry have substantially narrowed.
   – World-class firms take about 11-15 hours to assemble
     (per vehicle).
   – Product development cycle time down to 18-36 months.
• Flexibility and variety are becoming increasingly important.
  Leanness + Flexibility is being called Agility.
   – The key is to plan for variety carefully, and maximize
     differentiation while containing component
• Some recent observations suggest that there are limits to
  leanness [Cusumano, 95, SMR].
   – Pushed too far, a lean enterprise may become fragile!
   – Striking a balance continues to require fine judgement,
     and is an eternal challenge!!
TPS Takeaways
    • Leanness: Elimination of Waste
      – Jidoka, Heijunka (flexibility)
      – Tight link between production and detection
    • Implementation is very pragmatic
      –   Planning and forecasting present
      –   Opportunity for recovery
      –   Exception handling
      –   Line does not stop very long
    • Applications to Other Domains
      – Services
      – Software
TPS Takeaways
    • Appropriate Technology Utilization
    • Leanness: Elimination of Waste
      – Jidoka, Heijunka (flexibility)
      – Tight link between production and detection
    • Implementation is very pragmatic
      –   Planning and forecasting present
      –   Opportunity for recovery
      –   Exception handling
      –   Line does not stop very long
    • Applications to Other Domains
      – Services
      – Software                                      29
Key Takeaways
• The choice between Push and Pull is a false choice. What
  managers need is a blend of Push and Pull to take advantage of
  the planning capabilities of Push and the excellent control
  capabilities of Pull.
   – Use planning for long lead time items, deciding buffer sizes,
     managing transitions&changes
   – Use Pull for normal production of one particular product in
     set volumes

   Toyota illustrates how to pragmatically combine planning and
      • Entire line does not stop for each andon pull
      • Line segments are buffered from each other (with small
        size buffers)
      • Close communication with suppliers about plans.
      • Immediate problem resolution and prevention.
Introduction to Quality

   Different Dimensions of “Quality” and
           Competing On Them
   Total Quality Management
   Re-Engineering
 Old-Fashioned Product Design Attitude

• Design product for maximum performance.
   – Producibility is unknown until after launch.
   – Design can be modified after launch to make it
• If you can build one prototype, you can easily build
• Product quality is manufacturing’s job.

• What are the problems with the old attitude?
   – Does design have impact on quality?

 One of the difficulties: Tolerances!

• How tolerances evolved?
   – Because it is very difficult to make parts consistently to the
     same dimension and because we don’t want to throw away
     many parts that exhibit only a minor deviation from the target
     dimension, a tolerance window was specified within which
     products could lie without being thrown away.
   – The tolerance window was merely being tolerated, the desire
     was to make parts to the target dimension. Over time, this
     idea degenerated and the general impression in the shop
     floor was that the goal is to be within the tolerance window.
   – Sony plant in San Diego and Tokyo

Get As Close to Target As Possible

  • Taguchi’s main argument : the customer hardly
    cares about whether a product falls just outside
    or inside the tolerance window.
    – “Customers do not give a hang about staying in
      spec… For customers, the proof of a product’s
      quality is in its performance when rapped,
      overloaded, dropped, splashed. You gain nothing
      in shipping a product that barely satisfies
      corporate standards over one that just fails…”
    – In other words, a firm must strive to set good,
      robust targets and get as close to this target as
      possible. Taguchi’s solution:
             The Notion of Quality Loss
Quality Loss

  • The costs of a product’s failure in the field
    (“external failure”) are much greater than fixing it
    in-house or even preventing failure altogether.
  • For assembled products, failure often arises from
    tolerance stack-ups.
     – Even when individual components are within the
       tolerance limits, subassemblies fall out of
       tolerance due to stack-up of tolerance limits.
       Taguchi’s Quality loss functions penalize all
       deviations from target values including those
       within the tolerance limits.
Quality Loss Function

            Conventional View of
            Quality Loss

           Taguchi’s Quality Loss

     Lower         Target       Upper
     Tolerance                  Tolerance
     Limit                      Limit
Not all Deviations from Target are Equal

   • Who’s the Better Shot: Sam or John?
             Sam                               John

    Consistent deviations (such as that of Sam’s) can be much more
   easily adjusted to target than scattered deviations.
       –“There is much higher probability of catastrophic stack-up from
       random/scattered deviations than from deviations that show
Setting Targets
    • While the shop-floor should be focused on
      getting as close to the target as possible, it is
      impossible to always keep deviations to zero.
    • Designers should ensure targets are set such
      that deviations from them do not lead to large
      performance fluctuations. “Design Quality in”


                          Design Parameter Values
Robust Design

• Robustness is the property of a product to function under
  a variety of operating conditions.
   – Even amidst uncontrollable disturbances in the shop floor,
     and in the customer's environment.
   – These uncontrollable disturbances coming from outside are
     called noise.
• What the customer wants from the product is highsignal
  despite noise.

               A robust product has a high signal
                          to noise ratio

Lessons Learned
• Lessons from the Goal: Translating financial measures
  into operational measures, prioritized problem-solving.
• Toyota System in Action: Pragmatic Implementation;
  Balancing planning and control; Holistic thinking.
• Design has huge impact on quality (Taguchi’s main
   – Make sure strong, robust designs enter the factory floor.
   – Ensure parameter target values are set so as to
     maximize robustness
   – Tolerances are merely being tolerated; Customers want
     the product to be closer to target. Seek to get as close to
     target as possible
• Questions: How does Re-engineering and TQM fit in
  this picture?
• Challenges in Mass Customization. The National
  Bicycle case.

To top