Strategic Operations Management and Productivity Analysis

Adam Smith and the Foundations of Productivity

Adam Smith laid out the fundamentals of labor specialization in the 18th century. Productivity is defined as the ratio of output to factor of input. To calculate the percentage change in productivity, use the formula: (New System - Current) / Current = %.

Productivity Measurement and Quality Issues

• Quality: Quality may change while quantity or output remains constant.
• External Elements: External factors may distort measurement accuracy.
• Units: Precise and consistent units of measurement may be lacking.

Ten Strategic Operations Management Decision Areas

1. Design of goods and services
2. Managing quality
3. Process and capacity design
4. Location strategy
5. Layout strategy (facility arrangement)
6. Human Resources and job design
7. Supply chain management
8. Inventory, material requirements planning, and Just-in-Time (JIT)
9. Intermediate and short-term scheduling
10. Maintenance

Key Variables for Labor Productivity

To maintain and enhance labor productivity, several factors are essential:

• Basic education appropriate for the task.
• Proper diet of the labor force.
• Social overhead that makes labor available (e.g., transportation).
• Maintaining and enhancing skills while technology and knowledge change rapidly.

Operations Frontier and Competitive Attributes

The Operations Frontier consists of those who dominate in both price and rating. Competitive Attributes include price, response time, variety/differentiation, and quality. Operational Competency is measured by costs, cycles/delay, process flexibility, and technical innovation.

Process Strategies and Little's Law

Types of Processes

• Product-Focus: High volume, low variety, high fixed costs, and low variable costs. It involves few inputs with more output.
• Process-Focus: Low volume, high variety, high flexibility, and bursty demand. It involves some customization, many inputs/outputs, skilled personnel, high costs, and low equipment utilization.
• Repetitive: Assembly lines using modularity; typically more input than output.
• Mass Customization: Postponing differentiation to the end of the process. It uses many pre-made parts for various output versions, rapid process design, tight scheduling, and a responsive supply chain.

Little's Law for Estimating Averages

Little's Law is used for estimating averages involving Cycle Time (T), Inventory (I), and Throughput (R). Units are typically measured per hour.

• R = I / T
• Inventory Turns = 1 / T

Throughput determines the units a system can hold, receive, store, or produce in a period. It determines fixed costs and whether demand will be satisfied.

Capacity Management and Bottleneck Analysis

Bottleneck Identification

A bottleneck is the lowest effective capacity in a system.

• Process time of station: The time (t) for one unit at one station.
• Process time of system: The longest time (t) in the system (the bottleneck).
• Process cycle time: The time for one unit to go through the total system (the longest path).
• System process time: Time of bottleneck / Number of panels.
• System capacity: The inverse of the system process time.

Design and Effective Capacity

• Design Capacity: The maximum theoretical output expressed as a rate. Utilization = Actual Output / Design Capacity (taking labor hours into account).
• Effective Capacity: The capacity a firm expects given current constraints; it is often lower than design capacity. Efficiency = Actual Output / Effective Capacity.

Matching Capacity to Demand

To align capacity with demand, firms can:

• Make staff changes.
• Adjust equipment (purchase or sell).
• Improve processes to increase throughput (R).
• Redesign products.
• Add process flexibility to meet preferences.
• Close facilities.

Forecasting Methods and Strategies

Types of Forecasts

• Long-term (Strategic): Product/service design, facilities, and capacity. Uses qualitative methods for 3+ years.
• Medium-term (Tactical): Staffing, inventory levels, and financial planning. Uses quantitative methods for 3 months to 3 years.
• Short-term (Operations): Scheduling, paths, and demand. Uses both quantitative and prescriptive methods.

Qualitative Forecasting Methods

Used when little data exists, requiring intuition and experience:

• Grass Roots: Opinions from customer-contact personnel.
• Expert Consensus: Collective opinion of experts.
• Delphi Method: A panel of experts, decision-makers, and staff.
• Sales Force Composite: Estimates from salespeople aggregated for reasonableness.

Quantitative Forecasting Methods

Used when the situation is stable and historical data exists. Time-series models focus only on past values.

• Moving Average: Used when little to no trend is present. (Sum of previous demands) / n.
• Weighted Moving Average: Used when a trend is present; older data is less important.
• Exponential Smoothing: A weighted moving average where weights decline exponentially. Most recent data is weighted most. Requires a constant α (0.05 - 0.50).

New Forecast Formula: Last Period Forecast + α(Last Period Actual - Last Period Forecast).

Trend Projections and Seasonality

Linear trends use the least squares method: y = a + bx.

Seasonal Forecast Steps:

1. Find average historical demand for each season.
2. Find average demand over all seasons.
3. Find the seasonal index for each season.
4. Estimate next year's total demand (divide by 12).
5. Divide the estimate by the number of seasons, then multiply by the index.

Forecasting Formulas and Error Measurement

Trend-Adjusted Exponential Smoothing:
Ft = α(Actual Last Period) + (1 - α)(Forecast Last + Trend Est. Last)
Tt = β(Current Forecast - Last Forecast) + (1 - β)(Trend Est. Last)
Forecast with Trend = Ft + Tt

Standard Error and Correlation

The Standard Error of Estimate treats the forecast as a point estimate (the mean of a probability distribution).

The Correlation Coefficient (r) measures the strength of the relationship, where r^2 is the percentage of change in y explained by x.

Multiple Regression

y = a + b1x1 + b2x2

Tracking Signal and Bias

The Tracking Signal is the ratio of cumulative forecast errors to the Mean Absolute Deviation (MAD). A positive signal means demand is higher than forecast; a negative signal means the opposite. Bias occurs when there are consistent highs or lows from actual values.

Measuring Forecast Error

• Mean Absolute Deviation (MAD): Sum of |Actual - Forecast| / n.
• Mean Square Error (MSE): Sum of (Forecast Errors)^2 / n.
• Mean Absolute Percentage Error (MAPE):

Seasonality with Trend

• Simple: Multiply linear trend parameters by a given seasonal factor.
• Data-Driven: Deseasonalize data, compute trend with estimates, find the seasonal factor, and solve for forecasts.