Name: inventory-forecasting
Author: w95

Inventory Forecasting

Overview

Master inventory optimization by accurately forecasting demand, calculating optimal stock levels, and reducing carrying costs while avoiding stockouts. Effective forecasting balances supply availability with storage costs.

Demand Forecasting Methods

1. Historical Trending

Simple Moving Average (best for stable demand):

Month | Actual Demand | 3-Month MA | 6-Month MA
-------|---------------|-----------|----------
Jan | 1,000 | - | -
Feb | 1,100 | - | -
Mar | 950 | 1,017 | -
Apr | 1,200 | 1,083 | -
May | 1,150 | 1,100 | 1,067
Jun | 1,300 | 1,217 | 1,117
Jul | 1,250 | 1,233 | 1,150
Aug | 1,100 | 1,217 | 1,183
Sep | 1,400 | 1,250 | 1,250
Oct | 1,300 | 1,267 | 1,267
Nov | 1,500 | 1,400 | 1,308
Dec | 1,450 | 1,417 | 1,350

Average Annual Demand: 1,250 units/month
Forecast for Jan Year 2: 1,417 (3-month MA)

Weighted Moving Average (recent data more important):

Recent months get higher weights:
- 3 months ago: 20% weight
- 2 months ago: 35% weight
- Last month: 45% weight

Forecast = (Sep × 0.20) + (Oct × 0.35) + (Nov × 0.45)
         = (1,400 × 0.20) + (1,300 × 0.35) + (1,500 × 0.45)
         = 280 + 455 + 675
         = 1,410 units

2. Seasonal Adjustment

Identify Seasonality:

Month | Actual | Annual Avg | Seasonal Index
-------|--------|-----------|----------------
Jan | 800 | 1,200 | 0.67 (slower)
Feb | 900 | 1,200 | 0.75
Mar | 1,100 | 1,200 | 0.92
Apr | 1,400 | 1,200 | 1.17 (faster)
May | 1,600 | 1,200 | 1.33 (peak)
Jun | 1,500 | 1,200 | 1.25
Jul | 1,400 | 1,200 | 1.17
Aug | 1,300 | 1,200 | 1.08
Sep | 1,100 | 1,200 | 0.92
Oct | 1,000 | 1,200 | 0.83
Nov | 1,200 | 1,200 | 1.00
Dec | 2,000 | 1,200 | 1.67 (holiday peak)

Average index: 12.07 / 12 = 1.00 (verification)

Seasonal Forecast (using seasonal index):

Base forecast (trend): 1,250 units
Seasonal forecast for May: 1,250 × 1.33 = 1,663 units (peak season)
Seasonal forecast for Jan: 1,250 × 0.67 = 838 units (slow season)
Seasonal forecast for Dec: 1,250 × 1.67 = 2,088 units (holiday)

3. Regression Analysis

Linear Trend with Demand Driver:

Regression: Demand = Base + (Growth Rate × Month) + (Seasonal Factor)

Data analysis:
- Base demand: 1,200 units
- Growth rate: +25 units/month (+2% monthly growth)
- Seasonal adjustments: (as above)

Forecast for Month 25 (13 months forward):
= 1,200 + (25 × 25) + (seasonal index for that month)
= 1,200 + 625 + seasonal
= 1,825 + seasonal (if month has 1.1 seasonal index: 1,825 × 1.1 = 2,008 units)

4. Demand Driver Analysis

Identify what drives demand:

Demand Driver | Correlation | Lead Time | Action
--------------|-------------|-----------|-------
Marketing spend | 0.85 (strong) | 2 weeks | Forecast up 30% if large campaign planned
Sales pipeline | 0.92 (strong) | 1 month | Request deals closing in next 30 days
Competitor actions | 0.60 (moderate) | 1 month | Monitor competitor price cuts
Seasonal events | 0.75 (strong) | 3 months | Plan for holidays, back-to-school, etc.
Economic indicators | 0.40 (weak) | 2 months | Monitor consumer confidence index
Industry growth | 0.70 (moderate) | Continuous | Factor in industry growth rate

FORECAST ADJUSTMENT:
If large marketing campaign planned in 2 weeks (30% demand lift expected):
- Normal forecast for weeks 3-6: 1,250 units/week
- Adjusted forecast: 1,250 × 1.30 = 1,625 units/week (4-week impact)

Safety Stock Calculation

Purpose: Buffer against demand variability and supply delays

Basic Safety Stock Formula:

Safety Stock = Z × σ_L × √(L)

Where:
Z = Service level factor (95% = 1.65, 99% = 2.33)
σ_L = Standard deviation of demand
L = Lead time in periods

EXAMPLE:
Average monthly demand: 1,250 units
Demand std dev: 150 units (high variability)
Lead time: 2 months
Service level target: 95% (Z = 1.65)

Safety Stock = 1.65 × 150 × √2
             = 1.65 × 150 × 1.41
             = 350 units (safety buffer)

Service Level Trade-offs:

Service Level | Z Factor | Safety Stock | Stockout Risk | Carrying Cost
--------------|----------|--------------|---------------|---------------
90% | 1.28 | 272 units | 10% | Lower
95% | 1.65 | 350 units | 5% | Moderate
99% | 2.33 | 495 units | 1% | Higher
99.9% | 3.09 | 656 units | 0.1% | Very High

TRADE-OFF: 95% service level is typical balance of stockout risk vs. cost

Reorder Point (ROP)

When to reorder (balance lead time demand + safety buffer):

ROP = (Average Daily Demand × Lead Time) + Safety Stock

EXAMPLE:
Monthly demand: 1,250 units
Daily demand: 1,250 / 30 = 41.67 units/day
Lead time: 14 days (2 weeks from supplier)
Safety stock: 350 units (calculated above)

ROP = (41.67 × 14) + 350
    = 583 + 350
    = 933 units

ACTION: When inventory hits 933 units, place new order

ROP Sensitivity to Lead Time:

Lead Time | Demand During LT | Safety Stock | ROP | Comment
-----------|------------------|--------------|-----|--------
1 week | 292 units | 350 | 642 | Tight; need good forecast
2 weeks | 583 units | 350 | 933 | Standard lead time
4 weeks | 1,167 units | 350 | 1,517 | Long lead; high ROP
8 weeks | 2,333 units | 350 | 2,683 | Very long lead; expensive to carry

Implication: Shorter lead times allow lower ROP and less inventory carrying cost.

Economic Order Quantity (EOQ)

How much to order (balance order costs vs. carrying costs):

EOQ = √(2 × D × S / H)

Where:
D = Annual demand (units)
S = Order cost per order ($)
H = Annual holding cost per unit ($)

EXAMPLE:
Annual demand: 15,000 units (1,250/month × 12)
Order cost: $50 per order (processing, shipping, receiving)
Holding cost: $10 per unit per year (20% of $50 unit cost)

EOQ = √(2 × 15,000 × 50 / 10)
    = √(150,000)
    = 387 units per order

Order frequency: 15,000 / 387 = 39 orders/year (weekly ordering)
Average inventory: 387 / 2 = 194 units (plus safety stock)

EOQ Trade-off Analysis:

Order Quantity | Order Freq | Order Cost/Yr | Carrying Cost/Yr | Total Cost | Note
--------------|-----------|---------------|------------------|-----------|------
100 units | 150 orders | $7,500 | $500 | $8,000 | Too frequent
387 units | 39 orders | $1,950 | $1,935 | $3,885 | OPTIMAL (EOQ)
500 units | 30 orders | $1,500 | $2,500 | $4,000 | Higher carrying
1,000 units | 15 orders | $750 | $5,000 | $5,750 | Bulk discount? Consider if available