Methodology & Metrics — Banana Club

1

Sell-Through Rate (STR)

Data Layer

STR (%) = ( Units Sold ÷ Units Received ) × 100

What it measures: How efficiently inventory is being converted into revenue. A higher STR indicates strong demand alignment.

Why it matters: STR below 40% triggers governance review before replenishment. It's the primary signal for identifying slow movers and potential dead stock.

Threshold logic:
≥ 70% → Healthy
50–69% → Moderate
< 50% → At Risk

Worked Example

SHIRT-BLU-M at Store A — Delhi NCR
Units received last season: 200
Units sold: 156
STR = (156 ÷ 200) × 100 = 78%
→ Healthy. Eligible for replenishment.

2

Days of Cover (Projected Coverage)

Data Layer

Days of Cover = Available Inventory ÷ Avg. Daily Sales Velocity

What it measures: How many days current inventory will last at the current rate of sale. This is the single most important metric for triggering replenishment.

Decision trigger: When Days of Cover falls below the Dynamic Reorder Level (expressed in days), the system flags a replenishment need.

Thresholds:
≤ 1 day → Critical (stock-out imminent)
2–3 days → Low Stock
4–7 days → Monitor
> 7 days → Adequate

Worked Example

SHIRT-RED-M at Store B — Mumbai South
On-hand: 12, In-transit: 0, Returns: 0
Available = 12
Avg. daily velocity: 12 units/day
Coverage = 12 ÷ 12 = 1 Day
→ Critical. Immediate replenishment required.

3

Available Inventory (Effective Stock)

Data Layer

Available = On-Hand + In-Transit + Sellable Returns + DC Stock (if allocated)

What it measures: Total stock position including all sources that can fulfil demand within the planning horizon.

Key distinction: "On-hand" alone can be misleading. A store with 5 units on-hand but 50 in transit is not in crisis. Available Inventory gives the true picture.

Worked Example

Store A — Delhi NCR
On-hand: 30 + In-Transit: 10 + Returns: 2 + DC Stock: 150
Available = 192 units
→ Adequate. Despite low on-hand, total position is strong.

4

Dynamic Reorder Level (Target Stock)

Intelligence Layer

Target Stock = Forecasted Demand × (Lead Time + Review Period) + Safety Buffer

What it does: Calculates the inventory level at which replenishment should be triggered. Unlike static min/max, this adjusts dynamically based on changing demand and supply conditions.

Key inputs:
• Forecasted Demand — predicted units/day for the next cycle
• Lead Time — days from order to receipt (store-specific)
• Review Period — how often the system reassesses (typically 1 day in Banana Club)
• Safety Buffer — grade-adjusted protection (see #5)

Worked Example

SHIRT-STO-EXAMPLE at Store B (Grade A)
Forecast: 15 units/day
Lead Time: 2 days, Review Period: 1 day
Safety Buffer: 23 units (15% of forecast for Grade A)
Target = 15 × (2 + 1) + 23 = 68 units
On-hand at 80 → Coverage OK. But predict it crosses in 5 days.

5

Safety Buffer (Grade-Adjusted)

Intelligence Layer

Safety Buffer = Forecasted Demand × Grade Multiplier × Demand Variability Factor

Principle: Higher-performing stores get deeper protection. Grade A stores serve more customers and have higher opportunity cost per stock-out.

Grade multipliers:
Grade A (Delhi, Mumbai, Gurgaon): 15% of forecast — highest protection
Grade B (Bangalore, Hyderabad, Pune): 10% of forecast — standard protection
Grade C (Nashik, smaller cities): 5% of forecast — lean buffer

Worked Example

Same SKU, different stores:
Forecast: 120 units
Store A (Grade A): Buffer = 120 × 15% = 18 units
Store C (Grade B): Buffer = 120 × 10% = 12 units
Store E (Grade C): Buffer = 120 × 5% = 6 units
→ Grade A gets 3× the protection of Grade C.

The variability factor further adjusts based on demand consistency. Highly erratic demand gets a higher buffer regardless of store grade.

6

Open-to-Buy (OTB) Utilization

Governance Layer

OTB Utilization (%) = ( Budget Used ÷ Total Budget ) × 100

Capital control: OTB is the budget cap for new procurement (Purchase Orders). It protects working capital by limiting how much new inventory can be ordered.

Governance rules:
< 75% → Healthy — POs auto-eligible
75–90% → Caution — planner review recommended
> 90% → Near limit — only critical POs allowed
100% → Blocked — no new POs until budget reset

Worked Example

Monthly OTB Budget: ₹10,00,000
Used so far: ₹8,92,000
Remaining: ₹1,08,000
Utilization = (8,92,000 ÷ 10,00,000) × 100 = 89.2%
→ Caution. Next PO triggers planner review.

OTB only governs Purchase Orders (new procurement). DC Allocations and STOs do not consume OTB budget since they move existing inventory.

7

Forecast Accuracy

Intelligence Layer

Accuracy (%) = ( 1 − |Actual − Forecast| ÷ Actual ) × 100

What it measures: How close the system's demand predictions are to actual sales. Expressed as MAPE (Mean Absolute Percentage Error) inverted for readability.

Industry benchmarks:
> 85% → Excellent for fashion retail
70–85% → Acceptable
< 70% → Needs model recalibration

Worked Example

Forecasted demand (last week): 120 units
Actual sales: 108 units
Error = |108 − 120| ÷ 108 = 11.1%
Accuracy = 100 − 11.1 = 88.9%
→ Excellent. Model is well-calibrated.

8

Store Grading Logic (A / B / C)

Governance Layer

Grade = f( Revenue Contribution, Footfall, Sell-Through Rate, Location Tier )

What it does: Classifies stores into performance tiers. Grade influences safety buffer depth, replenishment priority, and allocation depth.

Grade definitions:
Grade A — Top 20% by revenue. Tier-1 metros. High footfall. Get deepest assortment and highest safety buffers.
Grade B — Middle 50%. Tier-2 cities. Standard allocation and buffers.
Grade C — Bottom 30%. Smaller cities. Lean inventory, tighter controls. Replenishment deprioritized in constrained supply scenarios.

Current Network

Grade A: Delhi NCR, Mumbai South, Gurgaon
Grade B: Bangalore, Hyderabad, Kolkata, Pune, Chennai, Chandigarh, Ahmedabad
Grade C: Nashik, Hyderabad-2

9

Lifecycle-Based Replenishment Control

Governance Layer

Replenishment Eligibility = f( Lifecycle Stage, Sell-Through Trend, Aging )

Principle: Not every SKU should be replenished. Lifecycle stage controls how aggressively the system recommends new stock.

Stage definitions (from PDF):
Launch — New arrivals. Conservative initial allocation. Monitor sell-through closely.
Growth — Strong demand trajectory. Aggressive replenishment. Maximize availability.
Peak — Demand plateauing. Standard replenishment. Watch for inflection point.
Decline — Sales declining. Reduce allocation depth. Flag governance review before any PO.
Clearance — End of life. Auto-replenishment completely restricted. Recommend markdown, STO to outlet, or liquidation.

Decision Gate Impact

JACKET-GRY-M → Lifecycle: Clearance
System action: All replenishment requests automatically blocked.
Governance message: "Lifecycle in Clearance — auto-replenishment restricted."
→ Planner directed to markdown or liquidation actions instead.

Lifecycle stages are assigned automatically by the intelligence engine based on sales velocity trends, aging data, and seasonality signals. Planners can manually override if needed.

10

Margin & Profitability Validation

Governance Layer

Margin Status = f( Contribution Margin %, Discount Dependency, Return Rate )

Principle: Replenishing unprofitable inventory destroys working capital. The margin filter ensures every stock movement is value-accretive.

Status definitions:
Healthy — Margin ≥ 25%. Full replenishment eligibility.
Moderate — Margin 15–24%. Replenishment allowed but flagged for review.
At Risk — Margin < 15%. Governance blocks replenishment. Requires pricing review before any stock movement.

Decision Gate Impact

SHIRT-RED-XL at Store G — Nashik
Contribution margin: 8% (below 15% threshold)
System action: Replenishment blocked by governance.
Message: "Margin at 8% — below minimum threshold. Deferred until pricing review."
→ Forces pricing correction before more capital is deployed.