E-Commerce Inventory Forecasting: Building an Integrated Model for a Fashion Brand

This case study is based on a real client engagement. All client identifying details, commercial terms, and specific product characteristics have been desensitised to protect confidentiality. The dollar values, order quantities, and model outputs are illustrative — they reflect the methodology and structure used in the engagement but have been adjusted with hypothetical numbers. The inventory forecasting methodology, financial model structure, and analytical framework are presented as they were applied.

Introduction

In early 2024, the founder of a fast-growing direct-to-consumer fashion brand reached out with a problem that's common among e-commerce operators: they had too much of the wrong stock and not enough of the right stock, and couldn't figure out whether the issue was purchasing, forecasting, or both.

The business sold 120+ SKUs across multiple product categories through their own website and wholesale channels. Revenue was growing 30% year-on-year, but cash was getting tighter. The founder had several spreadsheets — one for sales, one for purchase orders, one for bank balances — but none of them talked to each other.

What started as a "can you fix the inventory doc" request turned into a full financial model rebuild covering inventory forecasting, cash flow, and P&L integration.

The Brief

The client needed visibility on three questions:

1. How much of each SKU should I order, and when? — The current approach was reactive: order when stock ran low, often rush-ordering at higher cost.
2. What does my cash position look like over the next 6 months? — Purchase order commitments were made without checking the bank balance impact.
3. Is the business actually profitable at the product level? — Gross margin by SKU was known, but net profitability after warehousing, shipping, and marketing allocation was guesswork.

Methodology

Step 1: Sales Velocity and Seasonality

We started with 18 months of historical sales data, cleaned and categorised by:

• Product category (core basics, seasonal collections, limited editions)
• Channel (DTC website, wholesale, marketplaces)
• Sales velocity (fast, medium, slow movers based on turnover ratio)

For each category, we calculated:

• Average monthly sales volume
• Seasonal index (monthly multiplier based on historical patterns)
• Growth trend (month-over-month)
• Reorder point (lead time demand + safety stock)

Step 2: Inventory Policy Design

Working with the founder, we established inventory policies per category:

The key insight: core basics had reliable demand patterns and long supplier lead times (10–12 weeks from overseas factories), so carrying 6 weeks of safety stock was efficient. Seasonal items had shorter lead times from local suppliers but more volatile demand, so we held less safety stock and reordered more frequently.

Step 3: Cash Flow Integration

The inventory forecast fed directly into a monthly cash flow model:

• Inflows: projected sales revenue by channel (with payment timing — credit card = instant, wholesale = 30 days)
• Outflows: purchase order payments (triggered by order date + supplier lead time + payment terms), warehousing, shipping, marketing, overheads
• Working capital loop: the model tracked prepayments to suppliers, inventory on hand (by value and category), and receivables from wholesale partners

Step 4: The "MOQ Trap" Analysis

The most valuable insight came from analysing the trade-off between supplier minimum order quantities (MOQs) and cash flow. Several products had MOQs of 500–1,000 units per style/colour — but monthly demand for some colour variants was as low as 20–30 units.

We modelled two scenarios:

The 8% unit cost premium for smaller runs was cheaper than the cash tied up in slow-moving inventory. This finding reshaped the purchasing strategy.

Outcome

The client received:

• A driver-based inventory forecasting model covering all 120+ SKUs
• An integrated 3-way financial model with inventory as a driver
• A purchasing calendar showing recommended order timing and quantities for the next 6 months
• A cash flow forecast with 6-month forward visibility
• A supplier negotiation brief recommending smaller, more frequent runs

The model has been in monthly use since delivery. The founder now reviews two pages each month: the purchasing calendar (what to order and when) and the cash flow forecast (can we afford it?). Stock-out incidents on core basics dropped from 8–10 per quarter to 1–2. Cash flow visibility eliminated the surprise "can we pay the supplier this month?" conversations.

Key Takeaways

1. Inventory is a cash decision, not a purchasing decision. Every order consumes cash today in exchange for future revenue. The trade-off needs to be quantified.
2. SKU-level forecasting is work, but it pays off. The first build requires effort, but once the model is set up, monthly updates take 30–60 minutes.
3. MOQ savings are often an illusion. The unit price discount from bulk ordering looks good on a purchase order but destroys return on capital when stock sits unsold.
4. E-commerce brands need a 3-way model. Sales velocity, inventory purchasing, and cash flow must be connected. Separate spreadsheets miss the working capital feedback loop that causes most e-commerce cash crises.