Case Study
Rebuilding Forecasting and Supply Planning for a Growing Footwear Brand
THE PROBLEM
A fast-growing footwear brand ran its entire forecasting and planning operation on spreadsheets and disconnected data. As the catalog grew and buying windows got tighter, the same two problems kept showing up: how demand was forecasted, and how that forecast turned into buying decisions.
Forecasting
Forecasting sat outside any real planning system, so planners pieced it together by hand. That created a few recurring problems:
- Sales history, the item master, and inventory all lived in NetSuite, but nothing brought them together for forecasting, so planners had to pull and reconcile the data manually every cycle.
- With no correction for out-of-stocks, a stockout looked like falling demand rather than an empty shelf, so the forecast reflected past availability instead of real demand.
- Seasonality and holiday peaks were judged by eye rather than modeled, so recurring demand swings weren't captured consistently from cycle to cycle.
- New products had no sales history to draw on, so their first forecast had nothing to build on, and every NPI forecast came down to a judgment call.
- Promotions and sales-target adjustments had to be entered manually every cycle, with no safe way to test them.
- Size curves and warehouse placement were built ad hoc in spreadsheets, separate from the forecast itself.
Supply Planning
On the buying side, the whole model ran through Excel and Power Query every cycle, which was a manual and tedious process:
- Every buy was rebuilt in Excel from scratch, and placing a single order took two to three days of manual validation.
- Container fill, warehouse splits, and safety stock all ran on rules of thumb, and one-time buys below the minimum order quantity were tracked informally and easy to lose.
- Lead times lived in a separate system and had to be keyed in by hand. Factory breaks for holidays such as Chinese New Year were handled manually.
THE SOLUTION
Forecasting
The brand moved to Pint Planning, an on-demand forecasting engine that natively integrates with NetSuite. Item, sales, and inventory data now sync automatically. The forecasting system uses an ensemble of time-series and attribute-based-forecasting (ABF) models to accurately capture seasonality and holiday-driven demand peaks. It also supports New Product Introduction (NPI) forecasting by leveraging demand patterns from the most closely matched existing products. Correction for out-of-stocks is built in, so demand is read from actual sell-through rather than limited by past stockouts. Promo adjustments, sales-target overlays, and forecast versioning are built into the product: each adjustment sits in its own layer, and versioning protects the baseline, so planners can make a change or test a scenario without overwriting the working forecast.
Supply Planning
Buying moved to OnePint Supply Planning, which generates purchase order recommendations, with the option to export them to NetSuite directly, using statistical safety stock and geo-demand warehouse allocation. A demand-supply gap engine looks across the full planning horizon, with minimum order quantity, case-pack, and Chinese New Year blackout logic built directly in.
Every buy is also grounded in Agent-generated size curves. Instead of size splits maintained manually, OnePint builds each product's size distribution from past out-of-stock-corrected sales history and uses AI-based similarity matching to borrow a comparable product's curve when a product's own history is thin. Those curves then govern how each purchase order is allocated across sizes, so orders reflect how customers actually buy. Planners stay in the loop through side-by-side draft PO review, so the automation does the heavy lifting while human judgment still signs off. The application also enables geo-placement by using historical sales to determine how inventory should be distributed across fulfillment nodes. It identifies the optimal node for each past sale based on the customer’s Zipcode and uses this information to calculate the recommended inventory distribution for each product.
Business OUTCOME
With forecasting and planning operating as a single connected system, the business impact was both immediate and measurable:
- A demand signal planners can trust. Forecasting now produces a reliable, auditable demand signal that corrects stockouts, tracks accuracy through MAPE and WMAPE, and feeds directly into supply planning. Sales targets and promotional plans are built into that signal, so planning decisions stay aligned with the business plan.
- Faster buying cycles. Automating validation and PO submission cut the buying cycle from two to three days to one, so planners can respond to demand changes much faster.
- Higher availability with less inventory. In-stock performance has moved from the 80% range toward a target of 95%+, now backed by statistical, variability-based safety stock. At the same time, better size-curve allocation is targeting a 15–20% improvement in inventory turns, cutting excess stock and MOQ-driven overstocks without hurting service levels.
- Smarter planning, every day. Spreadsheet-based planning has been replaced with a NetSuite-native process. Attribute-Based Forecasting, a single source of truth for lead times, and real-time sell-through monitoring keep plans current as demand changes, giving planners more confidence in every inventory decision.