NetSuite demand planning is ready for a useful pilot when item history, location data, lead times and supply assumptions are reliable enough to explain the suggested orders. Start with a small, representative item population and compare the plan with business knowledge before allowing suggestions to drive purchasing. A forecast calculated from inconsistent data can look precise while recommending the wrong stock.
Oracle's Demand Planning documentation distinguishes demand plans from supply plans and requires Advanced Inventory Management. It identifies projection methods, historical and future periods, and location selection where applicable. Confirm which planning capability your account uses before adopting instructions from another planning module or an older implementation.
Agree whether the pilot is intended to improve replenishment timing, seasonal preparation, purchasing visibility or assembly supply. Choose a measurable decision, such as identifying items whose expected demand exceeds supported supply within lead time.
List the people allowed to adjust a plan and the evidence they must retain. Sales may know about a promotion; purchasing may know that a supplier has changed its minimum order. These inputs should be visible assumptions rather than undocumented edits made after the calculation.
Define the planning grain. Item-level totals may hide a shortage at one warehouse and excess at another. Equally, planning every small location separately can be misleading if the business actually replenishes through a central hub. The model should reflect the supply decision being made.
Review sales and demand history for missing periods, duplicate imports, cancelled orders, one-off contracts and abnormal returns. Determine what the selected calculation actually includes. Do not clean data by deleting inconvenient business events; retain the original record and document any planning adjustment separately.
Identify periods when stockouts suppressed shipments. A month with few sales might reflect weak demand, unavailable stock or an inactive sales channel. Those explanations lead to different forecasts. Ask the commercial owner to identify promotions, customer losses and product substitutions that the history alone cannot explain.
For new products, decide how the initial expectation will be formed and when it will be reviewed. A related product's history can be a useful assumption, but it should be labeled as such and checked against pack size, market and launch timing.
Make sure the historical series and supply quantities use compatible units. If the legacy system stored cartons and the new system stores pieces, preserve the conversion explicitly. NetSuite's units framework uses a defined base unit with transaction defaults. The planning team should understand which unit appears in each source and output.
Review predecessor and replacement items. If demand moved from an old SKU to a new SKU, treating them independently may understate the new item's future requirement. Conversely, combining distinct products merely because their descriptions match can exaggerate demand.
Check inactive, discontinued and non-replenished items. Marking an item as a planning exception should have an owner and reason. Otherwise, obsolete products can repeatedly re-enter purchasing discussions because they still have historical sales.
Examine lead times from the business's actual ordering and receiving experience, including manufacturing and transport where relevant. Separate normal lead time from exceptional disruption. Decide which value the current planning policy uses and how often it is reviewed.
Confirm preferred suppliers, minimum quantities, pack multiples, open purchase orders, planned production and existing stock. Define how overdue supply is treated. An old purchase order with an unrealistic arrival date can make a shortage appear covered when it is not.
Review safety-stock assumptions with the planner and finance team. Higher buffers consume cash and storage. The system configuration should reflect an approved service policy, not an unexplained number inherited from a spreadsheet.
A planner reviews a seasonal item with monthly shipments of 300, 310, 90 and 320 units. The low month initially suggests a sharp drop in demand. Investigation shows the item was unavailable for most of that month and 200 units of customer demand were cancelled or delayed.
The team preserves the recorded history, documents the stockout and chooses an approved planning treatment. It also discovers that the supplier's quoted lead time is 14 days, while the last several ordinary receipts arrived closer to 28 days after ordering.
The pilot compares a plan using the inherited assumptions with a reviewed plan. The purpose is to understand the suggested order timing and exposure, not to claim that one forecast will be correct. The item remains subject to planner approval until the inputs and exception handling are accepted.
For each pilot item and location, record:
Use “ready,” “ready with a documented assumption” and “blocked” as review outcomes. A blocked item might lack a reliable conversion or have unresolved duplicated history. Keep it outside automated purchasing decisions until the issue is fixed.
Retain the original forecast and compare it with observed demand using the same grain and period. Review bias and absolute errors, but also investigate whether the supply decision was useful. A forecast can be numerically close while arriving too late to influence purchasing.
Expand the population when planners can explain the data, assumptions and resulting suggestions. Demand planning becomes dependable through a repeatable review process, with the system performing calculations and accountable people deciding how those calculations should affect stock commitments.