Demand Planning for CPG Companies

CPG demand planning is the practice of forecasting and managing the demand for packaged consumer goods that move through retailers, distributors and online channels, and it differs from demand planning in other industries because most of the demand a manufacturer sees is not consumer demand. It is retailer ordering, shaped by promotions, shelf resets, inventory targets, forward buying and the retailer's own supply chain. A CPG planner has to forecast what shoppers will buy, how retailers will turn that into orders, and when the factory and the suppliers must be ready. This guide covers the features that make CPG planning distinctive: the gap between sell-in and sell-out, promotions and their after-effects, new products and distribution gains, forecasting hierarchies, collaboration with customers, shelf life and service measures. It includes a worked promotion example with checked arithmetic. It is part of demand planning and inventory optimization, and it leads to demand planning for beauty and personal care brands and luxury goods. The financial side is covered in cash flow management.

Sell-in and sell-out

Sell-in is what the manufacturer ships to customers. Sell-out is what shoppers buy from the retailer, measured at the point of sale. The two differ because retailers hold inventory and adjust it. When a retailer builds stock ahead of a promotion, sell-in rises before sell-out. When it reduces stock after the promotion, sell-in falls below sell-out. Over a long period, the two should converge, but over weeks they can be far apart. A planner who forecasts only from shipments sees the retailer's ordering behavior mixed with consumer demand and cannot separate the two. A planner who also uses sell-out data, where customers share it, can see the true consumption and forecast the replenishment orders from it. Where point-of-sale data is not available, syndicated market data, panel data or the customer's own reports may fill part of the gap. The key is to ask which signal drives the decision. Consumer demand drives the underlying trend. Retailer inventory policy drives the timing of orders.

Promotions and what follows them

Promotions are the main source of volatility in many CPG categories. A temporary price reduction, a display, a feature in the retailer's flyer or a multi-buy offer can raise sales for the promoted period well above the baseline. Three effects matter. The first is the lift during the promotion. The second is the dip afterwards, when shoppers who stocked up buy less. The third is cannibalization, in which the promotion takes sales from other products of the same brand or from competing items. A good forecast separates the baseline from these effects and estimates each from history.

Consider an item with baseline retail sales of 1,000 units a week (illustrative). A promotion runs in week 2. Sales are 1,000 in week 1, 2,500 in week 2, 800 in each of weeks 3 and 4, and 1,000 in weeks 5 and 6. Over six weeks the item sells 1,000 + 2,500 + 800 + 800 + 1,000 + 1,000 = 7,100 units, against 6 × 1,000 = 6,000 units at baseline. The incremental volume in the promotion week is 2,500 - 1,000 = 1,500 units. The post-promotion dip is 2 × (1,000 - 800) = 400 units. The net lift is 1,500 - 400 = 1,100 units, or 1,100 / 6,000 = 18.3 percent over the six weeks. The headline lift of 150 percent in the promotion week (1,500 / 1,000) overstates the true effect on volume by a factor of more than eight, because part of the lift was borrowed from the following weeks. If the promotion also cannibalized 200 units of another item in the same brand, the net gain for the brand falls to 900 units. The economics of the promotion have to be judged on the net result, after the cost of the discount.

Forward buying and retailer ordering

Retailers do not simply order what they sell. They order to maintain a target level of stock, adjusted for promotions, season and logistics. A retailer expecting a 2,500-unit week will order more than 2,500 units if it wants stock to remain at its target, and may order before the promotion starts to ensure shelf availability. Manufacturers often offer trade terms that encourage forward buying, which pulls sell-in into the promotion period. The planner should model the order pattern: when are orders placed, how far ahead of the promotion, and how quickly do orders return to normal afterwards? Over a year, the pattern of orders by customer often reveals regular cycles related to the retailer's replenishment system and promotion calendar. Customer-specific forecasting captures these patterns far better than a single forecast for the brand.

New products and distribution gains

New product introductions and distribution changes have a large effect on CPG volumes. A new item can start with a pipeline fill, in which retailers order initial stock for their stores and warehouses, followed by a period of repeat orders determined by actual sell-out. If the planner forecasts from the first orders only, the forecast will overstate the run rate. A gain in distribution, such as adding a retailer or listing in more stores, changes the baseline and needs its own assumption. Forecasts for new items usually draw on analogs: similar launches by size, price, category and customer. They should be updated quickly as early sell-out data arrives. See forecasting sales for a new product.

The forecasting hierarchy

CPG assortments are large. A single brand may have many sizes, flavors and pack types, sold to dozens of customers from several distribution centers. Forecasting every combination independently is noisy, and forecasting only at the brand level is not actionable. Most teams use a hierarchy. They forecast at a level where the signal is strong, such as brand and customer, then split to item, pack and location using proportions that reflect recent mix. Promotions and launches are added at the level where they happen. Statistical models handle the baseline, and planners overlay information the model cannot see. Measure accuracy at the level where decisions are made, usually item and location for supply planning, and at an aggregate level for financial planning. See SKU-level forecasting.

Collaboration with customers

Collaboration with retailers helps close the gap between sell-in and sell-out. Some retailers share their promotion calendar, inventory positions and sales forecasts. Sharing them lets the manufacturer plan production for actual need and lets the retailer improve availability. The collaboration works best with a small number of agreed measures, a regular meeting and a clear owner on each side. Where full collaboration is not possible, the planner can still request the promotion calendar and plan around it. A forecast that ignores the customer's plans is likely to be wrong in the weeks that matter most.

Shelf life and freshness

Many CPG products have a shelf life, and customers set rules about the remaining shelf life they will accept on delivery. Stock that sits in a warehouse for too long can become unsellable to key accounts even if it is still within its date. This turns excess stock into write-offs faster than the nominal shelf life suggests. The plan should consider age as well as quantity, and the production schedule should reflect minimum remaining shelf life requirements. See excess and obsolete inventory.

Service measures in CPG

Retailers often measure suppliers on delivery performance, such as the proportion of orders delivered in full and on time. Failure may bring penalties or lost shelf space. The planner should know how each customer defines service, because the definitions differ in how they treat partial deliveries, lateness and promotional orders. Service-level targets in the plan should match these measures. For high-priority customers or promotions, safety stock may be set higher. See in-stock rate.

Trade spend and the forecast

Trade promotion budgets and the volume forecast are two views of the same plan. If the promotion calendar assumes a certain lift and the forecast assumes another, the two will not reconcile. Include finance in the process so that the volume that results from the planned promotion spend is the same volume that goes into the supply plan and the revenue forecast. After each promotion, compare the actual lift with the planned lift, and update the library of promotion effects. See promotion forecasting.

Capacity, changeovers and batch sizes

CPG factories often run in batches with changeovers between products, which creates minimum run sizes and cycle times. The demand plan must be translated into a production schedule that respects these constraints. A forecast that implies a different flavor every day may be technically correct but impractical to produce. Close collaboration between planners and production schedulers helps find the best balance between batch size, inventory and service. Plan also for packaging materials and ingredients with long lead times, which may be the true constraint on flexibility.

What to measure

A balanced scorecard includes forecast accuracy and bias at the level of decision, forecast value added by step, fill rate and on-time delivery by customer, inventory days by category, aged stock, promotion forecast accuracy and new-product forecast accuracy. Review it monthly. See forecast value added.

Common mistakes

Forecasting from shipments alone mixes retailer behavior with consumer demand. Ignoring post-promotion dips leads to overstocking in the weeks after events. Treating all customers the same hides their different ordering patterns. Planning new products from the first pipeline orders overstates demand. Forecasting at too fine a level creates noise, and too coarse a level hides what operations need. Failing to link the promotion calendar with the supply plan is a frequent cause of stockouts and write-offs.

FAQs

What makes CPG demand planning different?

Most demand the manufacturer sees is retailer ordering, shaped by promotions, retailer inventory policies and shelf life, not just consumer purchases.

What is the difference between sell-in and sell-out?

Sell-in is what the manufacturer ships to retailers. Sell-out is what shoppers buy from the retailer. They diverge when retailers change their inventory.

How should promotions be forecast?

Estimate the baseline, the lift during the event, the dip afterwards and any cannibalization, using history for similar events.

Why is a customer-level forecast useful in CPG?

Each retailer has its own ordering pattern, promotion calendar and service rules, so customer-level forecasts capture timing that a brand-level forecast misses.

How do I forecast a new CPG product?

Use analogs for the run rate, separate the initial pipeline fill from repeat orders, and update quickly as sell-out data arrives.