Demand Planning vs Merchandise Planning vs Supply Planning

What separates the three disciplines

Demand planning, merchandise planning and supply planning are three different questions asked about the same season. Demand planning asks what customers will buy. Merchandise planning asks what the business will own to serve that demand, what it will cost, and what it must return. Supply planning asks whether those units can be produced and landed on the dates the plan assumes. The three overlap because they share a calendar, a product hierarchy and — where the same people wear more than one hat — the same desk. But each optimizes something different, and each is blind to something the others can see clearly.

Demand planning

Demand planning is the discipline that turns a view of future consumer demand into a committed, quantified plan of what should be available to sell: how many units of which item, in which location or channel, in which week. It works in units and in time. Its natural artifact is a demand plan by item and location, phased across the selling period, with the assumptions that produced it stated explicitly. Its core judgment is about the shape of demand — when it arrives, where it concentrates, and how quickly it decays — not about whether the business can afford to serve it.

One caveat about the term, because it decides how the rest of this guide reads. Demand planning is used more broadly in some organizations and in much of the software category — stretched to cover the buy commitment, the receipt flow and the allocation rules as well. This guide uses the narrow sense throughout: demand planning produces units by item, location and week, and the buy, the receipt phasing and the allocation sit with merchandise planning, because those are the decisions the money constrains.

Merchandise planning

Merchandise planning is the discipline that decides what the business will own, in money and in units, and holds it accountable to a return. It sets the sales, margin, inventory and receipt plan, then resolves that plan into an assortment: how many options, at what depth, in which classes and channels. It works in dollars first and units second, which is the reverse of demand planning. Its artifact is the merchandise financial plan and the assortment plan that sits beneath it. Its core judgment is about trade-offs under a constraint — breadth against depth, margin against volume, this class against that one — because the money is finite in a way that demand is not.

Supply planning

Supply planning is the discipline that converts an agreed buy into vendor commitments, production slots and delivery dates that the business can actually stand behind. It works in lead times, capacity and calendars. Its artifact is a production and delivery schedule by vendor and purchase order, with in-transit visibility against it. Its core judgment is about feasibility and risk — whether a factory can take the volume, whether the fabric is committed, whether the sailing makes the floor set. Supply planning is the one of the three that routinely tells the other two they are wrong, and when it does, the constraint is real.

The three-way comparison

The last row is the useful one. Each discipline's blind spot is another discipline's entire job, so no single function can see a whole failure from where it sits. That is why the postmortem question worth asking after a season that missed is which handoff dropped the constraint, not which team was wrong.

Where forecasting sits

Forecasting is an input to all three. It is not a fourth discipline standing next to them, and organizing it as one is a reliable way to produce numbers nobody uses.

A forecast produces an estimate. A plan produces a commitment. The difference is not pedantry — it determines who is accountable when the number is wrong. A forecast that nobody has committed to is a piece of analysis; a plan is a promise with money attached. The same underlying statistical work can feed all three disciplines, in three different resolutions:

• →Demand planning consumes a unit forecast at item and location level, phased by week.
• →Merchandise planning consumes a sales, sell-through and margin forecast at class or category level, which is what sizes the buy and phases the receipts.
• →Supply planning consumes a forecast of production requirements and their timing, usually well ahead of either of the others, because raw material commitments are made before anyone knows much.

In fashion the forecasting problem has a particular character, covered at length in demand forecasting for fashion. The short version is that forecast accuracy at the individual style level is a weaker lever than it looks, and the compensating levers — buying shallower on unproven styles, holding open-to-buy back for in-season reaction, tightening the reforecast cadence — belong to merchandise planning, not to the forecast.

The overlaps that cause the confusion

Three overlaps account for the recurring argument. Naming them explicitly is what ends it.

The buy quantity. Demand planning derives a unit requirement from expected demand. Merchandise planning derives a unit buy from a financial plan and a target sell-through. These two numbers are computed from different directions and they will not agree. Neither is wrong. The reconciliation is the actual work, and it belongs to whoever owns the money, because the money is the binding constraint.

The receipt and flow plan. Merchandise planning phases receipts to support the sales plan and hold inventory to a target. Supply planning phases deliveries to match production and freight reality. Both produce a week-by-week receipt schedule and both believe they own it. In practice the merchandise plan sets the intent and the supply plan sets the feasible, and the gap between them is a decision, not an error to be smoothed over silently.

Replenishment on core. For never-out-of-stock programs — basics, evergreen core, core colorways — demand planning genuinely owns the algorithm: rate of sale, coverage, reorder timing. But the inventory it calls for is still capital, and it still sits inside a merchandise plan's inventory target. Core replenishment planned against that target belongs with the rest of the buy; a replenishment system left to run unbounded against its own service-level logic will quietly consume open-to-buy that was committed elsewhere.

The handoffs, in sequence

The sequence below is the pre-season path. It is a loop, not a line, and the return arrows matter more than the forward ones.

1. →Step one — financial targets set the envelope. Revenue, margin and inventory targets come down from the business plan. This is where integrated business planning either happens or doesn't.
2. →Step two — merchandise planning sizes the season. Sales and margin by class and month, an inventory target, and the receipt plan that connects them. Open-to-buy falls out of this.
3. →Step three — demand planning shapes the units. The financial plan becomes a unit requirement, phased across weeks and split by location, channel and size.
4. →Step four — assortment resolves options and depth. Breadth against depth inside the class budget: how many options, how deep behind each, which are proven and which are risk.
5. →Step five — supply planning tests feasibility. Vendor capacity, minimums, material lead times, production slots, freight mode and dates.
6. →Step six — the plan comes back changed. Minimums round quantities up. Capacity moves a delivery. A fabric commitment forces an earlier decision on color. Every one of these changes the money.
7. →Step seven — in-season, the loop runs weekly. Actual sell-through updates the forecast, which updates the demand plan, which pressures the merchandise plan, which — where lead time allows — pressures supply.

Here is a small illustration of the sixth step, the plan coming back changed — the return arrow that is easiest to skip. The numbers below are illustrative round numbers chosen to show the arithmetic, not benchmarks.

Those extra thousand units are not a rounding detail. They were bought with money the merchandise plan had allocated somewhere else, and they will land in the same weeks as everything else. The handoff is only complete when the supply constraint has been priced back into the merchandise plan — either by taking the money from another class, or by accepting the higher inventory and the exit risk that comes with it, deliberately.

Why the vocabulary differs from company to company

This is the actual reason people search for this comparison. The three disciplines are stable; the words are not.

In a large vertically integrated brand, these are three teams. Demand planning, merchandise planning and supply planning each have headcount, a system, a cadence and a director. The handoffs are formal — a monthly S&OP or IBP cycle, with a published calendar and a decision forum. The risk in this shape is not confusion about who does what; it is latency, because a constraint discovered on the factory floor may wait weeks for a meeting that can price it.

In a mid-market or emerging brand, these are three hats on two people. A merchandise planner does the demand planning implicitly, inside the assortment plan — the size curve, the store profile and the weekly phasing are demand planning, whatever the job title says. Sourcing or production does the supply planning inside a purchase order tracker. This is a perfectly sound operating model, and it is faster than the formal one. Its specific failure mode is that the three questions collapse into a single number in a single workbook, so when a season misses, nobody can separate a bad demand assumption from a bad margin assumption from a late delivery.

In a grocery or hardline retailer, "demand planning" often means what an apparel brand calls merchandise planning. The underlying reason is product behavior: a grocery assortment is largely continuous, so the central problem is item-store replenishment at high frequency, and the function that owns it is called demand planning. That function ends up owning much of the assortment and inventory decision-making too. Meanwhile "merchandising" in that world usually names the buying and vendor negotiation function, not planning. An apparel planner and a grocery planner can hold an entire conversation about demand planning and mean two different jobs.

The collision has practical costs. Job titles do not transfer cleanly between the two worlds, and neither do software categories: a system built for item-store replenishment against a repeating assortment is solving a different problem from a system built to plan money and options across a seasonal line, even though both are sold as planning. Evaluations that skip this distinction produce a shortlist that mixes item-store replenishment engines with seasonal money-and-options planners — systems that were never solving the same problem.

The clearest way to see the difference is to read a supply-chain planning platform's own scope next to a merchandise planner's: our comparison with o9 Solutions walks the boundary in detail, and the same exercise is worth doing with whatever is on your shortlist. If you are earlier than that and still mapping the landscape, the category-by-category breakdown of merchandising software for apparel covers what each type of system is actually for.

The same three questions run in footwear, accessories and home and furniture; what changes is the vocabulary and which question dominates.

• →Footwear plans in pairs and size runs, and the wholesale prebook commits buy quantities long before the season — which moves the center of gravity toward supply planning earlier than in apparel.
• →Accessories and bags carry an evergreen core alongside hero colorways, so a single category can contain both a genuine replenishment problem and a seasonal one, with different planning logic for each.
• →Home and furniture plans configurations and finishes against container quantities and landed cost, with introductions timed to market weeks and lifecycles that outlive any season — so several clocks run at once rather than one. Supply planning is often the dominant constraint here, not a downstream check.

None of this changes the three questions. It changes which of them dominates, and therefore which one the company names its planning function after.

The fashion problem: statistical machinery needs a repeating item

Classical demand planning technique — time series decomposition, exponential smoothing, seasonality indices, safety stock formulas, reorder points — rests on an assumption that is usually left unstated: the item you are forecasting has a history, and that history repeats. A seasonal apparel assortment is built the other way. It turns over by design, so much of it has no history of its own.

A new style-color arriving for SS has no sales history. It may have no close analogue. It will sell for a defined window and then stop, whether it worked or not, and the decision that mattered most — how deep to go — had to be made before any of that was observable. Applying a reorder point to it is a category error: there is no reorder. Safety stock in its textbook form protects a replenishable item against lead-time variability, so a one-shot seasonal buy has nothing for it to protect; the equivalent lever is planned depth and a credible exit plan. The method for splitting a range by replenishability and buffering only the part that has history is worked through in safety stock for seasonal assortments on retail-plan.com.

What does still repeat, and is therefore fair game for statistical treatment:

• →Category and class level demand. The style mix churns; the class shape across the season is far more stable.
• →Core and carryover. Basics behave like the items the textbooks assume, and should be planned that way — with replenishment logic, not seasonal logic.
• →Size curves. Size distribution by class and by location repeats even when every style inside it is new.
• →Store and channel rate-of-sale profiles. Relative store and channel productivity moves far more slowly than style mix, which is why a location profile built last season is still usable this one.
• →Attribute-level performance. Silhouette, fabrication, price band and color family carry signal across seasons in a way that individual style history cannot.

The practical implication is not that demand planning technique is useless in fashion. It is that the technique has to be pointed at the parts of the assortment that repeat, and the parts that don't have to be handled with judgment plus a fast reforecast loop. A team that tries to run statistical demand planning over a whole seasonal line will get confident numbers on the styles that need them least.

Where a merchandise planning platform stops

This is the boundary statement, and it is worth being plain about rather than hedging.

A merchandise planning platform owns the money and the assortment. It plans sales, margin, inventory and receipts; it manages open-to-buy; it resolves the plan into options and depth; it produces the buy by style-color and size; and it allocates and reforecasts what was bought as the season trades. RetailNorthstar is a merchandise planning platform, and that list is its scope.

It stops at two boundaries. On the supply side, it stops where the question becomes factory capacity, raw material commitment, production scheduling, container consolidation or landed cost by freight mode. Those are supply planning problems, they need vendor and materials data a planning platform does not hold, and a system that pretends otherwise produces dates nobody in production believes.

On the demand side, the boundary runs through replenishment rather than around it. Core and carryover replenishment planned against the merchandise plan's inventory target, and allocated alongside the rest of the buy, is inside the scope. What is outside it is high-frequency item-store replenishment against an assortment that repeats indefinitely — the grocery-shaped problem — which is what a dedicated demand engine is built for and optimized around.

Between those boundaries, the sensible arrangement with existing systems is integration rather than replacement. The merchandise plan supplies the financial envelope and the assortment intent; the demand engine returns forecasts; the supply engine returns dates and feasible quantities; and both come back as constraints that get priced into the plan rather than as numbers that arrive after the decision. The value of a connected workflow here is not that one system does everything. It is that the constraint and the money live in the same model, so a vendor minimum or a late sailing changes the plan visibly, in the week it is discovered, instead of surfacing as an unexplained variance a month later.

If you have two people and three questions

If you are not choosing between three departments, you are deciding how much structure to impose on work that two or three people already do. Three habits are worth imposing:

Keep the three questions separate even when the same person answers all of them. Write the demand assumption down as an assumption — expected sell-through, expected weekly shape — rather than burying it inside a buy quantity. When the season misses, you can then test which assumption failed.

Make the supply constraint arrive before the buy is committed, not after. Vendor minimums, capacity and lead times are knowable early. A buy sized without them is a buy that will be revised by someone else's constraint, at which point the money has already moved.

Reforecast on a fixed cadence and let it reach the money. A demand reforecast that updates a unit plan but never touches the merchandise financial plan has changed nothing about what the business will own.

None of that requires three systems or three teams. It requires the three questions to stay visibly distinct, so that the answer to one of them can be wrong without quietly corrupting the other two.