How to plan wholesale and DTC together

Why booked and forecast cannot be added

A booked wholesale quantity is a customer purchase order with a style, a colourway, a size breakdown, a quantity and a ship window on it. It is close to a certainty. What can still go wrong is bounded and nameable: an account cancels, an account goes on credit hold, a late shipment triggers a chargeback or a cancellation clause. Those are discrete events with known consequences, and a plan can carry them as a stated allowance.

A DTC forecast is a different kind of object. It has no customer attached, no ship window, and no counterparty. It is a distribution — a most-likely figure with a spread around it that widens the further out you forecast and the newer the product is. The useful information is not the midpoint but the width. A style with an expected case of four thousand units and a plausible range from two thousand eight hundred to five thousand six hundred is telling you something specific about how much protection to carry, and that information is destroyed the moment the range collapses into a single planning number.

Adding a commitment to a distribution produces a total whose error is made of two incompatible kinds of uncertainty, and once summed they cannot be separated again. You can no longer answer the only question that matters when the season goes wrong: how much of this buy was contractually protected, and how much of it was exposed. That question has to be answerable in week two, not reconstructed in the post-season review.

The two signals also arrive on different clocks, which compounds the problem. The wholesale number is final months before the season. The DTC number is a forecast at commitment, a partial read at launch, and a fact only when it is too late to buy against. Planning them as one line forces the whole buy to be decided at the wholesale clock’s pace, which is how a DTC assortment ends up having been a wholesale decision. You can track how the booked side is landing against plan — booking attainment and fill rate — with the sell-in and bookings calculator.

What locks at each gate

A channel argument in season is an argument about a decision already taken, at an earlier gate, without being framed as a decision at the time. Writing the gates down turns that argument into a question with an answer. The sequence below is the general shape; the names change by vertical, but the structure does not.

Two gates deserve more attention than they normally get. Production commit is where the total stops being negotiable and the split is still free — which means it is the wrong moment to also decide the channel allocation, and it gets decided there anyway because the purchase order asks for a ship-to and the spreadsheet asks for a channel column. The split can wait until the units are physical, and waiting is worth real money, because a month of DTC signal will have arrived in the meantime.

The reorder cutoff is the other one, and unlike every gate above it it has no system of record. It is the last date on which a reorder can be placed and still land with enough trading weeks left to sell at full price — derived from supplier lead time and the remaining full-price weeks, which means it exists only if somebody works it out and writes it down. After it, the season is what you own. Every reserve conversation is really a conversation about the distance between the first genuine DTC read and that cutoff, and if the read arrives after the cutoff then the reserve was never a chase fund at all.

Eight steps, in order

Each step consumes the output of the one before it. The two easiest to skip are the pooling test and the pre-decided scarcity rule, and they are easy to skip for the same reason: both are cheap to write down before the season and nearly impossible to agree during it.

1. 1

   Keep the two demand signals apart on the page

   A booked quantity and a forecast are different kinds of object, and one demand figure destroys the difference. Build the wholesale line and the DTC line as separate rows with separate units of certainty, and never collapse them into a single demand figure before the buy is sized. Booked wholesale is a commitment with a cancellation risk attached. Forecast DTC is a distribution with a mean and a spread. Adding them produces a number carrying two different kinds of error that can no longer be told apart.

2. 2

   Lay both calendars on one timeline

   A channel argument in season is an argument about a gate that closed months earlier. Put prebook close, production commit, first receipt, wholesale ship window and DTC launch on the same line, and mark against each gate what locks and what stays open. Writing the gates down converts an argument about intent into a question about a date, and a question about a date has an answer.

3. 3

   Take the booked quantity at its committed level, with a stated haircut

   A haircut you can see is a decision; a haircut buried inside a lower forecast cannot be audited. Plan the wholesale line at the booked quantity less an explicit cancellation and chargeback allowance, written as its own line. When the season closes, a visible allowance can be judged against what actually happened. A hidden one leaves no way to tell whether the season missed or the plan was quietly cut.

4. 4

   Express the DTC number as a range before you use its midpoint

   The width of the DTC range, not its midpoint, is what sizes the reserve. Write the DTC line as a low, expected and high case, then use the expected case in the buy and keep the range visible next to it. A DTC forecast expressed as one number cannot tell you how much protection to buy, and the reserve then gets set by convention instead of by uncertainty.

5. 5

   Size one shared reserve instead of two channel buffers

   One pool is cheaper than two buffers only to the extent the two channels do not miss in the same direction at the same time. Hold a single pool of units that either channel can draw on rather than a safety buffer inside each channel plan. How much a pool saves depends entirely on how the two channels’ misses move together, so write that assumption down beside the reserve rather than letting the arithmetic imply it. How a buffer is sized in the first place is a separate problem; the question here is whether one can be shared across two channels at all.

6. 6

   Test whether pooling actually applies before you rely on it

   Pooling only works when the reserve unit fits whichever channel calls for it. Check three things: are the two channels asking for the same size run, the same colourways, and the same delivery dates? If any answer is no, that portion of the reserve is channel-specific and should be planned and named as such rather than counted twice.

7. 7

   Check every minimum order quantity against the combined line

   A factory minimum is a constraint on the combined line, not on either channel’s line. Run each colourway and each size run against the factory minimum using the combined wholesale and DTC quantity, not the channel quantity. Where the combined number still falls short, decide deliberately between buying the overage, dropping the option, or consolidating it into a colourway both channels ordered — and record which channel the decision was made for.

8. 8

   Pre-decide scarcity and the markdown gates, then reconcile to two P&Ls

   Both of these are cheap to agree before the season and nearly impossible to agree during it. Write the allocation rule for a season that beats the forecast before the season starts, and set the first DTC markdown date against wholesale sell-through rather than against the DTC receipt curve. Then restate the same buy as two margin lines — wholesale in IMU dollars at wholesale price, DTC at realised retail AUR net of fulfilment and returns — and manage the dollars, not the blended percentage.

Sizing one buy for two channels

One style, one colourway, one season, in units. Wholesale books six thousand at order cutoff. The DTC forecast has an expected case of four thousand with a plausible range from two thousand eight hundred to five thousand six hundred. Every figure here is constructed to show the mechanism — none of it is a benchmark.

The line that does the work is the reserve. Planned the conventional way, wholesale carries six hundred units against at-once and reorder demand from accounts, short-ship exposure and size-break substitutions, and DTC carries twelve hundred against a forecast it does not trust — eighteen hundred units of protection in total. Planned as one pool, the twelve hundred units stand behind whichever channel calls first. One pool of twelve hundred covers either channel’s miss, but not both channels missing in the same week — the six hundred units saved are the price of that assumption, and it should be stated rather than assumed. How much a pool actually saves depends on how the two channels’ misses move together. If they move independently, the pool that carries the same protection is larger than the bigger of the two buffers, and the saving is smaller than six hundred. If they tend to miss together — a weak season, a late delivery that hits both — the saving approaches nothing. Twelve hundred is the most favourable case, chosen here to keep the arithmetic legible; treat it as an illustration of the mechanism rather than a computed answer, and write the correlation you are actually betting on next to the reserve line in your own plan.

The cancellation haircut is the second line worth arguing about. Three hundred units come off the booked figure as a stated allowance rather than being absorbed into a quietly lower plan. The value is not the accuracy of the number — it is that the number exists somewhere it can be looked at. When the season closes, a visible haircut lets you ask whether it was the right size. A hidden one leaves you unable to tell whether the accounts cancelled more than expected or the plan was cut before they did.

And the reserve has a condition attached that the arithmetic does not show. Twelve hundred pooled units only work if a unit can go to either channel when called for. That is a question about size runs, colourways and dates, not about totals — and it is the question the next section is about.

The honest condition on a shared reserve

Pooling works because two uncertain demands are cheaper to protect together than separately, to the degree that they are not the same demand in disguise. It stops working altogether the moment the units stop being interchangeable, and there are exactly three ways that happens. The same argument on a different axis — pooling a buffer across SKUs at class level rather than across channels — is worked through in safety stock for seasonal assortments, which owns how a buffer gets sized; the question here is only whether one can be shared across two channels.

One clarification before the three, because it is the source of an apparent contradiction. The reserve is pooled in units — physical inventory in one building that either channel can be allocated. The budget stays bucketed by channel: wholesale open-to-buy and DTC open-to-buy are tracked separately so that a wholesale over-commitment cannot quietly consume the DTC buying budget. Pooled units and separate channel OTB dollars are not in conflict; they are two different objects, and RetailNorthstar’s omnichannel assortment planning guide covers the organisational side of running both.

Different sizes. Wholesale buys a size run for a door set and tends to take depth in the middle of the curve. DTC sells a broader, flatter curve, because it serves a national customer base rather than a set of doors with local size profiles. A reserve held in the middle sizes is genuinely shared. A reserve that ends up sitting in the tails belongs to DTC, whatever the plan calls it. In footwear the same problem appears twice over, since a wide-width unit cannot serve a standard-width order at all.

Different colourways. Accounts book what fits their floor, which is a subset of the range by construction — floor space is finite and each account edits to a different customer. A colourway that only one channel ordered has a channel-specific reserve by definition, and counting it in the shared pool overstates the protection on both sides. The core and hero colours are where pooling is real; the seasonal fashion colours are where it is decorative.

Different dates. A reserve held for a DTC read that arrives in November cannot also cover a wholesale ship window that closes in September. Those are two reserves that happen to be stored together. The test is not whether the units exist but whether they are unallocated on the date each channel would need them, and this is the failure that hides best, because the warehouse balance looks correct right up until the week it does not.

When an MOQ picks your assortment

Factory minimums are a constraint on the combined line, not on either channel’s line, and that difference decides options. Take a minimum of fifteen hundred units per colourway, illustratively. A fashion colour books three hundred in wholesale and forecasts four hundred in DTC. Combined demand is seven hundred against a minimum of fifteen hundred. There are four honest responses and no good one.

The distortion is structural rather than occasional. A minimum is easiest to clear in the colourways with the largest booked quantity, and the largest booked quantity is the wholesale one, because wholesale commits first and commits in depth. The options that survive the minimum are therefore the ones accounts chose for their own floors. Nobody decides this; it is the accumulated residue of a series of individually reasonable calls.

Run a few seasons of that and the DTC range is a subset of the wholesale line, selected by buyers optimising for a different customer in a different context. It is a perfectly good explanation for soft full-price DTC sell-through, and it is invisible from the marketing plan, because the decision that produced it was recorded as a sourcing constraint rather than as an assortment choice. The counter-measure is small: record, against every minimum decision, which channel it was made for. A column of answers that all say the same thing is the signal.

The plan creates it, not the market

When a brand discounts on its own site while accounts are still selling the same product at full price, the conversation that follows is about pricing. The pricing is the symptom. The cause is a calendar decision taken months earlier, and there are four specific places it gets taken.

The DTC markdown calendar was set against the DTC receipt curve. If the first promotion is scheduled at a fixed distance after your own first receipt, it can land in the account’s third week on the floor. The account is then competing with the brand on its own newness, using inventory it paid for, and the next prebook will be smaller. The fix is to gate the first DTC promotion on wholesale sell-through reaching an agreed point, which requires knowing the account’s floor-set date at the time you set the calendar — which is to say, at prebook close.

The second place is the launch date itself. A DTC launch at or before the wholesale in-store date hands the newness advantage to the channel that did not commit for it. The third is the reserve, which drifts to DTC by default because DTC can absorb units fastest and with the least friction — so the season ends with DTC long, and inventory that is long discounts, whatever anyone intended. The fourth is the buy itself: a DTC quantity sized on a full-price sell-through assumption that the exit date never supported was always going to end in a promotion, and the promotion was decided at commitment.

Where MAP applies — most sharply in outdoor and sporting goods — this stops being a relationship question and becomes a term of the arrangement the prebook was written under. Either way, the practical move is the same: put the wholesale sell-through checkpoint into the DTC markdown calendar as a gate, and name the first markdown date at prebook close rather than discovering it in week ten. If you want the way the two channels are positioned separately, RetailNorthstar covers planning for wholesale and planning for DTC as distinct operating problems.

Allocating units when the season beats the forecast

A style that beats plan is a good problem that gets handled badly, because it arrives in the week when nobody has time to think. The rule set belongs in the plan, agreed before the season, in this order.

The second rule is the one that gets argued about, so it is worth stating plainly. “Wholesale first because it is committed” is right contractually and is not automatically right commercially. It is right about the units where a short-ship carries a penalty, and it stops being right immediately after those. Beyond the penalised units, a wholesale unit and a DTC unit are two commercial options on the same piece of inventory, and the correct one is whichever earns more contribution dollars adjusted for the chance of realising the price. The mistake is applying a contractual rule to units that carry no contractual obligation, which is what happens when nobody separates the two before the week it matters.

The same unit, told two ways

Every unit in the buy has two possible margin stories, and they are not comparable until they are stated the same way. Both start from the same landed product cost, which the landed cost and margin calculator will build if you do not already carry it at that level. The wholesale story is IMU-driven at a lower price. The DTC story starts at realised retail AUR — the average actually achieved across full-price and promoted units, not the ticket price — and then gives back fulfilment, payment processing and returns. The example below carries one unit through both. The figures are illustrative.

Read the last two rows together and the point lands. The two channels sit less than a point apart on contribution percentage and nearly two to one apart on contribution dollars. Twenty-three dollars against forty-four, on the same physical unit, at margin rates that look interchangeable. A brand managing to a margin rate would conclude the channels are equivalent and allocate on convenience.

Blend them and it gets worse. Assume for the illustration that the buy clears in full at the realised prices above: 5,700 units through wholesale and 5,200 through DTC. Wholesale contributes 5,700 at twenty-three dollars, or 131,100. DTC contributes 5,200 at forty-four dollars, or 228,800. Total contribution is 359,900 on revenue of 256,500 plus 442,000, which is 698,500 — a blended rate of 51.5%. That figure sits between the two channel rates and describes neither, which is the whole problem: it is quoted in the meeting where a marginal-unit decision gets made, and it is the one number that cannot inform it.

The dollars are also not equally certain, which is the second half of the comparison — though the uncertainty sits in a different place on each side rather than in a different amount. The wholesale twenty-three is attached to a specific purchase order at a contracted price, and what can go wrong is counterparty risk: a cancellation, a credit hold, a chargeback. The DTC forty-four is an average across units that sold at full price and units that sold on promotion, so what can go wrong is the mix: the longer the unit sits, the more of the season’s volume lands on the promoted side and the lower the realised AUR turns out to have been. Wholesale carries price certainty and counterparty risk; DTC carries customer certainty and price risk. A blended percentage is precisely the format that makes both invisible.

Where the commitment signal comes from

The structure is the same everywhere: something commits early and something else stays open. What changes is which mechanism does the committing, and how much room it leaves. The tension is sharpest wherever a dealer prebook or a retailer purchase order is the dominant signal, because there the gate is set by the customer rather than by the brand. The entries below stay on the two-channel commitment question only; for how each vertical structures its hierarchy, depth and lifecycle, RetailNorthstar covers merchandise hierarchy by vertical.

Seven ways a two-channel plan goes wrong

None of these is a mistake anyone makes on purpose. Each is a reasonable simplification that stops being reasonable once there are two channels drawing on one buy — a total that hides a distinction, a buffer bought twice, an average that describes neither side. They are grouped here because they share a failure mode: the plan still balances afterwards, which is exactly why none of them surfaces until the season is already committed.

Where a two-channel plan goes stale

Everything above assumes the plan knows the current order book, the current production position and the current DTC read at the same moment. In a spreadsheet it does not, and the reason is structural rather than careless: bookings live in the sales system, purchase orders live with production, DTC demand lives in the commerce platform, and the plan lives in a file that reconciles all three by hand, on whatever cadence someone has time for.

That gap is expensive here in a specific way. A cancelled booking frees units the DTC side does not know exist. A slipped delivery moves a wholesale ship window that the reserve was sized against. A DTC week running ahead of plan is a reorder signal that has to be acted on before the reorder cutoff, which is a date that has to be derived rather than looked up. When the order book, the production position and the demand read sit on one data model, the reserve, the channel allocation and the wholesale-versus-DTC open-to-buy position all move the same day the underlying facts do — and the scarcity rule agreed before the season is applied while it is still cheap to apply.

• ✓Plan three lines, never one: booked wholesale net of a stated cancellation haircut, DTC at its expected case with the range visible, and a single shared reserve.
• ✓Size the reserve from the width of the DTC range, then test whether pooling actually applies — it fails on different sizes, different colourways, and different dates.
• ✓A pooled reserve is cheaper than two buffers only to the extent the two channels do not miss in the same direction at the same time. Write that assumption next to the reserve line; pool the units, keep the open-to-buy dollars bucketed by channel.
• ✓Check every factory minimum against the combined line and record which channel each decision was made for; that column is how a wholesale-shaped DTC assortment gets caught.
• ✓Channel conflict is created in the plan, at the DTC markdown calendar, the launch date, the default home of the reserve, and the sell-through assumption behind the buy.
• ✓Ship the penalised units first, then allocate by contribution dollars adjusted for the probability of realising the price — not by which channel committed first.
• ✓Reconcile one buy to two margin statements in dollars. In the worked example, a blended margin rate sits within a point of both channels while contribution per unit is nearly double on one side.

• Sell-in & bookings calculator — booking attainment and fill rate →
• How to build a buy plan — the line-by-line buy this sits on top of →
• How to plan receipt flow — phasing intake across the season →
• How to calculate size curves — where the two channels diverge →
• Safety stock for seasonal assortments — how a buffer is sized before you ask whether it can be shared →
• How to plan a peak-concentrated season — reserve lead time and the last-receipt date →
• How to set open-to-buy — the budget the buy spends, bucketed by channel →
• Landed cost & margin calculator — the cost side of both margin stories →
• Buy plan template — vendor, delivery window, receipt month and channel per line →
• Planning for wholesale on RetailNorthstar →
• Planning for DTC on RetailNorthstar →

Frequently asked questions

How do you plan wholesale and DTC from a single production buy?

Size the buy as three separate lines rather than one demand number: the booked wholesale quantity net of a stated cancellation allowance, the DTC forecast at its expected case, and one shared reserve sized to the width of the DTC range. Keep the three visible in the plan so anyone can see how much of the buy is contractually protected. Then set the channel split as a decision made at receipt, not at commitment, because the split is the last thing that stays genuinely open.

Can you add booked wholesale orders to a DTC forecast?

Not without losing the information that matters. A booked wholesale quantity is a near-certainty carrying cancellation and chargeback risk; a DTC forecast is a distribution with a mean and a spread. Summing them produces a total whose error is made of two incompatible kinds of uncertainty, so it can no longer tell you how much of the buy is exposed. Plan them as separate lines, apply a stated haircut to the booked figure, and use the width of the DTC range — not the total — to size protection.

What is a shared reserve, and when does pooling not help?

A shared reserve is one pool of units that either channel can draw on, held instead of a safety buffer inside each channel plan. It is cheaper than two buffers to the extent that the two channels do not miss in the same direction at the same time — which is an assumption about the buy, not a property of it, so state it where the reserve is sized. It stops working entirely when the units are not interchangeable: when the channels need different sizes or widths, different colourways, or the units on different dates. Test all three before relying on the pool, and plan the portion that fails the test as channel-specific stock.

When units are scarce, should wholesale or DTC get them?

Units against a confirmed PO inside its ship window, where a short-ship triggers a chargeback, come off the top — that is the price of having taken the booking. Beyond those, allocate by contribution dollars per unit adjusted for the probability of selling at that price, not by which channel committed first. "Wholesale first because it is committed" is right contractually and is not automatically right commercially, because a DTC unit can carry more contribution on the same piece of inventory. The cheapest lever is renegotiating the short-ship with the account before the ship date, which converts a chargeback into a conversation.

How do you compare wholesale and DTC margin on the same unit?

Restate the unit twice. The wholesale line is IMU-driven: wholesale price less product cost, less allowances and chargebacks. The DTC line starts at realised retail AUR — after promotion, not at ticket — less the same product cost, less fulfilment, payment processing and a returns provision. Compare the contribution dollars, not the percentages. In the illustrative example on this page the two channels land within a point of each other on margin rate while contribution per unit is nearly double on one side, and every marginal-unit decision depends on the dollars. The dollars are not equally certain either: the wholesale figure is attached to a specific purchase order, while the DTC figure is an average across units that sold at full price and units that sold on promotion.

How do you stop DTC discounting from undercutting wholesale accounts?

Treat the DTC markdown calendar as a wholesale commitment and set it at prebook close rather than in season. Gate the first DTC promotion on wholesale sell-through reaching an agreed point instead of on the DTC receipt curve, and set the DTC launch date relative to the accounts’ floor-set date rather than to your own first receipt. Where MAP applies, the pricing floor is part of the arrangement the prebook was written under. Channel conflict of this kind is created by the calendar, not by the discount.