MFP, assortment plan and open-to-buy: three plans, one set of numbers
The merchandise financial plan, the assortment plan and the open-to-buy are not three separate plans. They are one set of numbers viewed at three different commitments: the MFP commits the money, the assortment plan commits the product, and the OTB tracks the timing of what is left to commit. Everything that goes wrong between them goes wrong at the joins — which number feeds which, at what level of the hierarchy, and what is allowed to happen when one of them moves after another has been locked.
This guide is deliberately not a fourth definition. Each artifact gets two sentences and a link to the page that defines it properly. What follows is the reconciliation: the three tie-points that hold the set together, a worked example broken on purpose, who commits what and when, how the tie-points behave across apparel and adjacent verticals, and the four ways teams conflate the three.
Definition and quick answer
One row each: decision, grain, cadence, owner, input, output
Two sentences of definition each, then straight to the part that matters. Full treatments live on their own pages: merchandise financial planning, assortment planning and open-to-buy planning. For where all three sit inside the wider process, see what is merchandise planning.
| Artifact | Decision it makes | Its grain | Its cadence | Who commits it | What it consumes | What it emits |
|---|---|---|---|---|---|---|
| Merchandise financial plan (MFP) | How many dollars of sales, inventory, receipts, markdowns and margin, in which period. | Hierarchy node by period — commonly class or department by month. | Built pre-season, locked at sign-off, reforecast on a fixed cadence. | Planning director, with finance owning the top line. | Company and channel targets, restated prior year, stock-to-sales or weeks-of-supply targets. | A receipt plan in cost and retail dollars by period. |
| Assortment plan | Which products those receipt dollars buy, how wide the line is and how deep each option goes. | Option (style-colour) by depth by size or shade distribution, landed against a node and a delivery month. | Built to line review and buy milestones. Locked at the buy commitment. | Buyer and merchandise planner jointly; the merchant signs the line. | The MFP receipt plan, the line plan, cost sheets and an average-unit-retail assumption. | A unit total and a cost total per node per delivery. |
| Open-to-buy (OTB) | Whether budget remains to commit, in which period, and how much. | The same node and period as the receipt plan, drilled through to purchase order and delivery date. | Continuous. It moves on every order placed, receipt landed, cancellation and reforecast. | Nobody commits the OTB — it is a balance. The buyer draws it down, the planner maintains it. | The receipt plan as an opening budget, less on-order, less received. | Remaining open dollars and units — the authority for the next purchase order. |
Read down the last two columns and the dependency is obvious: what the MFP emits is what the OTB consumes, and what the assortment plan emits is what draws the OTB down. Three artifacts, two joins, and one balance that exists only to make the second join visible while there is still time to act on it.
The receipt plan is the open-to-buy opening budget
The MFP does not hand the buy a sales number or a margin number. It hands over one line: planned receipts by period, in cost and retail dollars, at a named node. That line is not a guideline and it is not a target. It is the opening balance of the open-to-buy, and every purchase order written against that node and period draws it down.
An assortment plan whose units cost more than the receipt plan carries is not a plan, it is a request. That distinction sounds pedantic until you watch what happens without it. A line that exceeds the receipt plan by any amount has an unfunded remainder, and the remainder does not disappear — it gets funded by something. Usually by a later delivery that quietly slips, or by inventory that arrives and sits, or by a markdown that nobody planned. The money was always going to be spent; the only question was whether the decision to spend it was made by a person or by the calendar.
The arithmetic of the join is one division. A receipt plan is dollars; an assortment plan is units. The bridge between them is average unit cost. Divide the receipt plan by the assumed average cost and you have the unit count the assortment plan is allowed to contain. If that number has never been calculated, the two artifacts are not tied at all — they are two documents about the same season that happen to sit in the same folder.
The direction of the tie also settles a common argument about sequence. The receipt plan comes first, but the assortment plan is allowed to push back on it, because the assortment is where the achievable average cost and the achievable option count actually live. That exchange — a top-down dollar ceiling meeting a bottom-up unit reality — is exactly the reconciliation described in aligning top-down and bottom-up planning. What is not allowed is for the push-back to happen silently, after the buy, in the form of a receipt plan that everybody has stopped looking at.
The option-by-depth grid must roll to the MFP at the same hierarchy node
The assortment plan is a grid: options across, depth down, cost and retail attached. Its totals are a unit count and a dollar figure. Those two totals have to land on the MFP unit and dollar lines — and they have to land at the same node. Not a node above. Not a node below. The same one.
A mismatched planning level produces a reconciliation gap that is structural, not arithmetic, which is why nobody can ever close it. If the MFP is maintained at department and the assortment plan is built at class, the two totals will differ by whatever the other classes in that department are doing. Both numbers are correct. Neither is wrong. There is no error to find, so the weekly hunt for the error finds nothing, and the team learns to describe the gap as a quirk of the reporting rather than as a decision they have failed to make.
The same failure runs the other way. An MFP maintained at class while the buy is committed at subclass gives you a plan that is met in aggregate and missed in every subclass, with no mechanism to say which subclass caused it. The aggregate is comforting and useless: it reports that the department landed on plan while three of its subclasses were bought long and two were bought short.
The rule that survives contact with practice is to maintain all three artifacts at the node where money is actually committed, then roll upward for reporting. Rolling up is free and lossless. Pushing down is neither — it requires an allocation rule that nobody owns and that changes every time the mix changes. The finer node is more work to maintain, and that work is the price of being able to answer, in October, which part of the buy caused the variance.
One practical test: pick a node at random and ask three people to state its receipt plan in units. If the planner answers in dollars at department level, the buyer answers in units at class level, and the allocator answers in units at style level, the artifacts are not tied — they are three parallel systems that agree only by coincidence.
The open-to-buy is a running balance, so it is meant to disagree
The MFP and the assortment plan are locked artifacts. They are built, signed off, and then changed only under change control. The open-to-buy is not like that at all. It moves every time a purchase order is raised, every time a receipt lands, every time a delivery is cancelled and every time the plan is reforecast.
This means “the OTB disagrees with the plan” is usually a description of the system working correctly. Take the illustrative node used in the worked example below, where the figures are chosen to divide cleanly rather than drawn from any brand: a receipt plan of $160,000 at cost and an open-to-buy showing $20,000 remaining are not in conflict — the second is the first minus everything committed against it so far. Reading that as a data problem, and raising a ticket to make the two numbers match, is one of the most reliable ways to destroy the only forward-looking control in the process.
There is a subtler trap inside the same fact, and it is the one that costs real money. Remaining open dollars are not the same thing as spare dollars. The $20,000 still showing as open is the part of the receipt plan that has not yet been converted into a purchase order — and in a healthy process it is already spoken for by units sitting in the assortment plan waiting to be ordered. A buyer who reads remaining-open as headroom and commits it to something new has not found budget; they have deleted units from the plan without telling anyone which ones.
The corollary is that a healthy open-to-buy trends toward zero as a period closes, and that an OTB which never moves is not disciplined, it is disconnected. If the balance is identical on the fifteenth and the twenty-eighth, either nothing was ordered or orders are not being written back to it. Both are worth investigating; only one of them is good news. How this balance behaves period to period, and how to read it, is covered on OTB planning.
One node, one month, three views — then break it on purpose
One class, one channel, one month in the middle of a season. The same money appears three times.
Illustrative figures, chosen because they divide cleanly. Not benchmarks, and not drawn from any brand.
View one — the MFP line for the month
| Plan line | Illustrative value | Where it comes from |
|---|---|---|
| Planned sales | $400,000 retail | Phased from the season sales plan for this node. |
| Planned markdowns | $40,000 retail | Planned rate for the class, not a residual. |
| Shrink | $0 | Set to zero here only to keep the arithmetic readable. |
| BOP inventory | $900,000 retail | Closing balance carried in from the prior month. |
| Planned EOP inventory | $860,000 retail | Set from the weeks-of-supply target for the following month. |
| Derived receipt plan | $400,000 retail / $160,000 cost | EOP + Sales + Markdowns + Shrink − BOP = 860 + 400 + 40 + 0 − 900 = 400 retail. At a 60% initial markup, cost = 400 × 0.40 = 160. |
View two — the assortment plan for the same node and delivery
| Element | Illustrative value |
|---|---|
| Options (style-colours) landing this month | 20 |
| Average depth per option | 400 units |
| Total units | 20 × 400 = 8,000 |
| Average cost per unit | $20 |
| Total buy at cost | 8,000 × $20 = $160,000 |
| Average retail per unit | $50 |
| Total buy at retail | 8,000 × $50 = $400,000 |
The two views reconcile exactly: $160,000 of buy against $160,000 of receipt plan, $400,000 at retail against $400,000 of planned receipts, and an implied initial markup of ($50 − $20) ÷ $50 = 60%, which is the rate the MFP used to convert retail to cost. That agreement is the tie-point holding, and it is worth noticing that it holds on three numbers at once — dollars, units and markup rate. A plan that reconciles on dollars but not on units has an average-cost assumption hiding in it that nobody has stated.
View three — the open-to-buy, part-way through the month
| Balance line | Illustrative value | What it represents |
|---|---|---|
| Opening OTB for the month | $160,000 cost | This is the receipt plan. It is not an independent number. |
| Less on-order (purchase orders placed, not yet landed) | $110,000 | 5,500 units already committed to vendors. |
| Less receipts already landed | $30,000 | 1,500 units in the building. |
| Remaining open | $20,000 | 1,000 units of the assortment plan not yet converted to a purchase order. |
Now break it: two options added at line review
The merchant adds two style-colours at line review, planned at the same depth and cost as everything else: 2 × 400 × $20 = $16,000 at cost. The assortment plan is now 22 options, 8,800 units and $176,000 at cost, against a receipt plan of $160,000. The buy exceeds the plan by $16,000 at cost and $40,000 at retail.
Here is where the third tie-point earns its place. The open-to-buy still shows $20,000 remaining open, and that number looks like room. It is not room. That $20,000 is the 1,000 units of the original twenty options that have not yet been turned into purchase orders. Spending it on the two new options funds them by silently deleting 800 units of planned depth from styles nobody has agreed to cut — and the deletion will be discovered when a vendor asks why the order never came, or when the size run arrives short.
The over-plan is $16,000 whether or not the open-to-buy currently shows a positive balance. Remaining open answers “can I raise a purchase order today”. It does not answer “does the line fit the plan”. Those are different questions and only the second one is about the assortment.
The honest responses, and what each one costs
Cut two existing options
Buyer with the merchant, because it is a line decision.Drop two planned options of equivalent cost. The buy returns to 20 options, 8,000 units and $160,000.
Whatever the cut options were carrying stops being covered — an opening price point, a size-run anchor, a colour that a wholesale prebook was written against. The dollars reconcile instantly; the coverage gap shows up in week six.
Shave depth across the line
Planner with the buyer, because it changes unit productivity assumptions.Hold 22 options and cut average depth from 400 to 360 units. 22 × 360 × $20 = $158,400, which is $1,600 inside the receipt plan.
Every option loses 40 units. Size runs break earlier at the same sell-through, the reorder decision arrives sooner than the lead time supports, and the markdown risk moves from too much inventory to lost full-price sales.
Move the receipt to the following month
Planner, because it changes the MFP inventory line in two periods.Push the two new options to the next delivery. This month stays at $160,000; next month absorbs $16,000 of its own receipt plan.
The current month EOP falls, so the following month opens lighter and the weeks-of-supply target for that month has to be re-solved. The units land four weeks closer to the season exit, so their markdown exposure rises even though nothing about the buy changed.
Change the receipt plan itself
Planning director, under change control.Raise the month receipt plan from $160,000 to $176,000 at cost and re-solve EOP and margin from the identity.
Legitimate, but it is a different decision made by a different owner. EOP inventory rises by $40,000 at retail, the weeks-of-supply target is no longer met, and the margin plan must be re-derived. What is not legitimate is raising the OTB without re-solving the MFP, which produces a budget that no longer reconciles to any inventory position.
Every one of these is a real answer. None of them is free, and the point of writing down which one was used is that at month end the variance has a sentence attached to it rather than a shrug. The failure is not choosing wrongly — it is the fifth response, where the two options are added, nothing else changes, and the arithmetic is resolved four months later by a markdown.
Who commits each artifact, when, and what happens if one moves after another is locked
Two of the three are locked artifacts with owners and versions. The third is a balance with a maintainer and no lock at all. Confusing which is which is the root of most change-control arguments in planning.
Merchandise financial plan
Pre-season sign-off, before the buy opens.What locked means: The sales, inventory and margin lines stop moving on opinion. Actuals replace estimates on the reforecast cadence, and every reforecast re-solves receipts from the same identity rather than editing them by hand.
Change control: A change to a locked MFP line is a planning decision with a named owner and a dated version. If the receipt plan moves, the OTB opening budget moves with it in the same action — otherwise the two drift and the buyer is working against a number that no longer exists.
Assortment plan
At the buy milestone, after line review and costing.What locked means: Options, depth and delivery are fixed for the purposes of committing purchase orders. Late additions are not free: each one has to be paid for out of the same receipt plan.
Change control: An addition after lock requires a named source of funds — a cut, a depth reduction, a moved delivery, or an approved change to the receipt plan. Recording which of the four was used is the whole of change control here. Teams that skip that record cannot explain, at month end, why the buy exceeded the plan.
Open-to-buy
Never. It opens when the receipt plan is released and runs until the period closes.What locked means: Nothing. A frozen OTB is not a stricter control, it is a dead one — it stops reflecting on-order and receipts, which is the only information it carries.
Change control: The OTB has no change control of its own, because it holds no decisions of its own. Every movement in it is an echo of a decision taken in one of the other two artifacts. If an OTB number looks wrong, the fault is upstream: either a receipt plan that was edited without a version, or a purchase order that was placed without being written back.
The sequencing question that comes up in every planning calendar is what to do when the MFP is reforecast after the assortment plan is locked. A reforecast that lowers the receipt plan does not automatically unmake commitments already placed — it reduces the remaining open balance, which is exactly what should happen, and it may take that balance negative. A negative open-to-buy is not an error state. It is the correct representation of a node that is committed beyond its revised plan, and it is the signal to run the same four responses against the periods that are still open.
The same three tie-points, seven different places they break
The artifacts do not change across merchandise categories. What changes is the grain of the assortment plan, how freely depth can move, and how big a step a single receipt makes in the balance — and each of those decides which tie-point fails first.
Apparel
The assortment plan grain is style-colour by size curve, so depth is a distribution rather than a number, and the tie to the MFP has to survive that translation. A planner who fits the unit count to the receipt plan has not finished: the same units spread across a size curve that does not match demand will break in the middle sizes while the tails sit. The hierarchy tie is the more common failure — the MFP maintained at department, the buy committed at class, and a translation step between them that no system performs. Season exit dates are hard, so a moved receipt is a markdown decision with a later date on it.
Footwear
Depth is quantised. The practical unit of an assortment plan is a size run, not a unit, so the three responses to an over-plan are not equally available: shaving depth by ten percent across the line often means dropping a whole run rather than trimming units, and dropping a run at the top or bottom of the range removes a customer segment rather than some inventory. That makes the cut-elsewhere response more attractive and the shave-depth response more expensive than it looks in the arithmetic. Carryover and model-year continuity also mean part of the assortment plan is effectively pre-committed before the season MFP is built, so the receipt plan opens partly spoken for.
Accessories and bags
There is no size dimension, so an option costs very little to add in units and the assortment plan fits the receipt plan almost too easily. The failure mode inverts: instead of an assortment that breaches the dollar ceiling, you get option proliferation inside it — more style-colours, each shallower, each carrying its own setup, sample and minimum-order cost that the dollar plan does not see. The tie-point that protects against this is the unit line, not the dollar line: an assortment plan that rolls to the MFP dollars while blowing through the planned option count and average depth has met the letter of the plan and abandoned its intent. Hardware and component lead times also make a moved receipt harder to reverse than the calendar suggests.
Home and furniture
Unit costs are high and unit counts are low, so a single container or a single SKU can be a material share of a month receipt plan. The OTB therefore moves in large steps rather than smoothly, and a two-week arrival slip does not shade the balance, it relocates it to the following period. Planning at a node fine enough to see that is the whole game: a collection-level plan hides which finish or which case good caused the movement. The option dimension is usually collection by finish, and depth is constrained at the low end by floor-sample and display commitments that have to be bought whether or not the arithmetic wants them, so the shave-depth response runs out of room earlier than in soft goods.
Outdoor
A large share of the assortment is committed at prebook, at a trade show, against dealer orders taken before the season OTB opens. That inverts the normal sequence: the assortment plan is partly fixed before the MFP receipt plan is fully phased, so the reconciliation runs backwards — the plan has to be built around commitments already made rather than the commitments being made inside the plan. The season exit is weather-dependent and unforgiving, which makes the move-the-receipt response the most dangerous of the three: a delivery pushed four weeks in a season that can end early is not a deferred cost, it is a markdown that has not been booked yet. Technical carryover softens this at the margin, because spec-stable product can be sold across seasons.
Health and beauty
The size-curve equivalent is the shade or variant range, and it behaves the same way at the tie-point: the assortment plan must roll to the MFP unit line at the same node, but depth within the range is not freely adjustable because dropping the extreme shades removes the reason the range exists. Period-after-opening and expiry turn depth into a perishability decision, so buying deep to hit an open-to-buy balance is a way of converting budget into write-off. Replenished lines also make the OTB behave less like a seasonal drawdown and more like a rolling budget, which changes what a healthy remaining-open number looks like: near zero at period end is correct for a seasonal buy and a stockout risk for a replenished one.
Sporting goods
Two rhythms sit under one plan. Hardgoods move on model-year cycles with long, largely fixed commitments; softgoods and apparel move on a seasonal calendar with size runs and in-season chase. If one MFP node covers both, the receipt plan averages a fixed commitment with a live one, the OTB balance reads as healthy while the flexible half of it is already gone, and no reconciliation can recover which half caused a variance. The fix is structural rather than analytical: plan the two rhythms at separate nodes with their own receipt plans and their own open-to-buy, then consolidate upward. That is the same discipline as separating wholesale from direct channels, applied to product type rather than route to market.
Two patterns run through all seven. The less freely depth can move, the more the over-plan conversation has to happen at option count rather than at depth — that is footwear, home and furniture, and beauty. And the larger a single receipt is relative to the month, the finer the node has to be for the balance to mean anything, because a balance that moves in one big step tells you nothing about which decision moved it. How the option-by-depth grid is built in each case is covered on assortment planning.
Four ways the three get conflated
Each of these is one artifact being asked to do another artifact's job. None of them produces an error message, and all of them produce a set of documents that review cleanly.
The assortment plan doing the MFP job
A line built from the bottom up with no dollar ceiling above it. Every option is individually defensible, the total is nobody's decision, and the first time anyone sees the number is when purchase orders exceed what the business intended to own. The symptom is an assortment review that never says no to anything; the cause is that there is no receipt plan for it to say no against.
The MFP doing the assortment plan job
A financial plan that balances perfectly and cannot be executed, because nobody has checked that the dollars can be spent on real product at real cost and real depth. Using the illustrative figures from the worked example above, a receipt plan of $160,000 at a $20 average cost is 8,000 units; if the line as designed only has room for 5,000, the plan is not tight, it is wrong, and the gap will be closed by whatever the buyer improvises. The test is cheap: divide the receipt plan by the assumed average unit cost and ask whether that many units have somewhere to go.
The OTB treated as a plan
Frozen at the start of the period, reported monthly, and never moved on order activity. A frozen open-to-buy is a stale copy of the receipt plan, which the MFP already holds. Its entire value is that it moves — that it drops when a purchase order is placed and rises when one is cancelled. Freeze it and it stops being a control, while continuing to look like one on the report.
Three artifacts at three different nodes
The MFP at department, the assortment plan at class, the OTB at subclass. Each is internally correct and no two of them reconcile, so a permanent gap sits between them that nobody can close — because it is structural, not arithmetic. No amount of checking finds the error, since there is no error to find. The only fix is to pick the node the buy is committed at and maintain all three there, then roll up for reporting rather than planning at the rolled-up level.
The diagnostic for all four is the same and takes about ten minutes. Ask for the receipt plan for one node and one month in dollars, the assortment total for the same node and month in units and dollars, and the open-to-buy balance for the same node and month. If the three cannot be put on one page because they are held at different levels, the problem is structural and no reconciliation will fix it. If they can be put on one page and they disagree, the disagreement is now a specific, answerable question — which is the entire purpose of keeping three artifacts instead of one.
Key takeaways
MFP, assortment plan and OTB — common questions
Is open-to-buy part of merchandise financial planning?
Open-to-buy is an output of merchandise financial planning rather than a separate discipline. The MFP produces a receipt plan by period; the open-to-buy is that receipt plan expressed as a running balance, reduced by what has been ordered and what has already landed. So the OTB is downstream of the MFP and depends on it entirely — but it is maintained continuously rather than planned, which is why it is treated as its own artifact in day-to-day work.
Which plan comes first — MFP, assortment plan or open-to-buy?
The merchandise financial plan comes first, because it sets the dollar ceiling the other two work inside. The assortment plan comes second and spends those dollars as options and depth. The open-to-buy opens last, at the point the receipt plan is released to buying, and then runs continuously for the rest of the period. In practice the first two overlap: the assortment plan sends back an average unit cost and an achievable option count that the MFP has to accept, and that exchange is the reconciliation between top-down and bottom-up planning.
Can you have an assortment plan without an MFP?
You can build one, but it will not be a plan in the sense that matters, because nothing constrains it. Without a receipt plan above it, an assortment plan has no dollar ceiling, no test for whether it is affordable, and no mechanism for saying no to an addition. It becomes a list of what the team would like to buy. The arrangement holds together only while one person carries the total in their head and no one else can commit money; it breaks at the point where a second person can raise a purchase order, because there is then no shared number for the two of them to be constrained by.
What is the difference between a receipt plan and an open-to-buy?
The receipt plan is a planned quantity of inventory landing in a period, stated in cost and retail dollars and fixed at plan sign-off. The open-to-buy is that same number as a live balance: receipt plan, less purchase orders already placed, less receipts already landed, equals remaining open. The receipt plan answers how much should arrive; the open-to-buy answers how much is still available to commit right now. They are the same money at two different moments, which is why they should never be maintained as two independent numbers.
At what level of the hierarchy should each plan sit?
All three should sit at the level the buy is committed at — the node a purchase order is written against and a receipt is released to. Whether that node is labelled class or subclass matters far less than that it is the same node in all three artifacts. The MFP can be reported at department and above, but it should be maintained where receipts are released, because that is where the assortment plan rolls to and where the open-to-buy is drawn down. Maintaining the three at different levels produces a permanent reconciliation gap: each artifact is internally correct, no two of them agree, and the difference cannot be traced because it is a structural mismatch rather than an error.
What do you do when the assortment plan does not fit the open-to-buy?
There are four honest responses and one dishonest one. Cut options elsewhere in the line; shave depth across the line; move the delivery into a later period; or change the receipt plan itself under change control, re-solving ending inventory and margin from the inventory identity. The dishonest response is raising the open-to-buy without changing the receipt plan behind it, which produces a spend authority that reconciles to no inventory position. Each of the four carries a different cost — lost coverage, broken size runs, later markdown exposure, or a changed inventory and margin plan — and the discipline is recording which one was chosen and why.
One set of numbers, in one place.
In RetailNorthstar the financial plan, the assortment and the open-to-buy sit on one data model at one node — add two options at line review and the receipt plan, the balance and the margin projection all answer in the same action. Thirty minutes shows you what that removes from month end.
Connected merchandise planning — live in weeks, not quarters.