What is product mix optimization?

Product mix optimization is the discipline of deciding how a fixed buy budget is distributed across categories, price tiers and product types — not which individual products to carry, which is assortment planning. It operates one level up, on proportions rather than choices, and its output is a set of weightings: what share of the buy each part of the range should receive.

What is the difference between mix effect and rate effect?

Rate effect is a change in blended margin caused by margins themselves moving — a cost increase, a markdown, a price change. Mix effect is a change in blended margin caused purely by the proportions shifting, with every individual margin unchanged. They are routinely confused, and the confusion is expensive: a blended margin that fell because a low-margin category grew is a completely different problem from one that fell because costs rose, and the two have opposite remedies.

Why not simply buy more of the highest-margin category?

Because margin rate is only one of three terms. What a category contributes is its margin rate multiplied by its sell-through multiplied by the volume it can actually absorb. A 70% margin category that sells through at 55% returns less per dollar committed than a 55% margin category selling through at 85%. Optimising on rate alone reliably overweights categories that look profitable per unit and underperform per dollar.

What constraints make an optimal mix unbuyable?

Minimum order quantities, which set a floor on how small a category's buy can be; fixture and floor-space capacity, which caps how much of a category can actually be presented; vendor consolidation, where moving volume out of a category weakens terms across everything sourced from that vendor; and range coherence, where a category exists to make the assortment make sense to a customer rather than to earn its own return.

How often should the mix be reviewed?

The weightings should be set pre-season alongside the option count and revisited at the same cadence as the merchandise financial plan. Reviewing mix in-season is usually too late to act on for the current season — the buy is committed — but it is the input that most improves the next one, which is why the post-season hindsight is where mix work earns its keep.

Does product mix optimization apply outside apparel?

The arithmetic is universal but the axes change. Apparel mixes across category, price tier and fabrication. Beauty mixes across franchise and format, where a gift set is a mix decision that consumes several components. Home and furniture mixes across collection and finish, constrained by container economics rather than floor space. The method transfers; the dimensions do not.

Product Mix Optimization: Where Assortment Margin Actually Comes From

Product mix optimization decides how buy dollars are distributed across categories and price tiers, not which products to carry. This guide separates mix effect from rate effect, shows why the highest-margin category is rarely the right place to add depth, and sets out the constraints that make a theoretically optimal mix unbuyable.

What product mix optimization decides

Product mix optimization is the discipline of deciding how a fixed buy budget is distributed across the parts of a range — categories, price tiers, fabrications, product types. It does not decide which products to carry. That is assortment planning, and it happens one level down.

The distinction is worth holding precisely, because the two get conflated and the conflation hides the decision. Assortment planning asks what should be in the range. Mix optimization asks what share of the money each part of the range should get. A brand can have an excellent assortment inside a badly weighted mix — every individual choice defensible, and forty per cent of the budget sitting in a category that cannot return it.

Mix is set at the level where the merchandise hierarchy has a financial rollup: division, department, class. It is expressed as proportions, and the proportions are the deliverable.

Mix effect and rate effect

The single most useful thing this discipline offers is a decomposition. Blended margin moves for two structurally different reasons, and a summary number cannot tell them apart.

Rate effect is margin moving because margins moved. A cost increase, a markdown taken, a price adjustment, a change in promotional depth. The mix is unchanged; the individual rates are different.

Mix effect is margin moving because the proportions moved. Every individual margin is identical to last year; a lower-margin category simply grew relative to the others, and the blended number fell.

A worked illustration, with round numbers chosen to make the arithmetic legible:

Category	LY share of buy	TY share of buy	Margin rate (unchanged)
Outerwear	30%	20%	62%
Knitwear	25%	25%	58%
Jersey basics	45%	55%	48%
Blended	54.7%	53.3%	—

Blended margin fell 1.4 points — weighting by share of buy — and not one category's margin changed. The entire movement is mix. A margin review that stops at the blended number will go looking for a cost problem that does not exist, and the actual decision — that the buy drifted ten points toward the lowest-margin category — never surfaces.

This drift is usually nobody's decision. It accumulates from individually reasonable calls: basics sell reliably, so the reorder goes there; outerwear is risky, so the buy is trimmed. Each choice is sound in isolation and the sum is a structural margin decline nobody made deliberately.

Why the highest-margin category is rarely the answer

The intuitive move — shift dollars toward the highest margin rate — fails often enough to be worth understanding properly. Margin rate is one of three terms.

What a category returns per dollar committed is approximately:

Contribution per dollar committed ≈ margin rate × sell-through rate × absorbable volume.

This is a way of thinking, not a canonical formula — the third term is a judgement rather than a measurement. What it does reliably is stop a margin ranking from being mistaken for a priority list.

Work it through. A premium category at a 70% margin that sells through at 55% before markdown returns meaningfully less per dollar than a core category at 55% margin selling through at 85%. The first looks twice as profitable in a margin report and returns less money.

The third term is the one most often ignored. Absorbable volume is the amount a category can take before the incremental unit stops selling — a function of customer demand, floor space, and how many options the category can support before they cannibalise each other. Doubling the buy in a category with a narrow customer base does not double its sales; it doubles its markdown.

This is why mix work has to be done against sell-through and capacity data, not against a margin ranking. A margin ranking is a list of what is profitable per unit. Mix needs to know what is profitable per dollar of budget.

The constraints that make optimal mixes unbuyable

An optimiser given margin, sell-through and volume will produce weightings that cannot be executed. The binding constraints are operational:

Minimum order quantities. A category's buy cannot go below the smallest producible order. A mix that says a category should take 3% of the budget may be saying it should be dropped, because 3% is under one MOQ. That is a legitimate conclusion — but it is a range decision, not a weighting adjustment, and it should be surfaced as one.

Fixture and floor capacity. A category weighted at 40% of the buy needs somewhere to be presented. Physical capacity caps the mix regardless of what the arithmetic prefers, and a buy that exceeds presentable capacity converts directly into stockroom inventory and markdown.

Vendor consolidation. Volume moved out of a category is often volume moved out of a vendor, and terms are negotiated at vendor level. Trimming a marginal category can worsen costs across everything else that vendor makes — a second-order effect the category-level arithmetic cannot see.

Range coherence. Some categories exist to make the assortment legible rather than to earn their own return. A brand known for outerwear that reduces outerwear to its standalone contribution has optimised itself out of the reason customers arrive. This constraint is real and is routinely dismissed as sentiment; it is not, but it does need to be argued explicitly rather than assumed.

The practical form is to run the arithmetic first, unconstrained, and then apply the constraints one at a time with the cost of each written down. That produces something more useful than an optimal answer: a clear account of what the constraints are costing, which is the input to whether they should be relaxed.

Doing the work

A workable sequence, run pre-season alongside the option count:

Decompose last season into mix and rate. Establish how much of the margin movement was proportion and how much was rate. Do this before proposing anything.
Build contribution per dollar by category, using margin rate, realised sell-through and an honest view of absorbable volume — not planned sell-through, which is the number that was already wrong.
Propose weightings unconstrained, and note where they differ sharply from last year. Large moves need an explanation that is not "the model said so".
Apply the operational constraints in order, recording what each one costs against the unconstrained answer.
Reconcile to the buy budget and the option count. Mix, option count and depth are one decision expressed three ways — changing the weighting without revisiting the option count usually just makes the same number of options thinner.
Write down what you expect, so the post-season hindsight has something to compare against. Mix decisions are only learnable in retrospect, and only if the expectation was recorded.

Across categories beyond apparel

The arithmetic transfers unchanged; the axes do not. Apparel mixes across category, price tier and fabrication. Beauty mixes across franchise and format, with the complication that a gift set consumes components that are themselves SKUs — a mix decision in one place consumes budget in another. Home and furniture mixes across collection and finish, constrained by container economics rather than floor space, so the MOQ constraint binds much harder. Footwear mixes across model and construction, with the size run acting as a multiplier on every depth decision.

In each case the method is the same: proportions, contribution per dollar, constraints applied explicitly. What changes is which dimensions the proportions run across — which is exactly what the merchandise hierarchy defines for each category.