What is a moving cost base?

A moving cost base is one where the input cost of goods changes materially between the point an item is costed and the point it is produced or sold, for reasons outside the brand's control. Precious metals and stones are the clearest case, but leather, cotton, and ocean freight behave the same way. The planning significance is that initial markup agreed at costing is an assumption with a shelf life rather than a fact, and any margin plan built on it inherits that uncertainty.

Why does a margin bridge find the problem too late?

Because a margin bridge is a reconciliation, and reconciliations run after the period they describe. By the time a cost move appears as a variance line, the goods have been made at that cost and often sold at the original retail. Every response that could have changed the outcome — repricing, respecifying, shifting mix — needed to happen while the season was still open. The bridge explains the loss accurately and cannot prevent it.

What are the levers when input costs rise mid-plan?

Four, and they expire in sequence. Change the specification — metal weight, stone size, plating thickness — which is available only before production is committed. Shift mix toward items with lower input-cost exposure, which is available while the buy is still open. Adjust retail, which stays available longer but is bounded by price architecture and the competitive set. And accept the compression knowingly, funding it from elsewhere in the plan, which is always available and is the only one left if the others were missed.

How do you carry cost as a planning input rather than a fixed attribute?

By separating the input-cost assumption from the item record, so that a change to the assumption re-prices everything that depends on it rather than requiring each item to be re-costed by hand. In practice that means the metal or material cost is a named parameter of the plan, items reference it with their own content weight, and initial and maintained markup are derived rather than stored. The test is simple: if changing one assumption requires editing more than one place, the cost is an attribute, not an input.

What is a re-plan trigger and how should one be set?

A re-plan trigger is a pre-agreed threshold of input-cost movement that automatically prompts a margin re-plan, rather than leaving the decision to whoever notices. It should be set as a percentage move in the input cost weighted by that input's share of total cost of goods, so that a large move in a minor input does not trigger while a modest move in a dominant one does. Setting it in advance matters more than setting it precisely, because the failure mode is not a wrong threshold but no threshold at all.

Planning Margin on a Moving Cost Base

Q: Is managing a moving cost base the same as hedging?

No, and conflating them is a common error. Hedging buys financial protection against a price move and is a treasury function with its own instruments, counterparties and accounting treatment. Planning against a moving cost base is about seeing the merchandising consequence early enough that a merchandising response is still available. A brand can do either, both, or neither, and a planning platform that claimed to hedge would be overreaching.

Most margin planning assumes cost stops moving once an item is costed. When it does not — metals, stones, leather, cotton, freight — the plan finds out in the margin bridge, after every response has expired. This guide covers carrying cost as a re-runnable input, the four levers and their expiry dates, setting re-plan triggers, and where planning stops and treasury begins.

The assumption nobody states

Almost every margin planning process contains an unstated assumption: that cost stops moving once an item has been costed. Initial markup is calculated at item creation, the plan is built on it, and thereafter margin is something to be monitored — tracked against plan, explained in the maintained markup bridge, and reported.

For a lot of retail that assumption is close enough to true. For any category where the input cost trades on a market of its own, it is not. Precious metals and stones are the clearest case and the reason this guide sits in the jewelry planning cluster, but the same structure applies to leather, to cotton, and — over the last several years, sharply — to ocean freight. If your cost of goods contains a meaningful share of something whose price is set somewhere you do not control, this problem is yours too.

Why the margin bridge is the wrong instrument

The margin bridge is an excellent tool doing a job it cannot do.

A bridge is a reconciliation. It runs after the period it describes, decomposes the gap between planned and achieved margin, and attributes it to causes: mix, markdown, cost, shrink. It is genuinely valuable for learning, and it is completely useless for prevention, because by the time a cost move appears as a variance line the goods have been manufactured at that cost and frequently sold at the original retail.

The window in which something could have been done closed months earlier. The bridge is a correct explanation of an outcome that was already fixed — which is why brands with excellent margin reporting still lose margin to cost moves year after year, and why the fix is not better reporting.

The four levers, and when each expires

When an input cost rises mid-plan, there are exactly four responses. What makes this actionable is that they expire in a definite order, so knowing when you found out determines which you still have.

1. Respecify. Change the metal weight, the stone size, the plating thickness, the material grade. This is the most powerful lever because it addresses the cost directly rather than passing it on — and it is the first to expire, because it is only available before production is committed and often before final samples are approved.

2. Shift mix. Move planned depth toward items with lower exposure to the input that moved. Available while the buy is still open, which is a meaningfully longer window than respecification. This is frequently the fastest real lever, because it requires no product change and no price change — only a different distribution of the same budget.

3. Reprice. Adjust retail. Available longest, and bounded by things outside the plan: price architecture, the competitive set, and how much of a rise a customer will absorb before the item stops selling. Worth noting that repricing protects percentage margin while potentially damaging unit velocity, so it is not a free lever even when available.

4. Accept. Take the compression knowingly and fund it from elsewhere. Always available, and it is what remains if the first three were missed. It is a legitimate decision when made deliberately — and it is what happens by default when the cost move is discovered in the bridge.

The entire argument for treating cost as dynamic is that levers one and two exist in month two and do not exist in month eight. Target costing is the natural instrument for lever one, because it works backwards from the margin the plan requires to the specification the product can afford.

Carrying cost as an input, not an attribute

The mechanism is unglamorous. Separate the input-cost assumption from the item record.

In most systems, an item carries a cost number. When the metal price moves, someone re-costs the items — which means the change is manual, partial, and slow, so it happens rarely and usually only for the largest items. The plan therefore contains a mixture of current and stale costs, and nobody can say which is which.

The alternative is that the input cost is a named parameter of the plan. Items reference it along with their own content weight, and initial and maintained markup are derived rather than stored. Changing the assumption re-prices everything that depends on it in one operation, which means it can be done often, completely, and as a what-if rather than as a commitment.

There is a simple test for which of the two you have: if changing one assumption requires editing more than one place, cost is an attribute rather than an input. Most spreadsheet processes fail this test, and so do a surprising number of planning platforms.

The immediate payoff is the ability to ask the question that matters — if gold moves by this much, what happens to this collection's margin and which levers are still open? — as an analysis rather than as a project.

RetailNorthstar has no jewelry or watch customers today — apparel is the flagship vertical and that is where its track record is. What it offers here is cost carried as a re-runnable planning input rather than a fixed item attribute, so a change to a material assumption re-prices initial and maintained markup across everything that depends on it while the season is still open. Commodity procurement, hedging instruments, and their accounting treatment sit outside the platform entirely.

Setting re-plan triggers before you need them

Most brands re-plan margin when someone senior notices a headline. That is a poor trigger — it fires late, it fires inconsistently, and it fires on salience rather than on exposure.

A re-plan trigger is a threshold agreed in advance: a percentage move in an input cost, weighted by that input's share of cost of goods, that automatically prompts a margin re-plan. The weighting is what makes it useful. A large move in a minor input should not trigger; a modest move in a dominant one should. A brand whose gold content is the majority of its cost of goods needs a tighter threshold than one where it is a small fraction, and the same threshold applied to both is wrong for at least one of them.

Two practical notes. Having a threshold matters more than having the right one — the failure mode in the wild is not a badly calibrated trigger but no trigger at all. And the trigger should name who re-plans and by when, because a threshold that fires into an unassigned queue is decoration. This is the same lesson that applies to any control: a check nobody owns and a check that does not exist behave identically.

The mix lever deserves more attention than it gets

Of the four levers, lever two is the most under-used relative to its power, and it is worth a paragraph of its own.

Shifting mix requires no product change, no price change, no customer-facing communication, and no renegotiation with anybody. It re-allocates planned depth within an existing assortment toward items whose margin is less exposed to the input that moved. In a collection with meaningful variation in material content — which most jewelry collections have — the aggregate margin effect of a mix shift can be substantial while every individual item's plan changes only modestly.

The reason it is under-used is that it requires the exposure to be visible at item level in the same view as the depth decision. If material content lives in a costing sheet and depth lives in an assortment plan, nobody can see which items to shift toward without building a third file. That is a data-model problem rather than an analytical one, and it is exactly the kind of gap that a connected plan closes.

Where planning stops and treasury begins

Worth stating clearly, because these get conflated in vendor conversations and in internal ones.

Hedging buys financial protection against a price move. It involves instruments, counterparties, margin requirements, and accounting treatment, and it belongs to treasury or finance. A brand may hedge its metal exposure, and if it does, the hedged cost becomes the input the plan uses.

Planning against a moving cost base is about seeing the merchandising consequence early enough that a merchandising response remains available. It does not remove the exposure; it preserves the options.

The two are complements, not substitutes, and a brand can sensibly do either, both, or neither. What is not sensible is assuming that because finance is "handling" commodity exposure, the merchandising consequence is handled too — hedging fixes a cost, it does not tell a merchandiser which items to shift depth toward.

The connected version

In a spreadsheet process, the metal cost lives with finance, the item specifications live with product development, the depth plan lives with merchandising, and the retail architecture lives with pricing. A cost move requires all four to be reconciled before anyone can evaluate a lever, which is why the evaluation usually does not happen until the bridge forces it.

In a connected model, the input cost is a parameter of the same plan that carries item specification and planned depth, so margin planning re-runs across the assortment plan in one operation, and the mix lever is visible in the same view as the exposure it would address. The question moves from a quarterly reconstruction to something a planner can ask on a Tuesday.

For the wider category picture — size runs, capital as the depth constraint, ownership status, and the serialisation boundary — see the jewelry and watch planning guide and the jewelry and watch industry page.

See how RetailNorthstar re-prices a margin plan from one changed assumption, while the levers are still open.

Book a Demo →

Related resources

Merchandise Planning for Jewelry & Watch Brands — The pillar this guide sits under
Planning the Gifting Calendar — Where occasion mix meets margin exposure
Jewelry & Watch Brands — RetailNorthstar — Platform fit for jewelry planning teams
Margin Planning — Margin as a planning decision rather than a reported outcome
Target Costing Formula — Working backwards from required margin to specification
Initial Markup — Glossary — The number a cost move erodes
Maintained Markup — Glossary — Where the erosion eventually shows up
Margin Optimization — Glossary — The wider discipline this sits inside