The cost of switching origin: when are two bulk wines genuinely interchangeable?

An illustrative situation. An offer arrives for a listing that is already working. Same type of wine, better price per litre, analytical results within specification. The saving per litre is the first figure. Its effect on the programme calls for a wider comparison.

Are we substituting the wine used in the agreed product, or changing the product the client buys?

These are different things, and they are best resolved in that order. Before comparing figures, you have to know whether you are comparing two suppliers, two origins or two commercial products.

Worth saying at the outset: switching origin can be a good decision. Sometimes it frees up margin, sometimes it solves a supply problem. What this piece proposes is not a defence of any origin, but a way of separating the questions so the decision rests on the right comparison.

The commercial brief governs the wine

Two requests that look alike can carry different constraints.

Supplying a red whose positioning accommodates several origins is a question of product and price. Replenishing a listing built as Argentine Malbec carries an additional constraint: the origin is part of what the account bought.

Where origin is built into the proposition, substituting it changes the proposition even when the liquid is similar. How much it weighs depends on the account. There is evidence that the association exists among professionals. A 2020 study worked with 201 wine trade professionals from 44 countries, recruited at ProWein 2018, and used word association to compare representations of wines from Argentina, France, Brazil and Switzerland. Malbec emerged as a prominent association for Argentina.

SourceRodrigues and colleagues, Food Research International 137, 109480, 2020

Worth reading for what it is. It documents mental associations in a self-selected sample at a professional fair — not a comparative tasting, not willingness to pay, not acceptance of a substitution, not sales. But it supports one practical conclusion: where a product carries a recognisable origin association, maintaining the sensory profile may not be enough to preserve the commercial proposition.

From statistical reference to executable comparison

Bulk moves volume, and the figure gets quoted to frame any price conversation. It is good for framing, not for settling.

The OIV reports that in 2025, wine in containers over ten litres accounted for 34.0% of volume and 7.3% of value in world exports, at an average unit value of €0.75 per litre. Against 2024, volume fell 3.8% and value 5.3%.

SourceOIV, State of the World Wine Sector in 2025, May 2026, page 18, table 4

That average pools origins, varieties, quality tiers, destinations and transactions with different specifications and commercial terms. It is not an executable offer, nor a price comparable with a sample on the table. And the aggregate decline does not mean every segment fell.

The same applies one step down. Public summaries of offers by variety and country — the kind market monitors publish from broker data — show a trend, not a quotation. They are historical offer references rather than verified transactions, and they rarely allow each observation to be normalised for specification, quantity and delivery terms.

A national average is not a substitute for a lot. Between the two sit a specification, an available volume and a date.

Three levels to resolve, in order

Two filters come before price, and they work best applied in sequence.

• Requirements that a lower price cannot override. Required country or geographical indication, declared variety, required certifications, destination compliance and written commitments to the account. If the candidate does not meet them, it is not a substitute: it is a modified product, and it needs the corresponding approvals before entering the comparison.
• Equivalence that has to be verified. Sensory profile and analytical range, the stability the journey and packaging require, treatment still pending at destination, available volume of that profile and the dispatch window. The availability of a wine category in a country does not establish the availability of the required lot.
• Economic differences that have to be measured. Wine, logistics, yield, duties, preparation at destination, approval work, materials, additional inventory and financing. Both alternatives have to be compared to the same operational and commercial point.

Comparing an already-approved supplier against a candidate without factoring in their outstanding tasks can overstate the saving.

What can change without anyone anticipating it

Label review and customs eligibility can also change the scope of the transition.

On labelling, the TTB specifies that changing a wine's appellation of origin requires a new COLA. Changing supplier does not in itself change the declared origin; whether a new COLA is needed has to be reviewed against the specific changes to the label and to the responsible operators.

SourceTTB, Allowable Changes Sample Label Generator

The OIV international labelling standard ties the country indication to grapes harvested and vinified there, and treats vinification or blending across countries separately. It is an international standard, not automatic clearance in every jurisdiction. And the place of packaging should not be confused with the origin of the wine.

SourceOIV, international standard for the labelling of wines, section 2.5

The second check is customs, and it gets confused with the first. A non-preferential certificate of origin is not the same as the proof required to claim a tariff preference under a trade agreement: the rules and documentation depend on the agreement and the market. Routing through a port does not change origin, and not every wine from a country pays the same.

SourceWine Australia, Export documentation, Proof of Origin and Rules of Origin

A difference in strength can reverse the comparison

Since 1 February 2026, the UK applies Alcohol Duty of £30.62 per litre of pure alcohol to wines between 8.5% and 22% ABV. The duty is calculated on volume and strength, and becomes due when the product is released for consumption.

SourceHMRC, Alcohol Duty rates, in force from 1 February 2026

Between a wine at 12.5% and one at 13.0%, the difference is £0.1531 per litre of wine, or £0.114825 per 750 ml bottle, before VAT and settlement rounding.

30.62 × 0.005 = £0.1531 per litre of wine

An alternative £0.10/l cheaper but half a point stronger ends up £0.0531/l more expensive, counting that duty alone.

This is not a penalty on nationality: at the same strength the difference disappears. It is a specification change with a tax consequence, and it is worth catching before signing. Nor is there a universal scale in half-point steps: this is an example, not a rule.

Alcohol duty and customs tariff are separate concepts, calculated separately. We covered the first in an earlier Journal piece.

How much better the offer has to be

The question is not how much is saved per litre, but how much volume it takes to recover what switching costs.

Three terms. Q is the volume in saleable litres within the same horizon. The net recurring saving per litre, d, compares avoidable future costs to the same point, including any difference in logistics, yield, duty and preparation. F is the additional transition cost that can still be avoided.

Incremental benefit = Q × d − F · Break-even = F ÷ d

Break-even only applies where d is positive: if the net recurring saving is zero or negative, raising volume does not recover the transition by that route. An example with hypothetical figures, attributed to no country and no real transaction. Product technically suitable, same acceptance and same sales. A purchase saving of €0.12 per litre, a recurring additional cost of €0.04 to reach the same point, and €6,500 of transition: €2,500 in additional trials, €2,000 in adaptation and new print runs, €2,000 in storage during the changeover. The net saving is €0.08 per litre.

Arithmetic assumptions, not market rates or regulatory costs. To isolate the effect of volume, the example holds transition costs constant; in a real operation they have to be recalculated if storage, lead times or required trials change. The euro figures should not be combined with the sterling box above as though they formed a single transaction.

Break-even sits at 81,250 litres. At an actual volume of 60,000 saleable litres, the same purchase difference does not recover the transition.

The result depends on the actual volume switched, not on the annual projection alone. Best not to multiply the saving by a projected volume if the opportunity is a single lot, or if there is no basis for assuming it repeats. Sensitivity matters as much as the result. At 120,000 litres and the same €6,500 of transition, the benefit changes sign according to the net saving.

Incremental benefit across 120,000 litres, with €6,500 of transition. The vertical line marks the point where the switch stops costing and starts returning.

Enabling an alternative is not the same as activating it

Break-even is calculated on the volume actually substituted. In practice there is an intermediate step worth separating, because it carries a different cost and a different logic.

A buyer can approve a second origin as contingency without moving the programme to it. That separates two kinds of cost from the volume that would transfer.

• Cost of enabling. The outlay required to hold a technically approved alternative: trials, documentation, samples and approvals.
• Cost of activating. What arises only if the switch goes ahead: material adaptation, transition inventory, label adjustments.
• Volume that would transfer. How much of the programme would actually move to the new origin.

Separating them changes the assessment. Enabling can be justified on security of supply even if it is never activated, and in that case break-even stops being the question: what is gained is operational flexibility, not a saving.

Approving an alternative can reduce the work needed to react to a contingency. Its usefulness also depends on the volume, the terms and the delivery capacity still being available.

Four ways the model misleads

The calculation is simple. The hard part is not loading it with things that do not belong.

• Brand investment already made is a sunk cost. It does not get charged to the switch. The demand that investment helped build may have future effects, but that needs separate evidence.
• Material already purchased is not counted twice. Separate past outlay, new replacement cost and recoverable value that is lost.
• Importer margin and retailer margin do not mix. Define who receives the saving and who pays for the transition.
• Profitability and cash are not the same. More inventory may require financing even where it is not a loss of the same amount.

And if the substitution changes selling price or demand, the model has to be extended: compare total contribution across both programmes rather than artificially holding volume constant.

What a tasting proves, and what it does not

Technical validation and commercial validation answer different questions.

The AWRI recommends tasting blind, with comparable presentation, randomised order and judgements recorded before they are discussed, and draws a clear line between detecting a difference and preferring a sample. Its triangle test asks which of three samples is the odd one out, two of them being the same wine.

SourceAWRI, Practical sensory evaluation considerations

SourceAWRI, Triangle

Failing to detect a difference in one trial does not demonstrate identity, let alone commercial equivalence. Panel size and design have to suit the purpose of the test, which is a methodological decision rather than a negotiation.

A technical committee can approve a sample that meets the brief. What the buyer needs to know is what they are accepting.

Which is why three assessments are worth separating: the analytical one, conformity with the agreed profile, and the commercial one, carried out with the real product information, label included.

Five possible sourcing decisions

A well-made comparison does not always end in a switch. It ends in one of these five decisions.

• Viable switch. The brief accommodates the new origin, product and documentation comply, the required volume can be delivered within the agreed window, and the advantage survives sensitivity on volume and cost.
• Limited trial. Compatibility is likely but acceptance or performance still need testing. Define scope, exit criteria and the cost of the trial, without committing to continuity that has not been assessed.
• New product. The alternative has potential but changes an essential characteristic. Assess it as a launch, not as a replacement.
• Keep the origin, review another variable. The origin is necessary for the brief. Work remains possible on supplier, specification or terms.
• Switch for resilience. Can be justified without a direct saving where it protects critical supply. State the objective and compare the cost of contingency alternatives.

• At Corbeau, projects are shaped around the target customer and market, and we work across different levels and styles of bulk Malbec. When comparing samples we assess alcohol, body, fruit intensity, and the presence and type of oak. Setting out the brief in those terms before discussing alternatives is what later makes it possible to tell whether an offer is comparable or is something else.