Forward buying by the trade
Forward buying is the trade purchasing ahead of a price increase or during a promotional allowance, holding stock, and selling it later at a better margin. It is rational behaviour and it distorts almost every series a brand looks at.
Reading a period that contains forward buying without recognising it produces conclusions about demand from a phenomenon about purchasing.
How does it show up in shipments?
As a spike ahead of a known price change or during an allowance window, followed by a trough while the bought stock is worked through. The pattern is distinctive once you look for it and invisible if you do not.
The trough is the part that causes trouble, because it looks like a demand collapse and arrives in the next reporting period.
How does it show up in scan data?
Mostly it does not, which is the useful part. Shoppers did not change their behaviour; the trade did. Scan data therefore continues roughly flat while shipments oscillate, and the gap between the two series is the inventory moving through the chain.
That divergence is diagnostic. A shipment spike with no scan movement is a trade event, not a consumer one.
What does it cost a brand?
Margin, mainly. Stock sold at an allowance price and then sold on at full price transfers margin to the trade without generating incremental consumption. The promotional investment bought inventory movement rather than shopper demand.
It also distorts forecasting and production, since a demand signal inflated by forward buying leads to over-production followed by a gap.
How is it detected?
By comparing shipment and scan series over a window long enough to contain both the spike and the trough. Pipeline inventory can be estimated as the cumulative difference, and a rising cumulative difference is stock building somewhere in the chain.
Where distributor withdrawal data is available, it locates the inventory directly and removes the guesswork.
What limits it?
Allowance structures that are tied to sales rather than to purchases, quantity caps, and shorter windows. Each reduces the incentive, and each is a commercial negotiation rather than an analytical fix.
The analysis supports the negotiation by quantifying the cost, which is the practical contribution a data function can make here.
How should periods containing it be reported?
With the effect identified and quantified rather than absorbed. A quarter that looks strong because of forward buying should say so, because the following quarter will look weak for the same reason and somebody will want an explanation.
Frequently asked questions
Is forward buying the same as pantry loading?
No. Pantry loading is shoppers buying ahead; forward buying is the trade buying ahead. They look similar in aggregate and they occur at different points in the chain, and the scan-versus-shipment comparison distinguishes them cleanly.
Can it be eliminated?
Not entirely, and it can be reduced through allowance design. Structuring incentives around depletion rather than purchase is the standard approach and it requires visibility of depletion, which is an argument for the data.
How do I estimate pipeline inventory?
As cumulative shipments minus cumulative measured depletions, adjusted for the coverage difference between the two universes. It is an estimate rather than a count, and its direction is usually reliable even when its level is not.
Does this affect share figures?
Scan-based share figures are largely unaffected, which is one of their advantages. Shipment-based share is severely affected and should not be used during periods containing known forward buying.
What about a price increase announcement?
It reliably triggers forward buying, which is why the announcement window is a known distortion. Planning the analysis around it in advance is better than explaining it afterwards.
How long does the trough last?
Until the bought inventory depletes, which depends on how much was bought and how fast it sells. A large buy in a slow-moving item can distort several periods.
Should the sales team be measured on shipments?
It is common and it creates exactly this incentive. Measuring on depletion where data allows aligns the measurement with the commercial reality, and it is one of the more consequential uses of scan data.
Does the independent channel do this?
Wholesalers serving it certainly do, and individual stores do to a lesser degree on deals. The wholesaler layer is where the inventory usually sits and it is the layer with the least visibility.
What is the most useful chart?
Shipments and measured depletion on one axis with cumulative difference below. It makes the whole phenomenon visible in a single view. Channel context is collected by NACS.
How should a forward-buy period be handled in a forecast?
Model the buy and the subsequent trough explicitly rather than letting the model learn them as demand. A forecast that treats a forward buy as a demand signal will repeat the spike in the following year.
### What should a brand actually do with this?
Quantify the margin transferred, once, and take it to the allowance negotiation. An abstract concern about forward buying moves nobody; a figure for what last year's allowance bought in inventory movement rather than consumption does.
Then structure the next allowance around depletion rather than purchase, which requires the depletion data to exist. That requirement is one of the clearer commercial arguments for measuring the channel at all.
Agree the measurement basis before the allowance period rather than during it. A depletion-based structure negotiated after a buy has happened is negotiating about a period both sides already read differently.
Working with retail data? See NRS Insights data solutions or contact NRS Insights.