Multiple unit pricing: how to measure a 2-for deal

What multiple unit pricing is

Multiple unit pricing is a shelf mechanic that gives the shopper a lower price per unit when they buy a stated quantity: 2 for $5, 3 for $10, buy two get one free. The single-unit shelf price does not move. The discount is conditional on quantity, which is the whole design and the whole measurement problem.

I have signed off on a lot of these from the trade finance side, and the pattern that got my attention was a 24 oz pasta sauce that came back from a 2 for $5 with a 72% unit lift and a gross profit line $2,461 worse than the four weeks before it. Both numbers were correct. The lift report and the P&L were reading the same promotion through different ends.

Must-buy and mix-and-match are different mechanics

The two configurations look identical on a shelf tag and behave completely differently in the data.

The mix-and-match case is the one that quietly breaks reporting. If the register applies the deal as a basket-level markdown, the line item for each SKU can still read $3.29 while a discount line elsewhere takes the money off. A per-SKU average-price read then shows no price change at all, and you get a lift that appears to have arrived for free. Before you measure any multi-buy, find out which of those two things your POS extract is showing you.

Mix-and-match also forces an allocation decision. If a shopper takes one sauce and one salsa on a 2 for $5, somebody has to decide how much of the $1.58 discount belongs to each SKU, and pro-rata by regular price, even split, and "charge it all to the anchor item" produce three different lift numbers for the same transaction.

Why unit lift on a 2-for always outruns dollar lift

The arithmetic is fixed and it is worth internalizing, because it explains almost every argument about whether a multi-buy worked.

A 2 for $5 on a $3.29 item is $2.50 a unit, a 24.0% cut. If units rise 72.2%, dollars rise by 1.722 times 0.760, which is 1.309, or 30.9%. Units and dollars cannot move together on a price cut, and the gap between them is exactly the discount depth. Anyone quoting the unit number without the dollar number beside it is quoting the flattering half of one calculation.

A worked 2 for $5

Marlow Creek pasta sauce, 24 oz, invented brand. Regular $3.29, unit cost $1.92, so $1.37 of gross profit at regular and $0.58 on deal. The mechanic is a must-buy 2 for $5 across 140 stores for two weeks.

Promo week 1 reconciles as 26 times $2.50 plus 6 times $3.29, which is $84.74, and gross profit as 26 times $0.58 plus 6 times $1.37, which is $23.30.

Averaged over the two promo weeks: 31 units, $82.24, $22.72 of gross profit per store per week.

• Unit lift: 31 / 18 minus 1 = 72.2%
• Dollar lift: 82.24 / 59.22 minus 1 = 38.9%
• Gross profit: down $1.94 per store-week, or $543.20 across 140 stores and two weeks

Note what the 6 units a week at regular price do. They are the single-unit tail: shoppers who wanted one jar and paid $3.29 for it. Because of them the effective price across all 31 units is $82.24 / 31 = $2.65, a 19.4% cut rather than 24.0%. If you had modelled the week by multiplying 31 units by the advertised $2.50 you would have reported $77.50 and a 30.9% dollar lift, understating the week by $4.74 a store. On a must-buy mechanic that tail is always there, and it is always ignored in the plan.

A quick diagnostic that costs nothing: on a genuine must-buy 2-for, qualifying quantities should skew even. If your POS shows a heavy odd-quantity tail, either the deal is misconfigured at the register or a large share of shoppers is declining it, and both change what the promotion was worth.

The pantry-loading tail

Look at the two weeks after the deal ends. The brand runs 12 and 14 against a base of 18. Those weeks are not a demand collapse. They are the deal being paid back:

• Gross incremental units in the promo weeks: (32 minus 18) plus (30 minus 18) = 26 per store, or 3,640 units across 140 stores.
• Units given back afterwards: (18 minus 12) plus (18 minus 14) = 10 per store, or 1,400 units.
• Net incremental: 2,240 units. The give-back is 1,400 / 3,640, which is 38.5% of everything the promotion appeared to add.

Multi-buy mechanics generate more of this than a straight percent-off does, for the obvious reason: the mechanic's entire purpose is to move two units where one would have gone, and a household that now has two jars of sauce is not buying sauce next week. That is the same pantry-loading effect that forward buying describes on the trade side, arriving one household at a time instead of one truckload at a time.

Run the gross profit over all four weeks and the picture completes. Promo weeks cost $543.20 against base. The two recovery weeks cost another $1,918.00, because those weeks sold at full margin but sold far too little of it. Total: $2,461.20 of gross profit for 2,240 net incremental units, which is $1.10 of margin given up per net unit gained. Whether that is a good trade depends on what else the deal bought, and it is a question nobody asks when the headline is "72% lift".

How to measure a multiple unit pricing deal honestly

• Read the observed price, not the advertised one. Effective price is promo dollars divided by promo units, and on a must-buy it will always sit above the per-unit deal price.
• Split deal units from regular units. If your extract cannot, ask for a discount-line breakout before you commit to a number.
• Set the baseline before you see the result. A clean pre-period, or a modelled baseline, decided in advance.
• Include the recovery weeks. Two weeks after is the minimum on a multi-buy. Stopping at the last promo week measures the borrowing and not the repayment.
• Report units, dollars and gross profit together. Any one of the three alone can be made to say the promotion worked.
• Watch the neighbours. A 2-for on one SKU pulls volume from its own siblings, so check the rest of the brand's line before you call it incremental.

The fuller version of that sequence is in evaluating promotions after they run, and discount pricing covers how a multi-buy compares to the other discount forms.

Where Scout fits

Scout reads your retailer POS and syndicated data and separates deal units from regular units, computes the effective price rather than the advertised one, and measures lift against a baseline that includes the post-promotion recovery weeks, so a multi-buy reports in units, dollars and gross profit at the same time. It also models the deduction cost of the deal when you are planning the next one.

One boundary, since this sits close to trade finance: Scout is the analytics layer upstream of deduction management. It models what a promotion should cost and what it returned. It does not match claims, work disputes, or run a cash-recovery workflow.

The short version

• Multiple unit pricing sets a lower per-unit price at a stated quantity while the single-unit shelf price stays put.
• Must-buy and mix-and-match are different mechanics. Mix-and-match often lands as a basket-level markdown, so per-SKU price looks unchanged and the lift appears to have cost nothing.
• Unit lift always outruns dollar lift by the depth of the discount: 72.2% units against 38.9% dollars on a 2 for $5 off a $3.29 regular.
• The single-unit tail lifts effective price above the advertised one, and the pantry-loading tail gave back 38.5% of the incremental units two weeks later.