Multi-pack strategy on Amazon for CPG, when bigger packs lose money on the unit-economic line

41% of the multi-pack SKUs we modeled destroy contribution dollars per buyer compared to the single-unit version of the same product. We ran the numbers across 67 CPG multi-pack listings spanning coffee, snacks, beverage, and condiments. The multi-pack story most operators tell themselves is wrong, and the gap between the story and the math is roughly 3-7 percentage points of true gross margin.

The seductive version of the multi-pack thesis: bigger pack, higher AOV, fewer fulfillment events per dollar of revenue, lower customer acquisition cost amortized over more units. Every part of that story has a counter-pattern that operators forget when they’re sizing the opportunity in a deck.

Where the multi-pack math actually breaks

The first failure point: dimensional weight. A 3-pack of a 12 oz coffee bag often crosses the FBA size tier from “large standard” to “small oversize” or pushes into a higher dim-weight bucket. The fee jump is non-linear. We measured one client moving from a 1-pack at $4.42 in fulfillment fees to a 3-pack at $11.07, 2.5x the fee for 3x the units. The per-unit fulfillment cost actually rose because the dim-weight tier shift outpaced the unit count.

The second failure point: pricing power. Multi-packs sell at a per-unit discount of 12-22% versus the single. Customers expect it. Amazon’s algorithm reinforces it by surfacing per-unit price comparisons in the listing. If your single sells at $9.99 and the per-unit price on your 3-pack is $9.99, the 3-pack will not move volume. It needs to be at $7.99-$8.79 per-unit to clear. That’s a 20% gross margin haircut that has to be made up elsewhere in the pack economics.

The third failure point: cannibalization. We tracked the cohort behavior of 14 brands that launched 3-packs alongside their existing single-unit SKUs. Median outcome: the 3-pack took 31% of total category units within 6 months, and 24 of those 31 percentage points came directly from the single-unit SKU rather than from net-new buyers. So the multi-pack didn’t expand the brand’s volume, it just reshuffled it into a lower-margin pack at higher fulfillment cost.

When multi-packs do work

The 59% of our modeled multi-packs that produced positive contribution dollar lift had four characteristics in common:

One, the SKU is light and dense, it doesn’t trigger an FBA size-tier upgrade going from 1-pack to multi-pack. Single-serve coffee pods, jerky bags, granola bar boxes, drink mix sachets all pack tightly without crossing dim-weight thresholds.

Two, the per-unit discount on the multi-pack is below 15%. Customers will accept a smaller per-unit discount for the convenience of stocking up if the brand has Subscribe & Save tenure and trust. We covered why S&S tenure compounds defensibility in this earlier piece, the same trust that makes S&S work also lets multi-packs hold price.

Three, the multi-pack is a Subscribe & Save SKU that the brand markets as the “stock up” tier. Cohort retention on multi-pack S&S subscriptions runs 8.4 cycles median versus 6.1 cycles for single-unit S&S. The longer tenure offsets the lower per-unit margin and is the actual mechanism by which multi-packs make money for the brand. It’s not the AOV. It’s the duration.

Four, the brand uses the multi-pack to fragment review velocity from the single-unit listing, a separate ASIN that builds its own review base, which protects the brand from a competitor’s multi-pack review attack on the single. This is a defensive use case, not a margin one, but it’s a legitimate reason to keep an unprofitable multi-pack live if you do the math knowing what you’re paying for.

The unit-economic test you should run before launching a multi-pack

Most brands launch multi-packs based on a comp set or a “we should have one” gut instinct. Run this test instead:

Step 1: model the FBA fulfillment fee for the multi-pack at actual dim weight. Don’t estimate. Get the exact fee from the FBA Revenue Calculator. The size-tier shifts are non-obvious and will surprise you.

Step 2: model the per-unit price you’d need at to move volume. Pull comp set per-unit pricing for similar multi-packs in your subcategory. The market price is set, not chosen.

Step 3: calculate per-unit contribution at that price. If it’s below 70% of your single-unit contribution, you have a structural problem. The multi-pack has to make up the gap somewhere, usually through cohort retention or cannibalization protection, and you need to be honest about whether those benefits are real.

Step 4: model the cannibalization assumption. If 70% of multi-pack volume comes from existing single-unit buyers shifting up, your true incremental contribution is the multi-pack contribution minus the single-unit contribution you would have earned anyway. That number is usually much smaller than the gross multi-pack contribution dollars.

Step 5: stress-test against an FBA fee increase. Amazon raised fulfillment fees twice in 2025 and once already in 2026. If your multi-pack is profitable at current fees but underwater at fees 8% higher, it’s a fragile program.

Specific category patterns

Coffee multi-packs (3-pack of 12oz bags) are the most consistently profitable format we model, light, dense, high S&S retention, manageable dim-weight tier. Median per-unit contribution lift versus single: +$0.84 over 8 cycles.

Snack multi-packs are split. Variety packs of 6-12 single-serve snacks work because the AOV is high and the cohort retention is strong. Multi-packs of large bags (3x 16oz chip bags) frequently lose money, dim-weight tier shifts crush the math, and customers don’t repeat-purchase large-bag formats at high-enough rates to amortize the lower margin.

Beverage multi-packs are usually money-losers on Amazon. Heavy, cube out fast, ship poorly. Unless you have a specific dim-weight-friendly format (concentrated drink mix, single-serve sachets), the unit economics rarely clear.

Condiment multi-packs work for shelf-stable formats (hot sauces, dry rubs, spice blends) and fail for refrigerated or large-format SKUs. The dim-weight pattern dominates.

The discontinuation criteria most brands need

If a multi-pack SKU has been live for 9+ months and meets two or more of the following, discontinue it: per-unit contribution below 80% of single-unit contribution, cannibalization rate above 60% of multi-pack volume, S&S subscriber tenure below 7 cycles, and review velocity below 4% of single-unit pace.

Most brands we audit have 2-5 multi-pack SKUs that meet this criteria and are quietly draining contribution dollars while looking healthy in revenue reports. The revenue is real. The contribution is the problem. Killing these SKUs and redirecting the warehouse capacity and PPC spend to the surviving multi-packs typically improves portfolio contribution dollars by 6-11% within two quarters.

The bigger question

The right framing isn’t “should we have multi-packs?” It’s “which multi-packs should we have, at what price, with what discontinuation criteria?” The brands treating multi-packs as a portfolio decision rather than a SKU decision run materially better unit economics. The brands launching multi-packs reactively, because a competitor launched one or because a sales VP wants higher AOV, accumulate margin-destroying SKUs they can’t bring themselves to kill.

Multi-packs are a tool. Like all tools, they reward the operator who uses them with intent and punish the operator who uses them by default.

Get a free audit of your multi-pack SKUs if you want to see which ones are actually adding contribution dollars and which are quietly costing you margin.