Kitting and assembly: how a pile of components becomes a pallet of finished SKUs.

Kitting, assembly, and co-packing are not the same thing.

The terms get used interchangeably and that costs money. Kitting is the act of combining multiple individual SKUs into a single, ready-to-ship unit with its own SKU. A three-piece skincare bundle, a subscription box with monthly product variation, a retail display containing a dozen of one item and six of another, all are kits.^[3]

Assembly is the preparatory work that has to happen before kitting can run. Applying FNSKU labels, inserting tissue liners, sealing inner cartons that arrived open from a vendor, attaching hang tags, pairing items that ship together inside a single bag.^[5] Assembly is what turns a raw component into a kit-ready component.

Co-packing is enclosing finished products in retail or shipping packaging. Shrink-wrap, a blister pack, a printed retail box. Co-packing makes a product shelf-ready. Kitting takes those shelf-ready products and combines them.^[4] Many programs use both. A skincare brand co-packs each tube into its individual carton, then kits a three-tube gift bundle for the holidays.^[6]

How kitting actually gets priced.

Kitting pricing comes in three structures. Per-kit all-in pricing rolls labor, materials, and overhead into a single number per finished unit, typically $0.50 to $8.00 depending on complexity.^[1] Hourly pricing charges $18 to $35 per labor hour plus materials at cost or with a small markup, which works well for variable jobs where component count or steps change between batches.^[1] Hybrid pricing combines a small base per-kit fee ($1 to $2) with a per-component charge ($0.15 to $0.40), a model that scales fairly when component count varies inside a single kit family.^[1]

For subscription boxes, the typical fulfillment fee runs $3 to $8 per box for mid-complexity work.^[1] Simple kitting (combining two to four pre-packaged items into one package) runs $1.50 to $3.50 per kit.^[1] Complex kitting (multiple components, customization, assembly steps) runs $3.50 to $8.00 per kit.^[1]

One published industry estimate puts the gap between quoted and actual kitting cost at 35% to 55% on average across the industry.^[1] The reasons are predictable: BOM changes that add a step, rework when components arrive damaged, tier complexity inside subscription boxes that was not in the original spec. The fix is detailed quoting (labor minutes by step, material at cost, overhead surfaced) and a quarterly review of actual versus quoted.

“Per-kit pricing is cleanest when the kit is stable. Hourly pricing is fairest when the kit changes. The wrong structure for the work creates friction every month.”

Subscription box assembly: the operational pattern.

A subscription box program runs on a calendar. Components arrive over a window (some from manufacturers in Asia, some from domestic vendors), get received and quality checked, then run through assembly and kitting in a concentrated build window before ship date. The build itself is short. The receiving, staging, and prep work consume the bulk of the labor budget.

The recurring operational risks are inventory timing, tier variation, and damage rates on inbound components. Inventory timing problems show up as a missing component the day of build, which forces a hold on the entire run. Tier variation (different boxes for different subscriber segments) doubles or triples the SKU count and the pick complexity. Damage rates show up at receiving when a vendor ships poorly packed components, and they need to be caught early because a damaged component discovered during build means tearing a kit apart to pull the bad unit.

BOM management: the document that keeps a kit clean.

A bill of materials (BOM) lists every component that goes into a kit, by SKU and quantity, with the right revision and the right pack-out instructions. For a stable three-piece skincare bundle the BOM is short. For a subscription box that changes monthly with five subscriber tiers, the BOM is a small spreadsheet that needs versioning, approval, and a pre-build sign-off before a single box gets assembled.

The most common BOM problems are quiet. A revision that updates the insert card but not the version number, so the build uses the old card. A swap from a 1.0 oz sample to a 1.5 oz sample that nobody told the warehouse about, so the box weighs 20 grams more and the shipping label rate changes. A vendor substitution that changed a component's UPC, which then fails the customer-facing scan check on receipt. None of these are catastrophic individually. They compound.

A working BOM process has four properties. It versions every change with a date and an approver. It assigns a unique kit SKU to each variant, including tier variants. It includes a pack-out diagram or photo so an operator can verify the build visually. It requires a first-article inspection on the first kit of each new run before the rest of the run executes. None of those steps are expensive. Skipping them is.

Retail display kits and the math of mixed counts.

A retail display kit is a shipper-ready unit destined for a shelf or endcap at a retailer. The kit might contain twelve units of one SKU, six of another, and a printed header card. The retailer wants the display assembled, shrink-wrapped, and palletized to their compliance specifications, which vary by chain. Walmart, Target, Costco, and Whole Foods each have their own pallet height limits, label placement rules, and case-pack requirements.

The cost driver here is mixed counts inside one shipper. A pure case of 24 units of one SKU is fast to pick, fast to verify, and cheap to handle. A mixed shipper with five SKUs in different counts requires more pick passes, more visual verification, and more potential for error. The kitting fee reflects that. So does the rework rate when a kit ships short on one SKU because the pick was wrong.

Ecommerce bundles: the simplest kitting program, run well or run badly.

An ecommerce bundle (buy two get one, a starter kit, a holiday gift set) is the most common kitting work and often the most poorly run. The temptation is to treat a bundle as a virtual SKU and pick the components at order time. That works for small volume. It stops working when bundle volume scales, because component-level picking costs labor every order, and because a bundle with three components is three times as likely to ship short as a single SKU.

The cleaner pattern is to forecast bundle demand, kit a buffer of finished bundles ahead of time, and replenish the bundle SKU as orders draw it down. Pick once at the bundle level and ship as a single SKU. The trade-off is inventory carry on the buffer and the cost of breaking up a bundle if demand shifts. For most stable bundle programs, the kitting buffer pays for itself inside the first quarter on labor savings alone.

Quality control inside the kitting line.

A working QC program inside kitting has three layers. First-article inspection at the start of every run, where one finished kit is opened, photographed, and signed off against the BOM before production scales. Inline visual verification at each station, so an operator confirms the right component is going into the right slot. End-of-line audit at a sampling rate, typically one in twenty for stable runs and one in five for new programs.

Defect rates on subscription boxes typically run between 0.2% and 1.5% depending on program complexity. The lower end requires investment (better receiving, photo SOPs, tighter BOM control). The higher end is what happens when those investments do not get made. Each defect is a customer service ticket, a replacement shipment, and a small dent in subscriber retention. The math gets ugly fast at scale.

When kitting belongs at a 3PL and when it belongs in-house.

Kitting belongs at a 3PL when the volume is variable, the labor pool needs to flex, and the work runs in concentrated build windows that do not justify a dedicated facility. That description fits most subscription box programs, most ecommerce bundles, and most retail launches. A 3PL spreads the labor pool across multiple clients, which means the ramp-up cost of a new program is small and the ramp-down cost when a program ends is zero.

Kitting moves in-house when volume is large enough to fill dedicated labor every day, the program runs continuously rather than in build windows, and there is meaningful intellectual property in the assembly process worth protecting. Even then, most brands keep an outsourced overflow capability, because launches and seasonal spikes overwhelm in-house capacity.

“The right answer depends on volume stability. Stable, high-volume programs in-house save money. Variable programs at a 3PL save money. Building one to handle the other gets expensive.”

How Warpspeed runs kitting and assembly.

Warpspeed runs kitting alongside the same fulfillment operation that ships the finished kits, which removes a transfer step and a handoff that often introduces errors. BOMs live in the WMS, not in a separate spreadsheet. First-article inspection runs on every new build with photo capture stored against the kit SKU. End-of-line audits sample against the BOM, and discrepancies route to a supervisor before the kit goes to the pick line.

For subscription box clients the calendar drives the operation. Inbound vendor receipts stage against ship date. The build window is sized against confirmed received inventory, not the PO, so missing units do not become a Friday surprise. Tier variation is handled at the kit-SKU level, with separate pick lists for each subscriber segment, so a Tier 2 box does not get built with Tier 1 components.

Related reading

• FBA prep services that ship clean to the FC
• Returns and reverse logistics, end to end
• Transportation management built around the warehouse