Pick-pack throughput benchmarks for apparel: what a healthy 3PL should hit

What does a pick pack throughput benchmark for apparel actually look like?

It is 6:40 AM at a 3PL in Carson. The DTC wave for a $22M contemporary brand released at 6 AM: 1,840 orders, 2,610 units, mostly singles with a small share of two and three-unit orders. By 9:30 AM the shift lead expects the wave closed, cartons on the outbound dock, and ASNs firing. At 10:15 AM only 61 percent of the wave is picked. The brand’s ops manager is on Slack asking why Shopify still shows 900 orders in ‘processing.’ The 3PL says staffing. The real answer is that the pick path has 14 SKUs slotted in the mezzanine that belong on the ground floor, and the wave logic batched singles with two-unit orders in a way that forces pickers to walk the same aisle three times.

This post is about what a pick pack throughput benchmark apparel operators should actually hold their 3PL to, by channel, and how to tell whether a miss is a labor problem or a systems problem. In almost every case I see, it is the second one.

How do you define pick-pack throughput in apparel specifically?

Pick-pack throughput is the number of units a picker-packer processes per labor hour, measured at a defined accuracy rate, segmented by order profile. The three pieces matter equally. Units per hour alone is a vanity number. Units per hour at 99.5 percent accuracy, broken out by DTC singles, DTC multi-unit, wholesale cartonized, and drop or gifting flows, is an operating metric.

Apparel breaks the generic ecommerce throughput benchmarks you will find on 3PL marketing pages for three reasons. First, size and color variants multiply SKU count without multiplying unit volume, which lengthens pick paths. A brand with 400 styles has 4,000 to 8,000 SKUs on the floor. Second, apparel order profiles bifurcate hard: DTC is mostly one and two-unit orders, wholesale is cartonized by size run with retailer-specific labeling, and drops are high-volume spikes with gift packaging and inserts. Third, returns volume is 20 to 35 percent on DTC and those units have to be inspected, re-tagged, and re-slotted before they count as pickable inventory again. A 3PL that reports a single blended UPH number is hiding at least two of these.

What are the actual benchmarks a healthy apparel 3PL should hit?

From what I see across the install base, these are the ranges that correlate with brands whose ops managers are not on Slack at 10 AM asking where the wave is:

• DTC singles and two-unit orders: 60 to 120 units per labor hour, picker-packer combined. The high end requires discrete order picking with a well-slotted fast-mover zone and a packing station within 30 feet of the pick finish line.
• DTC multi-unit (3+ units): 90 to 150 units per hour. Batch picking into totes, sort at pack.
• Wholesale cartonized by size run: 150 to 300 units per hour. Pick directly into the shipping carton, no re-pack, GS1-128 label applied at close.
• Drops, launches, and gifted orders: 40 to 80 units per hour. The ceiling is packaging complexity, not pick speed. Tissue, stickers, hand-written notes, and insert cards cap throughput before the picker does.
• Returns processing: 25 to 50 units per hour per processor through inspect, grade, re-tag, and putaway.

On accuracy: 99.5 percent order accuracy is the floor, 99.8 percent is healthy, 99.9+ percent is excellent and usually indicates scan-verified packing. Below 99.5 percent and you are generating returns and chargebacks faster than throughput gains can pay for.

On dock-to-stock: inbound units should be receivable and pickable within 48 hours of truck arrival for domestic, 72 hours for international containers. Beyond that, the inventory exists in your WMS but not in your sellable pool, and the oversell rate starts climbing.

Why do most 3PLs miss these benchmarks?

The honest answer, from running the operational debrief with new customers during their first 90 days on the platform, is that the misses almost never trace back to the pickers. The pickers are usually faster than the system lets them be. The misses trace back to four places, in roughly this order of frequency.

First, slotting is wrong for the current demand curve. Apparel velocity shifts weekly during drop and launch cycles, and most 3PLs re-slot quarterly at best. A tee that was a slow-mover in April becomes the hero SKU in June because a creator posted it, and it is still slotted in the back of the mezzanine. Pickers walk twice as far as they should. UPH drops 30 percent and nobody can explain why.

Second, wave logic batches the wrong orders together. Mixing DTC singles with wholesale cartonized picks in the same wave forces context switching and defeats the purpose of either flow. Mixing standard DTC with gifted or launch orders means the pack bench serializes on the slowest unit. The wave builder either does not segment by order profile, or the order feed from the brand’s commerce stack arrives without the metadata the WMS needs to segment.

Third, the order feed itself is broken. Shopify flags are missing or misinterpreted. Wholesale orders routed through a general order channel instead of a dedicated flow. Pre-orders released into waves before inventory is actually dock-received. This is where the breakpoint lives: the warehouse is executing what it was told to execute, but what it was told came from a disconnected set of systems that nobody is reconciling. BP5 of the 6 Breakpoints framework names this specifically. Warehouse execution looks like the problem because that is where the symptom shows up, but the cause is one or two breakpoints upstream.

Fourth, exceptions eat the shift. A missing SKU on a pick cart kicks off a research cycle that pulls a lead off the floor for 20 minutes. One short-pick cascades into a wave re-plan. Healthy operations have exception rates below 2 percent of picks. Unhealthy ones run 5 to 8 percent, and the UPH number on the daily report hides it because the exceptions are logged as ‘adjustments’ rather than failed picks.

How does channel mix change what you should expect?

A 3PL running pure DTC for a $15M brand should hit 90 UPH blended at 99.7 percent accuracy without heroics. The same 3PL running the same brand with a 60/40 DTC-to-wholesale split should hit 110 to 140 UPH blended, because the wholesale carton picks pull the average up, assuming the two flows are physically and logically separated.

If blended UPH drops when wholesale is added, something is wrong. Usually it is that wholesale orders are being treated as oversized DTC orders rather than cartonized flows, which means pickers are picking singles out of cases into totes and then re-packing into cases at the end. That is two touches where one belongs. The fix is a dedicated wholesale wave, a pick-to-carton path, and a retailer compliance step at close that handles GS1-128 labels, carton content labels, and ASN generation. A connected multi-warehouse and 3PL operating model treats these as two different fulfillment profiles against the same inventory pool, which is what the channel split actually requires.

Drops are the stress test. A $15M brand running a Thursday 10 AM drop will see 3,000 to 8,000 orders land in the first two hours. A healthy 3PL has a drop playbook: pre-staged packaging, a dedicated wave release schedule (release in four batches across the day, not one), a packing bench expanded with cross-trained labor from receiving, and a cutoff for same-day ship that is enforced in the order feed, not promised by marketing. Magnolia Pearl’s drop and same-day fulfillment pattern is a useful mental model: the operation is designed around the drop cadence, with international duty handling and returns integrated into the same flow, rather than bolted on after the launch.

When is the throughput problem actually an inventory problem?

This is where I would push back on most 3PL conversations. If pickers are routinely sent to a location and the unit is not there, the problem is not throughput. The problem is that the inventory record and the physical reality have drifted. Short-picks cascade into exceptions, exceptions eat the shift, UPH collapses, and the 3PL writes it up as a labor issue in the Monday report.

For a $15M brand running wholesale, DTC, and a 3PL, ops teams spend 6 to 9 hours per week reconciling inventory across Shopify, the 3PL WMS, and the wholesale order book, and the oversell rate at peak sits at 2 to 3 percent. That reconciliation time and that oversell rate are the same problem showing up in two places. When the three systems disagree, pickers get sent to pick units that are not there, customers get orders they cannot be shipped, and someone on the brand side spends their Tuesday morning in spreadsheets instead of planning the next drop.

The diagnostic question is: when a short-pick happens, how long before every downstream system knows about it? If the answer is ‘end of day’ or ‘when the ops manager catches it in the reconciliation,’ throughput is going to miss benchmarks no matter who is picking. A functional warehouse management module posts short-picks and cycle count variances back to the master inventory record in minutes, which lets the order and allocation layer re-plan before more waves release against bad data.

How should a brand hold a 3PL to these numbers?

Ask for the UPH report segmented by channel and order profile, weekly, with the accuracy rate attached. If the 3PL cannot produce that report, that is the answer. They are measuring the floor with a single blended number, which means they cannot tell you whether DTC is healthy and wholesale is dragging, or the reverse.

Ask for the exception rate as a percentage of picks, and the top five reason codes. If ‘location empty’ or ‘SKU not found’ is in the top three, the slotting or inventory record is the issue, not the labor. If ‘oversized item’ or ‘damaged on receipt’ is in the top three, that is a receiving and QA conversation.

Ask for the dock-to-stock time for the last 10 inbound shipments. If the median is above 48 hours for domestic, you are losing sellable days every time a container lands, and your oversell rate during peak is partly a function of that lag.

Ask for the wave composition logic. Specifically: do DTC singles, DTC multi-unit, wholesale cartonized, and drop or gifted orders go into separate waves, or does the WMS batch them together? If the answer is ‘together, with sort at pack,’ that is a throughput ceiling you will never break through regardless of staffing.

And set the operating standard explicitly in the SOW: 99.5 percent order accuracy minimum, 48-hour dock-to-stock, segmented UPH reporting weekly, exception rate under 2 percent, and a joint quarterly slotting review tied to the demand curve. If the 3PL will not sign to those, the conversation is already telling you what next peak will look like. The warehouse execution scorecard is a reasonable baseline to work from if you need a shared diagnostic.

The throughput number is downstream of three decisions you already made

If you take one thing from this: the pick-pack UPH number on the Monday report is downstream of slotting, wave logic, and the order feed. Three decisions, two of which the 3PL controls and one of which the brand controls. When UPH misses benchmark, work backwards through those three before you accept ‘we need more pickers’ as the answer.

The brands I see hit the benchmarks consistently are not the ones with the biggest 3PL contracts or the most automated warehouses. They are the ones whose order and inventory layer feeds the warehouse clean, segmented, metadata-rich orders, so the WMS can wave them intelligently and the pickers can execute without exception cycles. The warehouse is where the symptom shows up when any of that breaks. It is almost never where the fix lives.

For a buyer or ops lead evaluating a 3PL right now, the right question is not ‘what is your UPH.’ It is ‘show me your UPH for apparel wholesale cartonized orders at 99.7 percent accuracy, and show me how you got there.’ The 3PLs that can answer that question are the ones worth signing with. The ones that cannot are the ones you will be re-benchmarking in nine months.