Restock velocity: the metric that pays for your returns program

Restock velocity is the time it takes a resellable returned unit to go from receiving to being back in sellable inventory, and it is the single metric that decides whether your returns program recovers margin or bleeds it. Every hour a good unit sits in a grading queue is an hour it cannot be sold at full price, so restock velocity is where return-to-shelf time turns directly into returns margin recovery. If you only track one number on your returns operation, track this one.

What restock velocity actually measures

Restock velocity is a subset of total cycle time. Total cycle time covers every return, including the ones you liquidate, donate, or destroy. Restock velocity isolates the units that are worth putting back on the shelf and asks a narrower question: how fast does a resellable unit become resellable inventory again?

The two clocks start at the same place, the inbound scan, but they end differently. Cycle time ends at disposition. Restock velocity ends when the unit is available to sell, whether that is your own storefront, a marketplace listing, or a wholesale lane. That distinction matters because a return can be dispositioned as "restock" in minutes and still take days to reach a sellable state if the handoff back to inventory is slow or manual.

Why speed equals margin

Resellable inventory recovery is a race against three forms of decay, and each one compounds the cost of a slow shelf return.

• Demand windows. The customer who returned an item is often returning it inside a live demand window. Another buyer wants that exact SKU right now. Restock it in 48 hours and you catch that demand. Restock it in two weeks and the sale has moved to a competitor or to new inventory you had to purchase.
• Markdown decay. The longer a unit sits, the more likely it gets swept into the next markdown cycle. A B-grade jacket that could have sold at 85 percent of full price in week one often clears at 50 percent by week four, not because its condition changed but because the calendar did.
• Seasonal cliffs. Seasonal and trend items have a hard expiry. A returned swimsuit in August and a returned swimsuit in October are different assets. Slow restock velocity turns full-price inventory into liquidation lots.

We break down the full margin math in the real cost of a return, but the short version is that recovery value is a decaying function of time. Restock velocity is how you keep the recovered value high.

How to measure restock velocity

The formula is simple. The instrumentation is where teams fall short.

Restock velocity = timestamp(unit available to sell) minus timestamp(inbound scan), measured only across units dispositioned as restock.

Report it two ways. Use the median for your operational headline number, because averages get distorted by a handful of stuck edge cases. Use the 90th percentile to expose the tail, because that tail is where trapped margin hides. To calculate either honestly, you need event-level timestamps rather than a single "processed" flag.

Instrument these events at minimum:

1. Inbound scan at receiving. This is the start of both clocks. If your receiving is slow or batched, fix that first. See scan to search back to back receiving for how tight receiving keeps the clock honest.
2. Graded, when a condition grade and confidence score are assigned.
3. Dispositioned, when the routing decision is made.
4. Restocked, when the unit is available to sell.

At areturnz these fire as webhooks (return.received, return.graded, return.dispositioned, return.restocked), so the timestamps are captured as events rather than reconstructed after the fact. That is what makes the median and the P90 trustworthy, and it is what lets you see exactly which stage is eating your velocity.

Realistic benchmarks

Our median cycle time is roughly 48 hours from inbound scan to disposition across all returns. Restock velocity for resellable units should track close to that, and often beats it, because the highest-confidence resellable grades are the fastest to clear. A practical target ladder looks like this:

• Good: median restock velocity under 72 hours, P90 under 5 days.
• Strong: median at or under the 48-hour cycle-time median, P90 under 72 hours.
• Best in class: A-grade units cleared same-day or next-day because high-confidence grades are auto-accepted straight into the restock lane.

If your median restock velocity is measured in weeks, the problem is almost never the physical work of putting an item on a shelf. It is the decision-making in front of that shelf.

What slows restock velocity, and how to fix it

Queue time

Units pile up waiting for a human to look at them. Batched grading, understaffed receiving, and end-of-day processing all inflate the wait before any real work happens. The fix is continuous receiving and grading at the point of intake, not in batches.

Manual grading inconsistency

When two operators grade the same jacket differently, downstream routing becomes unpredictable and every ambiguous unit gets kicked to a review pile. Consistent, evidence-backed grades remove that friction. Our AI condition grade sits at roughly 99.6 percent match accuracy against operators, which means the grade almost never becomes the reason a unit stalls. Read how the scale works in the A/B/C/R AI condition grading guide.

Disposition ambiguity

If nobody is sure whether a B-grade item restocks or liquidates, it waits. Codified disposition rules that combine grade plus item attributes remove the guesswork and route each unit automatically, with operator overrides logged for the exceptions. That is the mechanism that converts a grade into a lane without a meeting. See turning grades into decisions with disposition rules.

Low-confidence handling

The real speed unlock is auto-accepting high-confidence grades. When a grade carries a high confidence score, it flows straight to restock with no human touch. Only borderline units get a human, so operator attention concentrates where it changes an outcome. We cover where to set that threshold in confidence scores on the line.

Grade to lane to velocity impact

The table below shows how the A/B/C/R grade drives both the typical disposition lane and the realistic restock velocity for each.

Grades A and B are where restock velocity is won. The faster and more consistently you can confirm those grades, the more full-price and near-full-price recovery you capture. B-grade units in particular reward speed, since many can be resold under your own brand rather than dumped into liquidation.

Turning velocity into a recovery program

Restock velocity is not just an ops KPI. It is the number that funds the whole returns program. Faster resellable inventory recovery means more units sold at full or near-full price, fewer markdowns, and less new inventory bought to cover demand you already had in hand. When you model the cost per return against your recovery lanes, velocity is the lever that moves the total the most. See how the lanes and per-return economics fit together on our pricing page, and for more on measuring recovery across the funnel, browse the returns economics library.

Frequently asked questions

What is a good restock velocity benchmark?

Aim for a median restock velocity at or under 48 hours from inbound scan to sellable, with a P90 under 72 hours. Best-in-class operations clear high-confidence A-grade units same-day or next-day by auto-accepting the grade straight into the restock lane.

How is restock velocity different from cycle time?

Cycle time measures every return from inbound scan to disposition, including liquidate, donate, and destroy. Restock velocity measures only resellable units and ends when the unit is available to sell, not when the routing decision is made. It isolates the units that carry recoverable margin.

What is the biggest cause of slow restock velocity?

Decision friction, not physical handling. Queue time before grading, inconsistent manual grades, and disposition ambiguity trap units in review piles. Evidence-backed grading with confidence scores and codified disposition rules removes that friction so good units move without waiting on a human.

Does faster restock velocity actually recover more margin?

Yes. Recovery value decays with time through demand windows closing, markdown cycles, and seasonal expiry. A resellable unit restocked in 48 hours routinely sells at a higher price than the same unit restocked in two weeks, so velocity converts directly into returns margin recovery.

How do I instrument restock velocity?

Capture event-level timestamps at inbound scan, graded, dispositioned, and restocked. With webhooks firing on each event, you can compute the median and 90th percentile per SKU or category and see exactly which stage is slowing you down, rather than reconstructing timing from a single processed flag.

Related reading: The True Cost of Returns Processing (And How to Cut It)