- A 3PL dashboard needs eight core metrics: SLA rate, volume, picker/packer performance, pick accuracy, return rate, cycle time, perfect order rate, cost per order.
- A perfect order rate of 97 to 98 percent is a reasonable experience-based target, not a fixed industry norm.
- Every metric belongs split by client - a blended number hides exactly the accounts causing trouble.
- 3PL KPIs are operational metrics, not retail ones - they measure your own performance, not your client's.
What is the SLA rate?
What it is. The share of orders shipped within the timeframe the contract promises. For a 3PL this is the anchor metric, because it sits directly on the contract rather than an internal target.
How you measure it. Per order: did the ship timestamp fall inside the deadline built from order release plus that client's cutoff rule? The rate is on-time orders divided by all qualifying orders, kept separate per client, because every client's cutoff and contract terms differ.
- Store a cutoff time per client - orders released after cutoff count against the next working day, not the same one
- Count in working days, not calendar days, or every weekend reads as a batch of missed SLAs
- Strip holds, backorders and cancellations out of the qualifying volume before you calculate the rate
What working days and cutoffs mean in practice is covered on its own in fulfillment SLA metrics, including where the clock should start and stop.
How do you measure shipment volume and throughput?
What it is. How many shipments a 3PL actually ships per day, per week and per client. This is the baseline for staffing, capacity limits, and whether a new client even fits.
How you measure it. Count the real ship timestamp, not the moment a packing slip was generated - the two drift apart regularly. Track orders and parcels as two separate numbers, because multi-parcel shipments otherwise distort any staffing calculation built on top. Keep it split by client so a spike at one account doesn't disappear inside the overall average.
Common pitfalls. Cancellations reversed after shipping need to be backed out. A pure monthly average also hides the daily peaks where capacity actually runs tight.
How is picker and packer performance measured?
What it is. How many picks or packs a worker completes per hour. This is the number staffing plans and client pricing are built on - and the one most commonly measured wrong.
How you measure it. Track picks per hour and packs per hour separately, since the two tasks run on different cycle times. The defensible formula is volume divided by actual clocked time from time tracking, not the span between the first and last scan - a span that hides every minute of waiting, staging and breaks as if it were productive work. The full mechanics, including the scan-span trap, are broken out in warehouse picker performance metrics.
A worker is clocked in for nine hours but scans their parcels inside a two-hour window. Scan-span math shows almost 50 parcels an hour. Measured against clocked time, the real rate is eleven. Both numbers are arithmetically correct - only one is honest.
How do you measure pick accuracy?
What it is. The share of picks completed without an error - right item, right quantity, right location. For a 3PL this is a quiet but expensive metric: every miss triggers a return, a complaint or a replacement shipment that costs time and trust.
How you measure it. Correctly picked lines divided by all picked lines over a defined window. The cleanest source is scan verification at the pick face; without a strict scan requirement, complaint rate is the fallback proxy. Split by client, because assortments with many similar variants - sizes, colors, near-identical SKUs - are structurally more error-prone than uniform catalogs.
Common pitfalls. Without scan verification, the number is really an estimate built off downstream complaints - it understates the real error rate, because not every mistake gets noticed or reported by the end customer.
How do you calculate the return rate?
What it is. The share of shipments that come back whole or in part. For a 3PL this isn't only a receiving cost - it's an early warning signal for pick errors or process problems in the operation itself.
How you measure it. Returns divided by shipments over a rolling window, commonly 30 days. Split by client, because return levels track the assortment closely - apparel behaves nothing like supplements.
- Define the reference window clearly - returns lag shipments in time
- Compare rates, not raw counts, or any volume increase reads as a returns problem
- Track return reasons alongside the rate to separate internal errors from causes outside your control
How do you measure cycle time?
What it is. The time from order release to ship. Where the SLA rate only knows "met or missed," cycle time shows the full spread: where the median sits, where the long tail starts, which orders are getting stuck.
How you measure it. The difference between ship and release timestamps, evaluated as a distribution rather than a single average. The median is more informative, since outliers otherwise pull a straight average upward. Worth splitting by client and order type.
Common pitfalls. A raw difference ignores weekends, public holidays and hold time. A cleaned cycle time with those effects removed is the number you can defend to a client - the raw figure is only good for an internal quick look.
What is the perfect order rate?
What it is. The share of orders that ran error-free across every relevant dimension: on time, complete, undamaged, correct paperwork. A single miss in any one dimension fails the whole order - which makes the perfect order rate the strictest and most informative roll-up of every other metric.
How you measure it. For each order, check whether every sub-criterion was met (SLA hit, quantity correct, no transit damage, documents match). Only if all criteria hold does the order count as perfect. The rate is that count divided by all orders in the period.
In our experience, not as a fixed benchmark: 97 to 98 percent is considered strong. Numbers below that usually trace back to one weak sub-metric - often the SLA rate or pick accuracy of one specific client, invisible in a blended average unless measurement is split by client.
How do you calculate cost per order?
What it is. What fulfilling an order costs on average - labor, packaging, shipping, allocated overhead. This is the number that decides whether a client is profitable or whether terms need renegotiating.
How you measure it. Relevant costs for a period divided by orders fulfilled in that same period, either as pure pick-and-pack cost or fully loaded. It only becomes reliable once throughput and volume are known per client - only then can the cost block be attributed on a fair basis.
Common pitfalls. A single blended calculation across all clients hides cross-subsidy: the large, simple client quietly carries the small, complex one, and nobody sees it. This is an accounting exercise, not a pure warehouse measurement - a reporting layer supplies the volume and performance basis for it without replacing bookkeeping. Connecting the systems these numbers come from is covered under warehouse system connections.
Why does every metric need to be split by client?
A 3PL serves multiple clients at once, each with its own contract, assortment and cutoff time. A blended figure across all of them is close to useless for running the operation, because it hides exactly the cases that need attention: a client with a tight cutoff and a complex assortment misses its SLA regularly, while a second one with a generous window and simple assortment drags the overall average back up. On average, everything looks fine. Case by case, something is on fire.
The same applies to cost per order, return rate and pick accuracy - each depends on the assortment and contract terms of its own client. Only client-level measurement shows which account is actually profitable, where a renegotiation is due, and where an SLA is genuinely at risk - before the client has to call and ask.
Why don't retail metrics work for a 3PL?
Metrics like average order value, margin per item, return cost against revenue, or customer lifetime value belong to the client, not the 3PL. They describe how well a product sells and how profitable the client's own business model is - questions a fulfillment operation has no direct control over.
The eight metrics above measure something else entirely: your own operational performance. Was it shipped on time, was it picked correctly, how productive was the team, what did fulfillment cost. A 3PL dashboard covers this second world, not the client's. Both sets of metrics exist side by side and answer different questions for different audiences - a client portal that shows each brand only its own operational numbers, without exposing the other set, is covered under a separate login per client.
Common questions
What KPIs does a 3PL dashboard need at minimum?
Eight core metrics cover most of it: SLA rate, shipment volume, picker and packer performance, pick accuracy, return rate, cycle time, perfect order rate and cost per order. All eight need to be split by client, because a blended number across customers hides exactly the accounts that are underperforming.
What is a good perfect order rate for a 3PL?
In our experience, not as a fixed benchmark: 97 to 98 percent is considered strong. The perfect order rate only counts orders that arrive on time, complete, undamaged and with correct paperwork - a single miss in any one of those four dimensions counts as a full failure for that order.
Why don't retail metrics work for a 3PL?
Metrics like average order value, margin or return cost per item belong to the client and say nothing about how the warehouse performed. A 3PL needs operational metrics that measure its own contract compliance and productivity: SLA adherence, pick accuracy, throughput per worker. Both sets of metrics exist side by side, but a 3PL dashboard only covers the second one.
Why does every 3PL KPI need to be split by client?
Because every client runs its own cutoff times, product mix and contract terms. A blended figure across all clients hides the fact that one account with a tight cutoff is missing its SLA regularly, while another with a generous window is pulling the average back up. Only client-level measurement shows where a contract is actually at risk.
Can a warehouse system produce these KPIs on its own?
Most give you the raw order and shipment data these metrics are built from, but not the finished, per-client calculation - cutoff rules, working-day calendars and exclusions still have to be defined and applied consistently. That definition work is usually the missing layer, not the data itself.