Mispick Rate: Measuring It and the True Cost of a Mis-Ship

The warehouse scorecard shows picking accuracy at 99.8 percent, and customer service is handling forty wrong-item complaints a week. Both numbers are true. The scorecard divides errors caught at the pack bench by orders shipped, while the complaints are errors that slipped past that bench, and nobody has added the two together or priced what each complaint costs once the return label, the replacement parcel and the credit note are paid.

A mispick rate that operations, customer service and finance all accept needs three decisions written down: what counts as a mispick, which denominator to use, and where each error was found. With those fixed, the same error log produces a rate per thousand lines, an escape rate and a monthly cost in dollars or rupees that reconciles to credits issued.

• Measurement topics: mispick definitions and exclusions, line, unit and order formulas, detection points, escape rate, the cost build of a mis-ship, and the commercial and GST paperwork for wrong goods.
• Tools included: an error-type list, a SQL query for the rate and escape share, a worked monthly cost in $ and ₹, and a setup checklist for the dashboard.

What counts as a mispick and what belongs in another error code

A mispick is a pick confirmation that does not match what was physically picked for that order line. Wrong item, wrong quantity and wrong lot or variant all qualify; a picker honestly confirming a short because the slot was empty does not. Keeping those boundaries fixed is what makes the rate comparable month to month.

• Wrong SKU: a different item from the ordered one, including the adjacent slot, a similar pack size or a superseded part number.
• Wrong variant or attribute: right style but wrong colour, size, lot or expiry where the order or rotation rule specified it.
• Over-pick: more units sent than ordered and confirmed; it costs margin quietly because customers rarely report it.
• Unconfirmed under-pick: fewer units sent than the picker confirmed; distinct from a short pick recorded at the slot.
• Unapproved substitution: another item sent without a substitution rule or customer consent; approved substitutions are not errors.

Several common errors look like mispicks but belong elsewhere. A right item in the wrong carton is a packing error; a right carton with the wrong label is a labelling or manifest error; a slot holding the wrong item because of a putaway mistake is an inventory error, even if the picker faithfully picks from it. Code them separately, or the picking team will be blamed for upstream faults and the fix will target the wrong process.

Line-based, unit-based and order-based mispick rate formulas

Use order lines as the default denominator, because a line is the unit of work a picker confirms. Units suit case-pick operations where quantity errors dominate; orders suit customer reporting. State the denominator on every chart, since the three can differ by a factor of five on the same data.

-- Monthly mispick rate and escape share from a pick-line log (illustrative schema)
-- Mispick rate (line)   R_line   = M / L
-- Mispick rate (unit)   R_unit   = U_err / U
-- Mispick rate (order)  R_order  = O_err / O
-- Escape share          E        = M_cust / M
--   L = order lines picked;  M = lines with a mispick from any detection point
--   U = units picked;  U_err = absolute unit difference on mispicked lines
--   O = orders shipped; O_err = orders with at least one mispicked line
--   M_cust = mispicked lines first detected by the customer
SELECT
  date_trunc('month', p.confirmed_at)                         AS month,
  COUNT(*)                                                     AS lines_picked,
  COUNT(e.pick_line_id)                                        AS mispicked_lines,
  ROUND(100.0 * COUNT(e.pick_line_id) / COUNT(*), 3)           AS mispick_rate_pct,
  SUM(CASE WHEN e.detected_at = 'CUSTOMER' THEN 1 ELSE 0 END)  AS escaped_lines,
  ROUND(100.0 * SUM(CASE WHEN e.detected_at = 'CUSTOMER' THEN 1 ELSE 0 END)
        / NULLIF(COUNT(e.pick_line_id), 0), 1)                 AS escape_share_pct
FROM pick_line p
LEFT JOIN pick_error e
  ON e.pick_line_id = p.id
 AND e.error_type IN ('WRONG_SKU','WRONG_VARIANT','OVER_PICK',
                      'UNCONFIRMED_UNDER','UNAPPROVED_SUB')
GROUP BY 1
ORDER BY 1;

Two rules keep the query honest. First, an error belongs to the month the line was picked, not the month the customer complained, so the figure for October restates when late complaints arrive; publish a provisional number at month-end and a final one thirty days later. Second, one line counts once even if both the packer and the customer flagged it, which the join on pick line enforces.

Order-based rate is the figure customers feel. One mispicked line spoils the whole delivery, so an operation shipping orders of eight lines can show a line rate of 0.37 percent and an order rate near 3 percent. That relationship is why the perfect order rate uses order-level counting, and why a mispick figure by line should sit beside it rather than replace it.

Mispick rate in a warehouse management system log is mispicked order lines divided by order lines picked in the same pick month, with each error dated to its pick and counted once; at 1,524 errors in 412,000 lines the rate is 0.37 percent, or 3.7 errors per thousand lines.

Where mispicks get caught: pack check, dock audit or the customer

Every mispick is found somewhere; the question is how far downstream. Stamp each error with its detection point, because an error found at the pack bench costs a re-pick while the same error found by the customer costs freight, labour and goodwill.

• Pick-confirm scan: the WMS rejects a wrong item barcode at the slot; these never become mispicks if the scan is of the item rather than the location.
• Pack verification: item-by-item scan or weight check at the pack bench catches most wrong-SKU errors but misses same-weight variants.
• Outbound audit: a sampled carton re-check at the dock; useful for measuring escapes statistically, not for catching them all.
• Customer: the error arrives as a complaint, a refused delivery, a deduction on remittance or silence when the customer simply keeps an over-pick.

Pack verification is the main internal net, and its design decides which errors escape. Our guide to scan-to-pack verification covers how to verify carton contents before the label closes it; this article measures what that check catches and what it misses, and prices the misses. Bench layout matters too: an item scan is only practical where the scanner reaches every unit without the packer setting it down twice.

Escape rate: the share of mispicks that leave the building

Escape rate tells you whether internal checks work; the headline mispick rate cannot. A site whose rate falls while escapes stay flat has improved picking but not verification, and the cost line will barely move because escapes carry most of the money.

Calculate escape share as customer-detected mispicks divided by all mispicks for the same pick month. In the worked US case below, 288 of 1,524 mispicks escaped, an escape share of 18.9 percent. In the Indian case, 270 of 1,170 escaped, or 23.1 percent. Report escapes per thousand lines too, which normalises for volume: 0.70 per thousand in the US case, 1.50 per thousand in India.

Mispick rate reporting in a WMS error log should carry a detection-point field on every error record, so escape share equals customer-found mispicks divided by all mispicks for the pick month; at 288 escapes in 1,524 errors the share is 18.9 percent and drives over 90 percent of mispick cost.

Customer detection is incomplete. Over-picks are rarely reported, and some business customers deduct the value from their remittance without opening a case. Reconcile the customer-found count against credits issued and against receivable deductions coded to wrong goods each month; the gap between the three is a measure of how much of the escape rate you are not seeing.

Cost of one mis-shipped line: return, re-ship, credit and stock

Price an escaped mispick as the sum of six costs, not as the value of the goods. The goods usually come back; the freight, labour, credit and stock correction do not. Internal catches cost a fraction of that, so the two must be priced separately.

• Return freight: the label or reverse pickup to bring the wrong item back, or zero if the customer is told to keep low-value goods.
• Replacement freight: often expedited to meet the original promise date, so it costs more than the first shipment.
• Handling labour: return receipt, inspection, restock or disposal, and the replacement pick and pack.
• Customer contact: agent time on the case, priced at a loaded cost per contact.
• Credit or deduction: any goodwill concession, plus any charge the customer applies under its vendor terms for wrong goods.
• Expected write-off: the share of returned items that cannot be resold, multiplied by their cost.

The stock error is easy to miss. When black shirts went out instead of navy, the system now shows twelve too many navy and twelve too few black. The next black order can short-pick, and the navy surplus can trigger a missed replenishment. Count both SKUs within the shift the error is found, or one mispick quietly produces a second error days later.

A mis-ship priced for a US distributor and an Indian distributor

The table prices one month for two sites using the six-part build. All figures are illustrative; replace each unit cost with your own carrier rates, labour rates and credit history before using the totals in a business case.

Escapes are 92 percent of the US cost and 94 percent of the Indian cost. That ratio is the business case for verification: cutting escapes from 288 to 150 saves about $14,270 a month in the US case even if the mispick rate itself does not move. Improving pick accuracy at the slot also matters, but its payback shows up mainly through fewer escapes.

Mis-ship cost models for an Indian distributor should include the GST paperwork: a Section 34 CGST credit note for returned goods, declared by 30 November after the financial year, and an e-way bill under Rule 138 when the return consignment exceeds ₹50,000; in the worked case each escape costs ₹740.

Legal and tax paperwork when the wrong goods reach a buyer

A mis-ship is also a contract event. In US sales of goods the buyer may reject and the seller may cure; in India the return and replacement need the right GST documents. Build those steps into the case workflow so the credit, return and replacement follow a known path.

Under UCC 2-602, a buyer rejecting goods must do so within a reasonable time after delivery and must seasonably notify the seller. A buyer who has taken possession must then hold the rejected goods with reasonable care for a time sufficient for the seller to remove them. Under UCC 2-508, where the time for performance has not expired, the seller may notify the buyer of its intention to cure and then make a conforming delivery within the contract time.

• US case record: date of delivery, date and channel of the customer's notice, date of the cure notice and the replacement shipment's tracking number.
• Return instruction: whether the buyer is asked to return, hold for collection or keep the goods, written in the case before any credit is issued.
• India credit note: Section 34(1) of the CGST Act allows a credit note where supplied goods are returned, and Section 34(2) requires it to be declared no later than 30 November following the financial year, or the annual return date if earlier.
• India return movement: Rule 138(1) requires an e-way bill where a registered person moves goods with consignment value above ₹50,000, including movement for reasons other than supply.

Tracing a wrong-colour shipment from pick to customer credit

Most escaped mispicks follow the same chain: a weak confirmation at the slot, a check that cannot see the difference, a slow return and an uncorrected stock record. The trace below follows one illustrative order through each step.

1. Day 1, 08:14: picker confirms order SO-77120 line 3, twelve units of TSH-NVY-M, by scanning location A-14-03; the WMS accepts a location scan, and the picker takes the adjacent black shirts, TSH-BLK-M.
2. Day 1, 08:31: the pack bench weighs the carton at 3.62 kg against an expected 3.60 kg; the variants weigh the same, so the check passes and the label prints.
3. Day 3, 14:05: the customer reports black instead of navy; case RMA-5521 opens, a return label is emailed, and replacement order SO-77120-R ships next-day.
4. Day 4, 09:20: a different order for twelve TSH-BLK-M short-picks, because the system still shows black stock that left on Day 1; a second customer case follows.
5. Control added: pick confirmation now needs an item scan for SKUs with colour or size variants, and every customer-found mispick triggers same-shift counts of both SKUs.

Error coding a system enforces and root cause a supervisor judges

Recording and counting mispicks can be fully rule-based. Deciding why a mispick happened, and whether the picker, the slot layout or the item master is at fault, needs a person who can look at the location. Use software to suggest patterns, not to assign blame.

Setting up a mispick dashboard that finance will sign off

Finance signs off a quality metric when it reconciles to money already booked. Tie the escape count to credit notes and deductions, tie unit costs to carrier and labour rates, and the mispick cost line becomes an accepted figure rather than an operations estimate.

Read mispick figures alongside productivity, since pushing pick rates without verification tends to raise both errors and escapes. Our guide to picks per hour covers how to measure speed on the same pick-line log, while other warehouse KPI guides sit on the logistics hub.

Teams that want the error log, case linking and cost reconciliation inside their WMS reporting can send us a short brief with site count, order profile and current error codes; we reply with a proposed dashboard build.