A vendor scorecard: on-time delivery, fill rate, damage, and margin

Ask the buyer which vendor caused the most trouble and the answer arrives immediately. Ask for the evidence and the story begins with one terrible spring. The story may be right. A scorecard makes it possible to see whether it describes the whole relationship or one memorable purchase.

Use one row per vendor per review period, with definitions your team can repeat. Record on-time delivery, first-shipment fill, rejected or damaged units, landed margin, and warranty outcomes. Keep the underlying counts beside the percentages so a pleasant phone manner and a small sample do not decide the result.

Define on time before calculating it

On-time performance needs an agreed milestone and comparison date. For this template, use complete receipt of the ordered goods by the baseline receipt date. The formula is POs completely received on time divided by POs due in the period with a usable baseline date, multiplied by 100.

Choose the baseline consistently. A requested delivery date measures performance against your request; an acknowledged date measures performance against the vendor's accepted estimate. Keep those measures separate if you use both. A vendor ship date cannot be compared directly with your warehouse receipt date without accounting for the different milestone.

Preserve the first agreed baseline and later revisions. Continually replacing the baseline with the latest delayed date can make a chronically late vendor look punctual. Record the updated estimate for operations while retaining the original comparison for the review.

If a vendor never supplied a usable date, mark that purchase as missing a date. Exclude it from the dated on-time denominator and report the missing count separately. Do not award an on-time result to an order that had no measurable promise.

Measure fill and rejection with explicit units

First-shipment fill measures how much of the ordered quantity arrived in usable condition in the first shipment. Use accepted units in that first shipment divided by ordered units for the same purchase cohort, multiplied by 100. State whether you measure units, lines, or complete POs; those produce different answers.

The rejection rate needs a different denominator. Use rejected units divided by all units presented for receiving, meaning accepted plus rejected units, multiplied by 100. Dividing rejected units by accepted units alone inflates the rate and becomes unusable when everything is rejected.

Separate reasons within the rejection total. Wrong specification, manufacturing defect, and transit damage are different problems. Keep carrier responsibility unresolved where the evidence is unresolved. A useful scorecard shows the customer and operational impact without automatically charging every broken corner to the maker.

The receiving-corrections guide helps keep the inputs accurate. Correct the record when the facts change, but preserve the reason so the next review can explain the difference.

Use landed margin and warranty outcomes together

Landed margin compares net merchandise sales with the associated landed cost of those sold goods. Calculate net merchandise sales minus landed cost, then divide by net merchandise sales. Match the cost to the sales cohort and include the recorded freight and other landed charges on a consistent basis.

Do not divide this quarter's sales by the cost of everything received this quarter if some of those receipts remain unsold stock. The dates may match on a calendar while the furniture does not. Returns and credits also need consistent treatment before vendors can be compared fairly.

For warranty, keep claim counts, claim costs, recoveries, and unresolved amounts separate. A claim count without the number of delivered units and their age can mislead: a long-established line has had more time to generate claims than one introduced recently. The warranty-report guide helps interpret that context.

Add responsiveness as an explained judgment

Responsiveness can be a simple assessment of whether the vendor answers clearly and follows through. Use the same questions for each review: did the acknowledgment arrive as agreed, were differences answered, and did the vendor resolve the claim or credit when promised?

Keep a short note beside any rating. “Hard to work with” is not a useful measure. “Acknowledgment corrections required repeated follow-up on the listed purchases” gives the rep something to address. The scorecard is a business discussion, not an annual prize for the best telephone voice.

Copy the quarterly scorecard layout

Use these fields as spreadsheet columns or a review table. The store sets its own targets; this template supplies no furniture-industry benchmark.

Add vendor, review period, purchase count, purchase value, data owner, agreed action, and next review date. Keep the detailed PO or claim references in a supporting tab so the summary remains readable.

If you want a weighted result, define the scoring scale before adding weights. For an example internal scheme, weight on-time performance 30%, fill 20%, rejection 20%, landed margin 20%, and warranty 10%. Convert each to a higher-is-better score using targets the store sets, then apply the weights. Keep responsiveness as commentary unless the team can rate it consistently.

Work through the counts before judging the vendor

Consider an illustrative quarter with 23 dated POs due. Fourteen arrived complete by their baseline date and nine did not. On-time performance is 14 ÷ 23, or about 60.9%. Another three purchases lack usable dates and are shown separately; quietly counting them as punctual would change the answer without improving delivery.

For a separate unit-based cohort in the same example, suppose 100 units were ordered, 80 were accepted on their first shipment, and 10 of 100 units presented across receipts were rejected. First-shipment fill is 80%; rejection is 10%. Those measures use different events and denominators, so one is not simply the opposite of the other.

There is no magic PO count at which the result becomes statistically reliable for every vendor. Show the sample and review its mix. A handful of complex special orders and a large repeat-stock program should not be compared as though they represent the same workload. Use several periods and the underlying cases before making a major buying decision.

Use GMROI for the stock investment

For stock programs, gross margin return on inventory is gross margin dollars divided by average inventory at cost over the same period. Keep customer-specific purchases and the stock program's inventory basis distinguishable so the result answers the floor-space question you intended to ask.

Say the example stock program generates $100,000 of sales and $65,000 of cost of goods sold, with average inventory cost of $25,000 during the quarter. Gross margin is $35,000 and quarterly GMROI is $35,000 ÷ $25,000, or 1.4. That is a quarterly result, not an annualized industry benchmark.

The equivalent calculation is turns × margin rate ÷ (1 − margin rate), when all inputs share the same cost basis and period. Here, turns are 2.6 and margin rate is 35%, giving 2.6 × 0.35 ÷ 0.65 = 1.4. The GMROI guide explains how to use that result alongside the role of a display sample.

Gather the underlying records in OrderPilot

OrderPilot holds the purchasing evidence rather than a built-in vendor scorecard report. Review the vendor's POs, requested delivery dates, and dated acknowledgments in the conversations. Copy the appropriate comparison dates and receipt outcomes into the scorecard using the definitions above. Receipt history supplies received and rejected quantities and the reasons behind corrections. Landed-cost records help establish the cost basis, while the associated order records provide the selling side. Keep a reference to the purchase behind each figure so a disputed result can be checked. On Professional, Analytics › Warranty provides claim information by vendor and product, including costs and recoveries. Use those records with the delivered cohort and unresolved-claim context rather than interpreting the highest raw count as the worst supplier automatically.

Share the evidence and agree a next action

Share the scorecard with the vendor when it supports a specific review. Give the rep the definitions, period, and purchase references, and invite corrections to the facts. Keep the store's confidential customer information out of the review materials where it is not needed.

Use red, amber, and green bands based on the store's own agreed targets, and retain the raw counts beside them. A color should direct attention to the next action, not conceal the reasoning. The negotiation guide turns that review into a concrete request about terms or service.

For a downloadable version of this layout, the workbook should have a definitions tab, a PO-and-receipts input tab, a cost-and-claims tab, and the quarterly summary. The complete fields and formulas above can be copied into that structure without waiting for a graphic. Return to the purchasing process when the scorecard reveals a missing handoff in the records themselves.