Dropshipping Supplier Scorecard: Grade Suppliers With Your Own Orders

Grade every dropshipping supplier from your own order data. Track five numbers, use keep or replace thresholds, and see what a slow, error-prone supplier really costs per 100 orders.

Dropship with Spocket
Ashutosh Ranjan
Ashutosh Ranjan
Created on
November 17, 2026
Last updated on
November 17, 2026
9
Written by:
Ashutosh Ranjan

A dropshipping supplier scorecard is a one-page grade for each supplier, built from orders you have already shipped instead of from the listing page. Most stores vet a supplier once, before the first sale, and never look again. Six months later that supplier is dispatching in four days instead of one, a few parcels in every hundred arrive wrong, and the owner only notices when refunds start eating the month. This guide shows how to pull the numbers from your own order history, which five to track, where to draw the line between keep and replace, and what the money looks like when you get it wrong.

Why your own order data beats the listing page

Supplier listings describe intent. Your order history describes behaviour. Stock levels drift, a warehouse switches carriers, a good account rep leaves, and a supplier that was fast in March can be slow by September. None of that shows up on a product page, and none of it triggers an alert.

Customers feel it first. In Baymard's latest study of why US shoppers abandon a checkout, 20% named slow delivery as a reason, second only to extra costs. If your product page promises 3 to 8 days and one parcel in ten takes 14, you pay for that gap in refunds, support tickets and reviews that mention waiting.

There is a legal angle too. The FTC's business guide to the Mail, Internet, or Telephone Order Merchandise Rule says you need a reasonable basis for any shipping time you state. That is general information, not legal advice, but your own delivery data is the obvious place to get that basis.

Pull the data out of your store

Start with an export. In Shopify, open Orders, filter to the last 90 days, click Export and choose the date range. The Shopify Help Center export guide lists the columns you need: Created at, Fulfilled at, Fulfillment Status, Refunded Amount, Vendor and Tags. Exports of more than 50 orders are emailed to you, so give it a few minutes. On other platforms, look for an order export with the same basics: order date, shipped date and refund total.

Open the file in a spreadsheet with one row per order. Add three columns of your own. Supplier comes from the Vendor field or your tags. Delivered date comes from the tracking page. Problem stays blank unless something went wrong, in which case you type wrong item, damaged, missing or late. Delivered dates are the slow part, so do not chase every one. A sample of 25 parcels per supplier gives a rough view of delivery times, nothing more.

Check what "Fulfilled at" means in your setup before trusting it. Some apps mark an order fulfilled the moment it is placed with the supplier, which makes dispatch look perfect while parcels sit on a shelf. If that is how yours works, use the date the tracking number first shows a carrier scan instead.

Wait for volume as well. With 20 orders, a single bad parcel moves a rate by five points, which is noise. Problem and refund rates mean little below about 50 orders per supplier, and the 90th percentile only means something once you have 50 or more delivered orders. Until then, grade on the median and mark the rest "too early".

The five numbers on a dropshipping supplier scorecard

1. Median dispatch time

Count business days from Created at to the day the parcel left the supplier. Use the median, the middle value, rather than the average, because one order stuck for three weeks drags an average badly while telling you little about a normal day. Compare the result with the processing time the supplier lists. A gap of more than a day is your first warning.

2. Delivery tail

Take the 90th percentile of dispatch-to-delivery days, which is the slowest one parcel in ten. In a spreadsheet that is =PERCENTILE(range,0.9). The customers who wait the longest tend to be the ones who open disputes and write reviews, so the tail predicts your complaints better than the middle does. Compare it with the window you promise on the product page.

3. Problem rate

Tally orders with a wrong item, wrong variant, damage or missing items, then divide by total orders. Tag these orders the day they happen, because by the end of the month nobody remembers which refund was which. Read the customer's message as well as the refund reason, since "not as described" often hides a packing mistake.

4. Supplier-fault refunds and cancellations

Count orders refunded or cancelled because of something the supplier did, including out-of-stock cancellations after the customer paid. Leave out change-of-mind returns, which belong to your product choice and not to the supplier. Divide by total orders for the rate.

5. Reply time

Log how many hours pass between your message about an order and a useful human answer. This one lives in your inbox rather than your export, so jot it down as you go. A supplier who takes four days to answer can turn a fixable parcel into a chargeback.

Where to draw the line

These thresholds are rules of thumb, sized for the example margin below, not industry standards. Adjust them to your margin: a cheap accessory with a wide margin can absorb more problem orders than a high-ticket item where one refund wipes out the profit on ten sales. The delivery line below assumes you promise 3 to 8 days from dispatch.

Starting thresholds for grading a supplier. Anything between the two columns is a watch.
MetricKeep ifReplace if
Median dispatch time2 business days or lessMore than 4 business days
Delivery tail (90th percentile)Inside your promised windowMore than 3 days past it
Problem rateUnder 3%Over 6%
Supplier-fault refunds and cancellationsUnder 5%Over 10%
Reply timeWithin 1 business dayOver 3 business days
One red line is a conversation. Two or more red lines is a decision.

A worked example with three suppliers

The numbers below are illustrative, not real suppliers. Picture three suppliers selling to the same store, each with 100 orders over the last 90 days and the same 3 to 8 day promise. Dispatch is counted in business days and delivery in calendar days. Run the five numbers and compare them with the table above.

Scorecard results for three suppliers, 100 orders each.
SupplierNumbersGrade
A (US warehouse)Dispatch 1 day, tail 6 days, problems 2%, refunds 4%, replies in 1 dayKeep
BDispatch 2 days, tail 9 days, problems 6%, refunds 8%, replies in 2 daysWatch
CDispatch 4 days, tail 15 days, problems 9%, refunds 12%, replies in 4 daysReplace
B sits one day past the promised window and exactly on the problem-rate line, so it earns a watch rather than a replacement.

Supplier C is tempting because it charges $2 less per order. Put that saving against the failures. Say each order sells for $40 and costs $22 at Supplier A and $20 at Supplier C, shipping included.

Profit per 100 orders, Supplier A against Supplier C.
LineSupplier ASupplier C
Price and cost per order$40 sale, $22 cost$40 sale, $20 cost
Refunded or cancelled orders per 100412
Profit per 100 orders$1,640$1,520
Assumes every failed order is refunded in full and the supplier does not reimburse you. Payment fees, return shipping and ad spend are left out, so real losses are larger.

The arithmetic is short. Problem orders and refunds overlap, since most problem orders end in a refund or cancellation, so count each failed order once using the refund figures, 4% and 12%. A normal order earns $18 at Supplier A, so 100 orders would earn $1,800. Each failed order refunds the customer and still costs you what you paid, so it swings $40 against plan ($18 of lost profit plus the $22 cost). Four failures leave $1,640. At Supplier C a normal order earns $20, so the plan is $2,000, and twelve failures at $40 each leave $1,520. The cheaper supplier loses $120 per 100 orders before you count support time, bad reviews and customers who never come back.

What to do with each grade

Nothing needs doing for a keep grade except continued logging. Do not renegotiate a supplier that is performing, and do not move volume away from it.

A watch needs a conversation. Send the supplier your numbers, name the orders behind them and ask what changed. Set a date 30 days out to rerun the scorecard, and in the meantime shift a share of your best sellers to a backup supplier so one weak link cannot sink the month. When you ask for credit on bad orders, bring order numbers, photos and tracking. The returns workflow guide covers who does what when a return or claim comes up.

Replacing a supplier needs a plan, not a slammed door. Stop advertising that supplier's products first, keep fulfilling open orders until they clear, and bring in the replacement through a sample order so you can inspect the product yourself. Our quality control guide has an inspection checklist for samples.

If your scorecard says replace, the next job is finding suppliers who can show what they deliver. Spocket's 7-day free trial gives you access to US and EU suppliers with 2 to 7 day shipping, which is long enough to order a sample and start a fresh scorecard from the first order. Spocket is our own product.

Make it a monthly habit

Block thirty minutes on the first Monday of the month. Export the last 90 days, refresh the five numbers, and write one line per supplier in a log with the date, the numbers and your decision. Add an extra review after any event that changes how a supplier works, such as a price change, a new warehouse or the run-up to a peak sales period.

Judge trends, not single data points, and compare busy-season results with the same weeks last year, since parcels can slow down before major sales periods. Keep the thresholds honest too: loosening a line every time a favourite supplier misses it defeats the point. One bad month for a supplier with a clean history deserves a question. Three declining months in a row deserves a replacement. The log also becomes your negotiating file and your record of why you promise the delivery times you do.

Start your dropshipping business today

Start Free 7-day trial

Dropshipping Supplier Scorecard FAQs

How many orders do I need before I grade a supplier?

Fifty orders is a workable minimum for a first grade. Below that, a single bad parcel moves a rate by several points and you end up reacting to noise. For a new supplier, mark the result "too early" and keep logging.

Is average delivery time good enough?

Not on its own. The average hides the slow parcels that generate disputes and reviews. Track the 90th percentile as well, which shows how long the slowest one order in ten takes to arrive.

What if my fulfilment app marks orders as fulfilled automatically?

Then Fulfilled at may show when the order was placed with the supplier, not when it shipped. Use the first carrier scan on the tracking number as your dispatch date, and spot-check a handful of orders by hand each month.

Should I drop a supplier after one bad order?

Usually not. One failure is a reason to ask what happened and to log it. A pattern across several weeks, or two or more metrics past your replace line, is what justifies a switch.

No items found.

Launch your dropshipping business now!

Start free trial
Table of Contents

Start your dropshipping business today.

Start for FREE
7 day trial
Cancel anytime

Start dropshipping

20M+ Product Catalog
Winning Products
AliExpress Dropshipping
AI Store Creation
Get Started — It’s FREE
Start dropshipping with Spocket