agentcocoon

Guides

What should a daily fulfillment report include?

By Drew Amato · Published

Every paid order that has not shipped, split into orders past your shipping promise and orders still inside it. Age each one in business days, group them by product and size so you can see what is actually stuck, and open with what changed since yesterday. Flag orders that cannot be packed, like missing sizes. Keep returns, cancels and holds in a separate exceptions report so they do not bury the orders that need to ship.

Since September 2026, an agent we built for our golf apparel brand has sent a fulfillment report every day. It reads Shopify, finds every paid order that has not shipped, and sends a summary by Slack and the full list by email. It only reports. It never touches an order.

This guide covers what is in that report, what each part caught, and the one setting we still have not settled.

What should be at the top of the report?

Two lines: orders past your shipping promise, and orders still inside it. Our promise is one business day, so anything older is urgent. Each line shows the same five numbers:

Section Orders Line items Value Average age Oldest
Urgent (past promise) 16 22 $3,991 10.3 business days 20 business days
New (within promise) 3 3 $488 under 1 business day 1 business day

That is our report from November 7. Value matters as much as count. On September 29 we had 14 urgent orders, but they were worth $5,711 because several were subscription boxes.

Directly under the table, the report says what changed since yesterday. On November 7 that was "urgent dropped from 18 orders to 16," and which two orders came off the list. One sentence of change is worth more than the whole table, because it tells you whether the team is catching up.

Which orders should the report include?

Paid orders that have not shipped, with a few rules so the list stays honest:

  1. Paid and unfulfilled only. Unpaid drafts and canceled orders are not fulfillment work.
  2. Skip orders placed in the last 15 minutes. They are not late yet, and leaving them out keeps the count from flickering.
  3. Age in business days. Our warehouse does not ship on weekends, so Saturday and Sunday do not count.
  4. Check held orders separately. Shopify leaves orders marked on hold out of its Unfulfilled list (Shopify Help Center), so a held order can sit unseen. Ours go to the exceptions report.
  5. Sort oldest first. Shopify sorts the Orders page by date by default, and warns that prepaid subscription orders ready to ship can drop to the bottom of the unfulfilled list (Shopify Help Center). Oldest first puts the worst problem on top.

Why group orders by product and size?

Because stuck orders cluster, and the cluster tells you the cause. On October 3, our urgent list jumped to 158 orders worth $34,086. Read as a list, that looks like a warehouse failure. Grouped by product, it was a subscription renewal batch from October 1. One variant, a men's XL box, was 41 of those orders and about $12,000.

The grouping also shows what a list hides. On November 7, the stuck items were three orders of one vest in one color and size, two of one chino size, and one bag. That usually points to stock or a size that has not arrived, not to packing speed, so it goes to whoever buys inventory.

What should the report flag as a problem?

Anything a person has to fix before the order can ship. The biggest one for us was orders missing information. During that October renewal batch, 22 urgent subscription orders had no gender or shirt size recorded, worth $5,471.56. No one could pick them until someone contacted the customer. The report called them out as their own group every day until they were resolved.

It also tracks how the queue moves:

Date Urgent orders Value
October 3 158 $34,086
October 5 171 $35,074
October 6 69 $16,896
October 7 43 $11,085

The October 5 jump was 13 Friday orders crossing the one-day line on Monday morning, with nothing shipped over the weekend. The report said exactly that, so no one had to ask why the number went up.

What belongs in a separate exceptions report?

Orders that are open for a reason other than shipping. When we added an 8am weekday report for these, the first run found 63 orders worth $16,137, and the oldest had been open 89 days. The top reasons were returns: approved returns with unused labels, used labels with nothing checked in, and refunded returns never closed. Seven orders were refunded but never canceled, so they still looked unfulfilled.

Each reason gets a count, an average age and an action, like "cancel the order" or "find it at the warehouse, restock, refund, close." One thing we learned: Shopify does not record "cancel requested" as a status. If customers ask to cancel by email or through a returns app, the report has to read that source too.

What can a date window get wrong?

It can hide your oldest orders. Our report only looks back 30 days to keep old test orders out. That also means a paid order that never shipped drops off the list on day 31, exactly when it is most overdue. Right now, three paid orders are off the report for that reason alone, and changing the rule is still waiting on a decision.

That matters beyond customer service. When a store states no shipping time, the FTC expects it to have a reasonable basis to ship within 30 days, and to offer the buyer a choice of accepting the delay or canceling for a prompt refund if it cannot (FTC). The better rule is to keep any paid, unshipped order on the list until it ships, and exclude test orders, $0 orders and your own orders by name instead of by date.

How should the report be delivered?

Short where people read it, complete where they can search it. Our Slack message keeps the summary, product groups and the oldest rows, and stops before it hits the message limit. The email has every line. For other reports our agents run, see What tools does an eCommerce AI agent need access to?

Common questions

What time should the report go out?

Early enough that the team can act on it the same day. Ours goes out at noon so the morning's picks are already counted, and a separate exceptions report goes out at 8am on weekdays. Pick the time your warehouse plans its day around.

Should the report count weekends?

Only if your warehouse ships on weekends. Ours does not, so ages are in business days and the urgent list looks the same on Saturday and Sunday. That stops a Monday panic over orders that only aged because of the calendar.

Who should get it?

The people who can fix what it shows. Ours goes to two people, by Slack for a quick read and by email for the full list. If the report goes to a channel no one owns, stuck orders stay stuck.

Can an AI agent build this report?

Yes, and it is a good first agent because it only reads orders and reports. Ours runs on Shopify order data every day without anyone touching it. The useful part is not the table, it is the short note on what changed and what looks wrong.

Sources

  1. Shopify Help Center: Marking your order as on hold
  2. Shopify Help Center: Managing and fulfilling subscription orders
  3. FTC: Business Guide to the Mail, Internet, or Telephone Order Merchandise Rule

About the author. Drew Amato is the founder of Mully, a golf apparel brand in Detroit, and of AgentCocoon. He builds and runs the AI agents behind Mully's own operations. Follow on X.