A buyer at a contract manufacturer once told us about her Friday afternoons. Around 4:30, an email lands from a distributor: the connector you ordered three weeks ago is now on a 14-week backorder, here's a substitute that might work, please confirm by Tuesday. By 5pm she's pinging the engineering lead about cross-references, by 6pm she's drafting a note to production that nobody will read until Monday morning. Her ERP says the PO is "open." Production planning is still building schedules around an arrival date that's no longer real. The line that was supposed to run Wednesday is going to stop, and right now she's the only person in the company who knows.
That gap, between a supplier sending bad news and the rest of the company catching up, is where backorder management actually lives. Not in a transaction code. Not in a portal. In whatever Slack thread, hallway conversation, or 7am huddle the buyer can squeeze in before production finds out the hard way.
What Does Backorder Management Actually Cover?
Backorder management is the buyer-side workflow for handling parts that won't arrive on the original promised date. It includes confirming the new lead time, deciding whether to substitute, requesting a cross-reference from engineering, communicating the change to production planning, updating the PO and MRP system, and tracking whether the supplier actually delivers on the revised commitment.
It's not glamorous. It's not what gets pitched in procurement software demos. But it's where buyers spend an enormous chunk of their week, especially in industries where one missing component stops a finished good from shipping.
Substitutions are the harder cousin of backorders. A backorder is just "later." A substitution is "different." Different part number, sometimes a different manufacturer, occasionally a different spec that requires engineering sign-off. And every substitution needs a paper trail, because if the new part fails in six months, you're going to be the one explaining what changed and why.
Why Does Production Always Find Out Last?
Because the news travels through email. The supplier emails the buyer. The buyer emails engineering. Engineering replies a day later. The buyer emails production. Production updates the build plan. By the time everyone is on the same page, you've burned three to five days, and the production schedule has already been built around an arrival date that was never going to happen.
Most ERPs have a delivery date field, but it doesn't update itself. When a supplier slips, that field stays at the original date until a buyer manually overrides it. SAP's MD04 (the stock and requirements list) will happily tell production planning that the part is arriving on the promised date even if the supplier has already told the buyer it isn't. NetSuite's "Expected Receipt Date" works the same way. The data is only as fresh as the last time someone touched it.
There's something fundamentally broken about a workflow where the supplier already told you the truth, but the system everyone else relies on still shows the lie. The information exists. It's just trapped in the buyer's inbox.
The Substitution Decision That Eats a Half Day
A real workflow we've watched at a metal fabricator with about 200 active POs:
Distributor emails: original part on 16-week backorder, substitute available now.
Buyer pulls the original part spec from the engineering doc system.
Buyer pulls the substitute spec from the distributor's PDF.
Buyer eyeballs both for the obvious differences (length, thread pitch, plating).
If anything looks off, buyer pings the design engineer in Teams.
Engineer replies: "Need to check with the customer, that's a controlled drawing."
Buyer emails the customer's procurement contact.
Customer's procurement contact emails their engineer.
Two days later, buyer gets a yes or no back.
Buyer issues a change order, updates the PO, notes the substitution in QuickBooks, and emails production.
Production updates the build plan.
That's 11 steps for one substitution on one PO. Multiply by the 8 to 12 substitution decisions a typical buyer handles in a week and you've used up most of someone's salary on coordination work that doesn't move a single part.
What gets cut when this stacks up? Sourcing new vendors. Negotiating better pricing. Anything that looks like actual procurement strategy. The buyer becomes a full-time substitution traffic cop.
Why ERPs and Portals Don't Solve It
The pitch from most procurement tools is "supplier portal." Get the supplier to log in, update the PO, flag substitutions there. In theory it routes the news automatically. In practice, supplier adoption is the wall every portal hits. Coupa's supplier portal has working flows for backorder updates, but if your supplier is a 12-person distributor in Ohio, they're not logging in. They're going to email you. They've been emailing you for 20 years and they're not changing.
A few tools in the space pull acknowledgements and date confirmations out of email when suppliers refuse to log into a portal. That's a real improvement over the SAP-native experience for date confirmation specifically. But the substitution side stays thin. The decision tree (engineer review, customer approval, change order, production update) still happens in the buyer's inbox.
Coupa, Ivalua, and SAP Ariba are all built for indirect spend and multi-step P2P approvals. They're not really designed for the messy "this part is on backorder, here's a slightly different one, can you decide by Tuesday" workflow that direct materials buyers handle constantly. Most AI sourcing tools focus on the front end (RFQs, bid comparison, supplier discovery) and leave the post-PO substitution work alone.
The closest thing to a real fix is something that watches your inbox, recognizes a backorder or substitution email when it lands, drafts the cross-reference request to engineering, drafts the production update, and posts the revised date back into the ERP without anyone having to retype it. That's where AI buyers actually move the needle. Not on placing the PO. On everything that happens after the supplier breaks the original commitment.
What Good Backorder Management Looks Like
A few principles, based on what works at procurement teams that don't seem to be drowning:
Email stays the channel, but stops being the system of record. The supplier still emails. The buyer still receives. But an AI agent reads the message, extracts the new ETA and any substitute parts, and writes that change to the PO and the MRP system automatically. The buyer's job stops being transcription.
Substitutions trigger a checklist, not a one-off email thread. When a substitute is offered, the system pulls the original spec, pulls the substitute spec, runs a diff on the obvious dimensions, and pre-drafts the engineering review request. Engineering still has to approve, but they're not starting from scratch.
Production sees the new date the same hour the buyer does. If the supplier slips a date, the build plan should reflect it before the next planning meeting. Not three days later when someone gets around to forwarding the email.
Every substitution is logged automatically. The audit trail (who approved, when, what changed) writes itself, because the next time that part fails or the customer asks why, you don't want to be reconstructing it from memory.
This isn't a vendor pitch. It's just what the workflow looks like when the buyer isn't the bottleneck. We've seen teams cut substitution turnaround from five days to one by automating the inbox-to-ERP loop. The buyer still owns the decision. They just stop owning the typing.
How to Audit Your Own Backorder Workflow
Pick a week. Count three things:
First, count how many of your open POs slipped a date or got a substitute offer. If you don't know off the top of your head, that's a signal. Buyers who are on top of this can usually rattle off the number.
Second, the time gap. From supplier email to production schedule update. If it's more than 24 hours, something is broken. Most teams land between 48 and 96 hours. The good ones are under 12.
Third (and this is the one teams hate looking at): substitution decisions without a written approval. Count change orders against substituted line items. If the gap is large, you're carrying audit risk you can't see until a customer or auditor asks.
Fixing these isn't usually a software problem. It's an information-flow problem. The software just helps you stop being the slowest part of the loop.
FAQ
What's the difference between a backorder and a stockout?
A backorder is a confirmed order that the supplier can't fulfill on the original date but will fulfill later. The supplier acknowledges the demand. A stockout is when the supplier has zero inventory and no firm restock date. Stockouts often turn into substitution decisions because waiting isn't an option, while backorders sometimes just need a date update and a production replan.
Should we always require engineering approval on a substitution?
For controlled drawings, customer-specified parts, or anything safety-critical, yes. For commodity hardware (standard fasteners, generic resistors, off-the-shelf cabling) most teams have a pre-approved cross-reference list that lets buyers swap without a full engineering review. The fastest teams document this list once and stop revisiting it for every individual substitution.
How fast should production hear about a slip?
Same day. The longer the gap between the supplier email and the production update, the more capacity gets planned around an arrival date that isn't real. Teams that update within 4 hours rarely get caught flat-footed. Teams that update twice a week (when someone has time to consolidate) are constantly fighting fires they could have prevented.
Can ERPs handle backorder management on their own?
Not really. SAP, NetSuite, Epicor Kinetic, and QuickBooks all have fields for expected receipt dates and supplier-confirmed dates, but none of them update those fields automatically when a supplier emails a slip. The data lives in the buyer's inbox until someone retypes it. That's the gap most procurement teams are trying to close.
How do we track substitutions for compliance?
Every substitution should generate a record that includes the original part, the substitute part, the requestor, the approver, the date, and the reason. Most teams keep this in a spreadsheet or a buried ERP note field, which works until an auditor or a customer asks for it. AI agents can write this record automatically as part of the substitution workflow, which removes the "I forgot to log it" failure mode.
What's the realistic time savings from automating this?
Based on what teams we work with tell us, somewhere between 6 and 14 hours a week per buyer once the inbox-to-ERP loop is automated. Most of that comes from killing the retyping and the manual production updates. The substitution decision still takes engineering time, but the buyer's coordination overhead drops sharply.
Lumari watches the inbox where backorders and substitution offers actually land, extracts the new dates and part numbers, drafts the review request, and pushes the revised PO data back into your ERP before anyone has to retype it. If your buyers spend their week chasing slips and writing production update emails, book a quick walkthrough and we'll show you what the loop looks like when the buyer isn't the bottleneck.
Share
