What Your MRP Exception Messages Are Actually Trying to Tell You

A material planner opens the MRP run Monday morning to 340 exception messages. Reschedule in, reschedule out, cancel, expedite. Nobody has time to read 340 of anything, so most get cleared without a second look. Three weeks later a line goes down for a part that was flagged the whole time.

Why the list gets so long

Exception messages aren't wrong, exactly. They're honest about a mismatch between the plan and reality. The trouble is the system doesn't distinguish a message that matters from one that's just noise, so they all show up with the same urgency.

• Lead times in the system don't match what suppliers actually deliver, so the system keeps "catching" problems that were baked in from the start.
• Safety stock and lot sizing rules haven't been revisited since they were set up, so routine demand swings generate exceptions that a better rule would absorb quietly.
• Engineering changes or substitute parts aren't reflected in the BOM promptly, so the system flags shortages for parts that are actually fine.
• One planner covers too many part numbers to review each message individually, so triage becomes "clear it and hope," not a real decision.

What gets missed when everything looks the same

Treating every message the same way means the two or three that actually predict a stockout get buried next to a hundred that don't matter. The planner isn't careless, the tool just hasn't told them which is which.

Where to look before automating anything

Before layering a tool on top, it's worth understanding why the list is long in the first place.

1. Pull a week of exception messages and sort by type. If one type (say, reschedule-out) dominates, that's usually a lead time or lot-size problem, not a planning-effort problem.
2. Pick the ten oldest recurring exceptions and check whether they resolved themselves without action. If most did, your thresholds are probably too sensitive.
3. Ask which exceptions actually led to a stockout or expedite in the last quarter, and work backward from there instead of from message volume.
4. Check how many part numbers one planner is responsible for. If it's in the thousands, no triage process fixes that alone.

Where automation actually helps

When it's a data fix

If most exceptions trace back to lead times or safety stock settings that no longer match reality, the fix is updating those parameters, not adding software. This is usually cheaper and faster than it sounds, and it's worth doing before anything else.

When it's worth automating

If the data is reasonably accurate and the real problem is volume, too many messages for one person to triage by hand, that's a fair case for rule-based filtering: surfacing only exceptions above a certain dollar or risk threshold, grouping related messages by part family, and routing the rest to a weekly batch review instead of a daily fire drill. It's a sorting problem more than an intelligence problem, and treating it that way keeps the fix simple.

The takeaway

A planner drowning in exception messages usually isn't undertrained or careless. The list is just too flat, every message looks equally urgent, so the ones that matter get lost in the ones that don't. Fixing the data behind the plan comes first. Filtering and automation are worth adding once that's done, not before.