VDA EDI Tools

VDA 4984 DELFOR Errors: A Field-Level Troubleshooting Playbook

VDA EDI Tools Blog · 2026-11-13

Most DELFOR problems do not arrive as error messages - they arrive as a phone call: “your delivery is short”, “the schedule shows double demand”, “we rejected your ASN because it references an unknown order”. This playbook maps the common symptoms back to the exact segment and element that caused them, in the order a debugger would check.

Case 1 - “Your cumulative quantity is wrong”

Symptom: the customer's scheduling team reports your cumulative-received counter disagrees with theirs by a fixed offset that never shrinks.

Where to look: line-level QTY+70 (cumulative quantity ordered) and QTY+1/QTY+3 (cumulative figures in SCC groups), plus the RFF+CT contract reference that anchors the counter.

Root cause: VDA 4984 expresses much of its demand as running totals, not per-date amounts. Your ERP consumed one message twice (a redelivery after an OFTP restart is the classic), or consumed a message partially after a mapping change. The offset is now baked into every subsequent calculation.

Fix: freeze consumption, compare your last successfully reconciled cumulative figure against the customer's, agree the corrected baseline in writing, and add idempotency: store the BGM-2 schedule number of every consumed file and refuse duplicates.

Case 2 - “The forecast doubled overnight”

Symptom: MRP shows roughly 2x demand versus yesterday for the same horizon.

Where to look: the SCC segment before each schedule QTY. A line like SCC+1 means firm; SCC+4 means forecast.

Root cause: your import treated every SCC+QTY+DTM triple as additive demand, but many OEM guidelines send the same quantities at multiple commitment levels - the firm 500 for Monday and the forecast 500 for Monday are the same 500, not 1,000. Summing across SCC types double-counts by construction.

Fix: pick the strongest SCC type your customer designates as authoritative (usually SCC+1 for the firm zone, SCC+4 beyond it) per date, and ignore weaker levels for that date. Document the precedence table per customer - BMW, VW-group and Brose-style guidelines differ in exactly this point.

Case 3 - “Every schedule line is one day early”

Symptom: all dates systematically shifted; your shipments keep arriving a day before the dock expects them.

Where to look: the DTM qualifier in each SCC group: 2 (delivery date), 10 (shipment/despatch date), 64/159 (window start), 63/158 (window end).

Root cause: the mapping read DTM+10 (despatch date) where the guideline meant DTM+2 (delivery date), and the customer has one transit day. When the window qualifiers 64/159 vs 63/158 are also swapped, every window collapses to a single day or stretches by 24 hours.

Fix: build a qualifier whitelist per customer and reject anything outside it at import time, rather than silently treating unknown qualifiers as “probably the date”.

Case 4 - “Message rejected: unknown ship-to”

Symptom: the OEM gateway bounces the file or bookings land in a manual queue.

Where to look: NAD+ST and the LOC segments with qualifiers 11 (unload point), 159 (dock), 172 (line), and the NAD+ST internal ID vs GLN ( qualifier 9 vs 92).

Root cause: LOC codes that are not in the plant master data - often copied from an older guideline revision, or a GLN sent where the guideline requires the internal DUNS-style code (or vice versa).

Fix: treat NAD/LOC code lists as master data: version them, diff them against each new guideline release, and validate at import against the list your onboarding contact confirmed in writing.

Case 5 - “Parser drops the whole file”

Symptom: your translator reports a syntax error on a file that “looks fine” in a text editor.

Where to look: the envelope first: UNB-1 syntax level (UNOA restricts character set - a stray umlaut or § breaks it), UNH-2 version (D:04A vs D:07A vs D:96A element positions differ), UNT-1 count, and any unescaped separator (release character ? missing before a literal + or ' in an IMD description).

Root cause: usually one of three: a German-locale date (13.11.2026) instead of format 102, a decimal comma in QTY, or a character outside the UNOA set.

Fix: validate the envelope and character-level rules before the semantic mapping - the VDA 4984 validator runs exactly that sequence locally, which turns a 40-minute translator spelunking session into a ten-second paste.

The meta-lesson

Every case above shares one pattern: the file parsed, so nobody looked at it, and the error surfaced days later as a logistics complaint. A structural check at receipt time - envelope, required segments, date formats, code values - is the cheapest control in the whole EDI chain, and it runs happily on every single message, not just the ones someone complains about.

Building your own first-response checklist

Adapt the cases above into a runbook your team can execute without an EDI specialist in the room. A solid template has five slots, filled in per customer:

Slot five deserves emphasis: redelivered files after network retries are the silent source of half the cumulative-quantity horror stories. A duplicate file should be a one-line log entry, not a doubling of demand. Keep the checklist next to the import job, version it with your mapping, and rehearse it once per quarter with a deliberately broken file - the ten minutes it takes beats every Monday-morning escalation.

Try it: Reproduce each failure locally: paste your DELFOR into the validator and watch the checks fire. Open the tool →

← All articles