How to Batch Process Multiple German Delivery Notes (Lieferschein) into One Receiving Log
Batch process multiple Lieferschein deliveries from different German suppliers into one consolidated receiving log spreadsheet in a single run.
Why One-at-a-Time Lieferschein Processing Doesn't Scale
The efficiency argument for batch processing seems obvious — "do 20 at once instead of 20 one at a time." But the real cost of single-document processing is not the time it takes to type. It is the context-switching cost between formats. Every time a receiving clerk opens a new Lieferschein from a different supplier, they must reorient: where is the Lieferscheinnummer on this layout? Is the Lieferdatum in the header or the footer? Is the item quantity in the third column or the fourth? Is "Pos." the line item number or the position in the warehouse? This mental remapping takes longer than the typing itself — and it is where most data entry errors originate.
When you process Lieferscheine one at a time, you also lose the ability to cross-reference. Was delivery note 2026-478 from Supplier A a full delivery or a partial one (Teillieferung)? The answer might be on delivery note 2026-479 from the same supplier, which you processed 15 minutes ago and already closed. Batch processing keeps all documents in view simultaneously, making patterns and discrepancies visible that single-document processing hides.
The math at scale: 20 delivery notes per shift × 8 fields per note × 3 minutes per note = 480 minutes of manual data entry per week. That is one full day of a warehouse employee's time spent typing numbers that already exist on paper. At German logistics wage rates, that is approximately €400–600 per week in pure transcription cost — before accounting for error correction, discrepancy investigation, and rework.
The German Warehouse Reality: Multi-Supplier, Multi-Format, Multi-Urgency
A typical morning at a German manufacturing or distribution warehouse receiving dock looks like this:
07:30 — An SAP-generated Lieferschein arrives with a truck from a large automotive components supplier. The PDF is a dense multi-page document with 8-column item tables, material numbers (Materialnummer) that are 12-digit internal codes, and a separate Lieferscheinnummer and Versandnummer (shipment number). The delivery dock needs the item-level data in the WMS by 08:15 so putaway can begin.
08:00 — Two Lexware Lieferscheine arrive from regional Handwerksbetriebe. These are clean single-page A4 documents with 4-column tables. They carry the same categories of information as the SAP document — Lieferscheinnummer, Lieferdatum, Artikel, Menge — but in a layout that shares zero pixel coordinates with the SAP format.
08:30 — A paper Lieferschein arrives in a driver's hand, filled out by a small Spediteur using a Durchschreibesatz (carbon-copy pad). The Lieferscheinnummer is handwritten in the top-right corner. The item table is a hand-drawn grid. The driver needs a signature so they can leave.
Three format generations — digital structured (SAP), digital simple (Lexware), paper (handwritten) — all within one hour, all needing to end up in the same WMS before the morning shift ends. This is not a hypothetical edge case. This is daily operational reality in German logistics, where 95,000+ companies use GS1 standards but the majority of SME suppliers have not adopted structured EDI messaging like DESADV. The gap between the structured future and the PDF/paper present is where batch extraction lives.
The receiving dock is not a data entry station. It is a verification checkpoint. When manual transcription consumes the time meant for inspection, the entire quality control function of receiving collapses into "open box, count roughly, sign paper."
How Batch Extraction Handles Format Variability at Scale
The core capability that makes batch Lieferschein processing viable across formats is semantic field matching. You define a set of column names once — "Lieferscheinnummer," "Lieferdatum," "Bestellnummer," "Artikel," "Menge," "Einheit" — and the AI applies the same column definitions to every document in the batch, regardless of format. This is the critical difference from template-based batch processing, which requires either identical document layouts (unrealistic) or per-supplier template configuration (unscalable).
For a deeper walkthrough of how semantic extraction works on individual German Lieferscheine — including field definitions, Computed Columns for discrepancy checking, and handling edge cases like handwritten annotations — see our guide to extracting German delivery note data to Excel. This article focuses on what changes when you scale from one document to a full shift's worth.
Batch processing introduces three additional requirements beyond single-document extraction:
Row-level document traceability. In a batch output, every row in the spreadsheet must identify which Lieferschein it came from. The Lieferscheinnummer column serves as the primary key. Without it, you have a pile of line items with no way to trace a discrepancy back to the document that caused it.
Consistent null handling. If a Lexware Lieferschein includes a Chargennummer (batch/lot number) column and a paper Lieferschein does not, the output column "Chargennummer" should show blank cells for the paper rows — not zeros, not errors, not guesses. Blank cells carry information (this supplier does not track batch numbers). Filled-in zeros would be misinformation.
Output row ordering. The output should preserve the logical order: rows grouped by Lieferschein, within each Lieferschein ordered by Positionsnummer. If you process 20 documents, the first 3 rows of your output should correspond to the 3 line items on the first Lieferschein, not a jumbled interleaving of line items from multiple documents.
Batch extraction is not just single-document extraction done 20 times faster. It is a different operation — one where cross-document consistency, traceability, and output organization matter as much as per-field accuracy.
From Receiving Dock to WMS: The Batch Workflow Step-by-Step
This workflow takes you from a stack of mixed-format Lieferscheine at the receiving dock to a structured spreadsheet ready for WMS import — with all the operational checkpoints that prevent errors from propagating downstream.
Collect and digitize all Lieferscheine. Gather every delivery note from the shift — PDFs from suppliers, email attachments, paper copies from drivers. Scan paper Lieferscheine using a document scanner or a phone camera (well-lit, flat, no shadows). Organize them in a single folder. The AI processes JPG, PNG, and PDF interchangeably in the same batch — a photo of a paper Lieferschein and a SAP-generated PDF sit side by side with no special treatment required.
Define columns once for all suppliers. Type your column names: Lieferscheinnummer, Lieferdatum, Bestellnummer, Absender, Positionsnummer, Artikelnummer, Artikelbezeichnung, Menge, Einheit. Optionally add Computed Columns — for example, "Differenz (Bestellt − Geliefert)" to flag quantity discrepancies — or an Inferred Column like "Format (options: SAP, Lexware, sevDesk, Papier)" to auto-classify each document's source system.
Upload the entire batch. Drag all 20+ files into the upload area. The batch processes as a single job — you do not need to start a new extraction for each supplier or each format. Processing speed depends on document count and page count. A batch of 20 single-page Lieferscheine typically processes in under 2 minutes.
Review the output for discrepancies before exporting. Scan the "Differenz" column if you added a Computed Column. Look for blank rows where critical fields (Lieferscheinnummer, Menge) are missing — these may indicate a document that was too degraded or had an unrecognizable layout. Spot-check the first Lieferscheinnummer from each supplier format against the original PDF. This review step catches errors before they enter your WMS.
Export to Excel or CSV. The output is one spreadsheet — one row per line item, Lieferscheinnummer repeated on each row for traceability, all dates normalized, all quantities in consistent numeric format. Save as XLSX for direct use or CSV for WMS import.
Import into your WMS or ERP. SAP Extended Warehouse Management (EWM), Lexware Warenwirtschaft, and most modern WMS platforms support XLSX or CSV import for goods receipt records. Map the spreadsheet columns to your system's receiving fields (Lieferscheinnummer → delivery note reference, Bestellnummer → PO reference, Menge → received quantity) and import. The extraction layer produces structured data; the integration layer connects it to your specific system. If your WMS supports file-watch directories or API endpoints, you can automate this last step entirely.
Files are processed securely and not stored.
Naming Conventions and Output Organization
Batch processing produces more data, which means data organization matters more. Here are the conventions that keep a 200-row spreadsheet manageable:
The Lieferscheinnummer is your primary key. Every line item row must carry the delivery note number of its source document. If the AI cannot read the Lieferscheinnummer — a smudged paper slip, a PDF where the number was cropped out — that row should be flagged for manual review rather than silently assigned a null key. A Computed Column "Lieferscheinnummer vorhanden (options: Ja, Nein)" auto-flags these rows.
Group output by document, not by field. The output should list all line items from Lieferschein A, then all line items from Lieferschein B, and so on. This grouping lets you quickly scan for documents that produced unexpected output (e.g., a 40-line Lieferschein that only produced 3 rows — something went wrong with the multi-page item table). Interleaved output — row from A, row from B, row from A again — is a symptom of a processing artifact and makes reviewing the batch nearly impossible.
Add metadata columns for batch traceability. Consider adding: "Verarbeitungsdatum" (processing date — when the batch was extracted), "Dateiname" (original filename — maps output rows to source files), and "Quellformat" (source format — SAP, Lexware, Papier, et cetera). These columns do not appear on the original Lieferscheine. They are batch-level metadata that makes the output self-documenting — useful when you revisit last month's batch and need to know which files produced which rows.
Teillieferung (partial delivery) handling in batch: If supplier A sends two Lieferscheine for the same Bestellnummer (partial deliveries), both appear in the batch output with different Lieferscheinnummern but the same Bestellnummer. Filter by Bestellnummer to see all partial deliveries grouped. A Computed Column "Gesamtmenge pro Bestellung" can sum the Menge across all rows with the same Bestellnummer, so you know at a glance whether the full order has been fulfilled.
FAQ: Batch German Delivery Note Processing
Can I mix paper scans and digital PDFs in one batch?
Yes. Photographed paper Lieferscheine, scanned paper copies, and digital PDFs from any ERP system can all be uploaded in the same batch. The only requirement is that scanned or photographed documents are well-lit and in focus — the same quality standard you would apply to any document you intend to read. If a paper Lieferschein is folded, creased, or partially obscured, extraction accuracy drops on the affected areas, but the rest of the document processes normally.
How does batch extraction handle multi-page Lieferscheine alongside single-page ones?
The AI treats each file as a complete logical document, regardless of page count. A 3-page SAP Lieferschein with 30 line items and a 1-page Lexware Lieferschein with 5 line items sit in the same batch. The output correctly produces 30 rows for the SAP document and 5 rows for the Lexware document, with the Lieferscheinnummer repeated on every row to maintain traceability. Header fields that only appear on page 1 of a multi-page document are correctly associated with all line items across all pages.
What export formats work with SAP EWM or Lexware import?
The extraction produces XLSX (Excel), CSV, and JSON. SAP EWM supports CSV and XLSX import for goods receipt via standard transaction codes (MIGO for goods receipt, VL31N for inbound delivery). Lexware Warenwirtschaft supports CSV import. Most WMS platforms accept at minimum CSV import for receiving records. If your system requires a specific column order or header format, you can reorder columns in Excel before import — the extraction output is a standard spreadsheet, not a proprietary format.
Is there a practical limit on batch size?
There is no hard technical limit, but operational best practice recommends batching by shift or by day — typically 20 to 50 documents per batch. Larger batches (100+) are technically possible, but the review step (Step 4 in the workflow) becomes unwieldy when scanning hundreds of rows for anomalies. If you process 200 Lieferscheine per day across multiple shifts, split them into shift-level batches so that review is manageable and discrepancies can be traced to specific time windows.
Does batch extraction satisfy GoBD traceability requirements?
GoBD requires that extracted data be traceable back to the original document. The Lieferscheinnummer column serves this purpose — every row in the output carries the unique identifier of its source document. Combined with the original PDF archive (which you must retain separately under GoBD), this creates a complete audit trail: output row → Lieferscheinnummer → original PDF. For additional audit readiness, the "Dateiname" metadata column maps output rows to source filenames. Note that the extracted Excel does not replace the original PDF archive — it supplements it.
When should I process Lieferscheine individually instead of in batch?
Batch processing is ideal when: you have a collection of Lieferscheine from one shift; you want cross-document consistency in the output; you need to reconcile partial deliveries across multiple documents for the same order; and you do not need immediate per-document output. Process individually when: you have a single urgent Lieferschein that needs immediate WMS entry while the truck is waiting; you need to do a detailed review of one complex or problematic document before processing the rest; or a supplier sends a Lieferschein after the day's batch has already been processed. Batch and single-document extraction are complementary, not mutually exclusive.