What is an electronic batch record (EBR)?

An EBR is the digital, executed record of one batch: materials and lots, steps with timestamps and signatures, process values, in-process checks, and deviations, captured as the batch runs against the approved master record. Unlike paper, it enforces sequence and limits during execution rather than documenting them after.

What is the difference between a master batch record and a batch production record?

The master batch record is the approved template for a product: steps, limits, and required checks (21 CFR 211.186 for drugs). The batch production record is the completed instance for one specific batch (211.188). An EBR system stores the master electronically and executes each batch against it.

What is review by exception in batch record review?

It means QA reviews only flagged exceptions, deviations, overrides, out-of-tolerance entries, because the system verified everything else at capture. It cuts review labor and release time while concentrating human judgment where it matters. A person still reviews every exception and still owns the release decision.

Do electronic batch records need to be 21 CFR Part 11 compliant?

Where a predicate rule requires the record and you keep it electronically, yes: unique logins, computer-generated audit trails, bound electronic signatures, and risk-based validation. Note there is no FDA certification for Part 11; compliance is a property of your implemented, validated system, not a product label.

Do food and supplement plants need batch records?

Batch production records are required for dietary supplements under 21 CFR Part 111, and food plants under FSMA keep monitoring, verification, and traceability records that serve the same role. The paper pain and the digital payoff, enforcement at capture and faster review, are the same as in pharma.

How long does it take to implement an EBR?

It depends on product count and validation scope, which is why credible timelines are per-product, not per-plant: one high-volume product digitized faithfully, a short parallel run, then cutover, repeated across the portfolio. Plants get in trouble when they attempt every product and a process redesign in one program.

Electronic Batch Records: A Field Guide

The batch record is the story of how a batch was made. On paper, someone reads every page. Electronically, someone reads only the interesting pages.

An electronic batch record is the digital version of a batch production record: it captures every material, step, process value, in-process check, deviation, and signature as the batch runs, enforcing the approved recipe in real time. Unlike paper, it can block out-of-sequence steps, pre-fill machine data, and support review by exception.

Batch records exist to answer one question with evidence: was this batch made the way the approved process says it must be made? Paper answers that question slowly, after the fact, at the cost of enormous review labor. This guide covers how a batch record is structured, what review by exception actually changes, where 21 CFR Part 11 fits, and a realistic migration path. It is part of our paperless manufacturing series.

What is an electronic batch record?

An electronic batch record (EBR) is the executed, digital instance of a master batch record for one specific batch. The distinction matters: the master batch record (MBR) is the approved template, the recipe, steps, limits, and required checks for a product. The batch production record (BPR) is the filled-out copy for batch 4711 specifically. In FDA's drug GMPs those are two separate requirements: master records under 21 CFR 211.186 and batch records under 211.188. An EBR system holds the master electronically and executes each batch against it.

Execution is what separates an EBR from a scanned form. Because the system knows the approved sequence, it can refuse to present step 12 before step 11 is signed, require a second signature exactly where the master demands one, reject a charge quantity outside tolerance at entry, and attach machine data to the step where it belongs. Paper describes the process; an EBR enforces it. The same logic applies beyond pharma: food, beverage, and supplement plants running batch production under their own GMP rules keep batch records too, with the same review pain.

The batch record is the batch's biography. The release decision is only as fast as the review of everything above it.

What is review by exception?

Review by exception means quality reviews only the parts of a batch record where something deviated, because the system attests that everything else executed within limits. On paper, QA reads every page of every batch record, and most pages say the same thing: normal, normal, normal. The reading exists to find the three entries that are not normal. An EBR inverts this: every in-limit, in-sequence, properly signed entry is verified at capture, so the review queue contains only exceptions, the late signature, the out-of-tolerance charge, the deviation and its resolution.

This is where EBRs pay for themselves. Review labor drops because the reading drops; release time shortens because the batch no longer waits for a page-by-page read; and review quality improves because attention concentrates on the entries that deserve it. It also changes behavior upstream: when a flagged exception visibly delays release, the plant gets serious about right-first-time execution instead of fixing records at review. To be clear about scope: review by exception does not remove the quality decision, it removes the searching. A person still judges every exception and still owns the release.

Review by exception concentrates human attention on the entries that need judgment.

Where does 21 CFR Part 11 fit?

Part 11 sets the conditions an electronic batch record must meet wherever a predicate rule requires the record: it does not add new record requirements of its own. The predicate rules, 211.186 and 211.188 for drugs, or the applicable GMPs for supplements and food, define what must be recorded. Part 11 defines what makes the electronic version trustworthy. The load-bearing requirements:

Access controls: unique logins, so every entry is attributable to one person.
Audit trails: computer-generated, time-stamped records of who changed what and when, with original values preserved.
Electronic signatures: bound to their records, with the signer's name, date and time, and meaning (performed, verified, approved) displayed.
Validation: evidence the system does what it claims, scaled by risk per FDA's 2003 scope-and-application guidance.

Two more anchors are worth reading before an implementation: FDA's 2018 data integrity Q&A guidance, which explains ALCOA expectations in plain terms, and EU GMP Annex 11 in EudraLex Volume 4 if you ship into Europe. Our GMP compliance and Part 11 guides go deeper on both.

How do you move from paper batch records to an EBR?

Move one product at a time, and digitize the record you have before improving it. A migration path that survives contact with a real plant:

Pick one product with volume. A frequently run, stable product gives you enough batches to prove the system quickly. Do not start with your most complex recipe.
Transcribe the master record faithfully. Same steps, same limits, same checks. Resist re-engineering the process during digitization; every change you make is a change you must explain and validate.
Wire in the enforcement. Sequence control, tolerance checks at entry, required signatures, and line clearance as a gating step before the run.
Connect machine data. Temperatures, times, and quantities flow into the steps that need them, replacing manual transcription, and manual second checks of transcription, entirely.
Run parallel briefly, then commit. A few batches on both paper and EBR to shake out the master record, then a clean cutover for that product. Indefinite parallel running doubles work and proves nothing new.
Turn on review by exception once trust is earned. Start with full review of EBR batches, measure how often review finds problems the system missed, and narrow to exceptions as that number goes to zero.
Repeat per product, fastest movers first.

The traps are predictable. Over-customizing the master record template per product multiplies validation work; keep one structure and vary the content. Treating the EBR as a documentation project instead of an execution system leaves you with expensive digital paper, records that look better but still enforce nothing. And skipping operator involvement guarantees a system that fights the way the floor actually works, which surfaces later as workarounds, and workarounds in a batch record are deviations waiting to be found by an investigator instead of by you.

How does an AI-native MES change batch records?

An AI-native MES compiles the batch record from what actually happened, instead of asking operators to narrate the batch by hand. Because Harmony AI connects machines, software, and paperwork in one system, most of a batch's story is already known: when the mixer started and stopped, what the temperature actually was, which lots were scanned, which checks were completed. AI agents assemble that into the executed record as the batch runs, attach the evidence to the right steps, flag exceptions the moment they occur rather than at review, and draft the deviation summary with its machine context already gathered, feeding CAPA when one is warranted. Operators sign for what they did and judged; they stop transcribing what sensors already measured. Lot-level traceability falls out as a by-product, because materials, machines, and steps are linked at capture rather than reconstructed afterward.

The implementation model matters as much as the software: we come on-site, in person, walk your process, and digitize your existing batch record in place, then connect machines and add the automation. No rip-and-replace, and your validated process stays your process. See the plant-floor progression in our CLS case study, and use our ROI calculators to put your own numbers on review hours and release-time value before you commit to anything.