Good Documentation Practices

This topic is part of the SG Systems Global Guides library for regulated manufacturing teams evaluating eBMR, MES, and QMS controls.

Updated December 2025 • good documentation practices (GDP), ALCOA+, data integrity, contemporaneous recording, corrections, audit trails, e-signatures, review by exception • Dietary Supplements (USA)

Good Documentation Practices (GDP) are the rules that make your records credible as evidence. In dietary supplement manufacturing, GDP is not “nice handwriting” or a training slide. It is the difference between a batch record that proves what happened and a batch record that merely claims what happened. GDP ensures records are created at the time of work, tied to identity (who/what/when), protected from silent changes, and structured so reviewers can trust the story without reconstructing it from memory.

Buyers searching for GDP are usually trying to close a painful gap: their shop floor execution is mostly fine, but their records are fragile. That fragility shows up in customer audits, internal investigations, supplier escalations, and “why did we do it that way?” debates after a deviation. A mature GDP program changes the culture from “make the paperwork look clean” to “make the evidence defensible.” It also unlocks the fastest ROI in eBMR/MES/QMS: review by exception, where QA reviews the outliers instead of rechecking every line item.

“If your documentation can be edited to look right, it can’t be trusted to be true.”

1) What buyers mean by Good Documentation Practices

Buyers aren’t asking for a poster that says “write neatly.” They’re asking for a system that produces audit-ready evidence at production speed. In supplements, GDP is the operating system of your compliance posture. It determines whether your batch record is defensible when something goes wrong and whether QA can release product without redoing the entire plant’s work in a conference room.

Practically, GDP means: records are created as work happens, tied to the correct lot and step, and protected from quiet edits. It also means reviewers can tell what happened without guessing. If the record is ambiguous, the investigation becomes slow and expensive. If the record is clear and controlled, most “issues” end in minutes instead of days.

2) Why GDP failures happen (and why “clean records” are a trap)

GDP failures happen because operations and documentation are often treated as separate jobs. The floor prioritizes throughput; QA prioritizes defensibility; both are correct. GDP bridges them by making defensible documentation the fastest path—not a separate after-hours chore.

Common failure modes:

• Backfilling. Operators write notes on scrap paper and enter them later. The record becomes non-contemporaneous.
• Overwriting. Values are “fixed” silently to make the record look right. Evidence is destroyed (data integrity risk).
• Ambiguous identity. “Lot 123” appears, but the record doesn’t show which container, which supplier lot, or which status applied.
• Missing meaning. A signature exists, but it’s unclear what was reviewed or what authority it represents.
• Version drift. A batch is run on an old master record or outdated SOP (document control failure).
• Paper habits in digital forms. Electronic systems allow the same shortcuts as paper (typing weights, editing after the fact, unclear timestamps).

3) ALCOA+ in plain language

ALCOA+ is the simplest checklist for evidence quality. If a record fails ALCOA+, it fails scrutiny—internal or external. Here’s ALCOA+ translated into practical shop-floor terms:

The point is not to “teach ALCOA+.” The point is to configure your system and your workflows so ALCOA+ happens automatically and exceptions are visible.

4) Which records matter most in dietary supplements

GDP applies everywhere, but payback comes from focusing on the records that drive release risk and traceability risk. In supplement manufacturing, the highest-impact record types include:

Then layer in supporting records that commonly break investigations: chain of custody, equipment calibration status (calibration status), and material status history (hold/quarantine). The more your system captures these links automatically, the less your team has to rely on memory.

5) Contemporaneous recording and time controls

Contemporaneous recording is where GDP becomes real. It sounds obvious until you watch a plant under pressure. If documentation can be done “later,” it will be done later. And once “later” is normal, you no longer know whether entries reflect reality or reconstruction.

Practical time controls that support GDP:

• Auto time-stamps for step completion, measurements, and approvals. No manual timestamps.
• Time windows for entries: a step must be recorded within a defined window unless a late-entry exception is opened.
• Hold-time linkage: if WIP is held, the hold duration and conditions are captured and visible (Hold Time Studies).
• Event ordering integrity: prevent “signing” steps out of sequence unless specifically allowed and governed.

These controls don’t slow the floor when designed correctly. They reduce rework. When people know they can’t backfill later, they document at the moment of work, and the system becomes the source of truth.

6) Identity controls: who, what, where, which lot/container

In regulated manufacturing, a number without identity is just a number. GDP requires that records tie every critical entry to identity:

• Who performed the action (unique user, no shared accounts).
• What was acted on (material ID, spec version, step ID).
• Which lot/container was used (scan-verified lot and container identity).
• Where it occurred (line, room, station, location history when relevant).

Identity failures are the root of most painful investigations: wrong lot used, wrong label issued, wrong container returned to stock, wrong version of a record executed. A strong system prevents these with scan verification and status enforcement. If a lot is on hold, it cannot be used. If a label is obsolete, it cannot be issued. This is where GDP overlaps with execution control: hard gating is a documentation practice because it prevents bad documentation from being created.

7) Corrections: how to fix errors without destroying evidence

Corrections are normal. Uncontrolled corrections are dangerous. A GDP correction rule is simple: never overwrite; always version. When a correction is made, the original entry remains visible, the corrected entry is captured as a new version, and the system records who/when/why. This is the digital equivalent of a single line-through and initial/date on paper.

Key correction rules that keep you out of trouble:

• Reason-for-change required for every correction, using controlled categories (typo, wrong selection, device error, late entry, reconciliation).
• Risk-based approvals: critical corrections require supervisor or QA approval with e-signature.
• No “cleanup later” culture: corrections are visible in the batch exception summary.
• Correction vs deviation clarity: if the correction reconstructs missing evidence or bypasses a required control, it’s a deviation (Deviation Management).

If you want the deeper workflow, see Batch Record Corrections (if/when you publish it), and align it with Audit Trail Software and Electronic Signatures (Part 11).

8) Late entries: when allowed, how governed, and how reviewed

Late entries are one of the fastest ways to lose data integrity because they break contemporaneous evidence. Sometimes they are unavoidable (system outage, network failure, device issue). GDP doesn’t pretend outages never happen. GDP defines how you handle them without lying to yourself.

A defensible late entry policy:

• Late entry flag is automatic and permanent when an entry is created outside the allowed time window.
• Reason required: outage, equipment failure, emergency interruption, safety event.
• Actual event time vs entry time: the system captures both, clearly separated.
• Approval required for late entries that affect release evidence (weights, test results, pass/fail outcomes).
• Trending: repeated late entries are a signal—either system uptime is poor or the workflow is unrealistic.

9) E-signatures and meaning: what you are signing and why

E-signatures are often implemented badly: users click “sign” like it’s a checkbox. GDP requires signature meaning. A signature must clearly represent what the signer reviewed and what authority they are exercising. For example:

• Execution sign-off: “I performed this step as instructed.”
• Verification sign-off: “I independently verified identity/quantity/result.”
• Disposition sign-off: “I approve/reject this lot based on evidence.”
• Correction approval: “I authorize this correction and acknowledge the reason.”

Signature meaning is enforced through role permissions (RBAC) and workflow design. A line operator should not be able to perform a QA disposition. A supervisor should not be able to override a critical quality gate without documented reason and authority. This is not bureaucracy. It’s how you prove control.

10) Audit trails: what must be captured, how to test them, red flags

Audit trails are the enforcement mechanism of GDP in digital systems. If you can’t rely on audit trails, you can’t rely on the record. At minimum, audit trails should capture:

• Before/after values for changes (what changed)
• User identity and timestamp (who/when)
• Reason-for-change (why)
• Approval chain for critical changes (who approved)
• Context (which lot/record/step was affected)

Practical audit trail tests (you should literally demo these with any vendor):

• Edit a critical value and confirm the original remains visible.
• Try to edit without a reason and confirm the system blocks it.
• Try to edit as an unauthorized role and confirm it is blocked.
• Export audit trail and verify it is readable, complete, and immutable.

Red flags:

• Audit trail exists but is not searchable. If you can’t find changes quickly, it won’t be used.
• Audit trail exists but is not exportable. If you can’t provide it in an audit, it’s not useful.
• Audit trail shows “field edited” but not before/after. That’s not evidence; it’s a log entry.
• Audit trail can be turned off. If audit trails are optional, they will be optional under pressure.

11) Document control and revision control for master records and SOPs

GDP is not only about what operators type. It’s also about what instructions they follow. If the wrong version of an SOP or master record is used, the batch record can be perfectly documented and still noncompliant. That’s why GDP is inseparable from document control and revision control.

Practical controls:

• Effective dating. Procedures and master records have clear effective dates; drafts are not executable.
• Obsolete control. Old versions are clearly marked and not selectable in execution systems.
• Training linkage. When a revision is approved, impacted roles must complete training before execution (tie to training matrix).
• Batch linkage. The batch record must record which version was executed.
• Change control rationale. Why the procedure changed is recorded and reviewable (change control).

These controls prevent a classic failure mode: “We updated the procedure, but the floor ran the old one for two weeks.” In digital systems, that should be impossible if configuration is correct.

12) Exceptions and deviations: when a documentation issue becomes a quality event

GDP is not about punishing documentation mistakes. It’s about deciding what matters. Some documentation issues are minor and can be corrected. Others are evidence that the process was not controlled and must be handled as a quality event.

Common “documentation issue” events that should trigger deviations:

• Missing required verification step (line clearance, label reconciliation, dual verification).
• Manual entry used where device capture is required (scale values, critical measurements).
• Out-of-sequence sign-offs that reconstruct the record.
• Repeated late entries without a valid operational cause.
• Changing identity fields (lot/container) after consumption or packaging has occurred.

When deviations occur, GDP ensures you can do impact assessment quickly: which lots were involved, which shipments were affected, and whether release should be held. This connects directly to Deviation Management, OOS Investigation, and CAPA.

13) Training and behavior: making GDP stick on the floor

GDP fails when it is taught as rules instead of as risk. People follow rules they understand. The floor doesn’t need a lecture on ALCOA+. The floor needs to understand two things:

• Why evidence matters. Records protect the team when a complaint occurs or a customer questions a release.
• What correct behavior looks like. When to stop and ask, when to open an exception, how to correct visibly.

Practical training moves that work:

• Use real examples from your own investigations (sanitized) to show how weak records create long, painful work.
• Train “how to correct” as a skill: reason-for-change, approvals, and what not to do.
• Make GDP the easiest path: scan-driven identity, forced required fields, device capture, short structured notes.
• Give supervisors coaching tools: how to respond to mistakes without creating “hide it” culture.

GDP maturity isn’t measured by how many SOPs you have. It’s measured by how consistently people behave under pressure.

14) Metrics: the GDP scorecard that actually predicts audit outcomes

You can measure GDP. The trick is measuring leading indicators, not vanity metrics. A practical GDP scorecard:

Other useful metrics:

• Missing required step rate (how often steps are skipped or forced complete).
• Deviation linkage (how many deviations are documentation-driven vs process-driven).
• Record completeness at first pass (how many batches are “ready for review” without back-and-forth).
• Audit trail review outcomes (how often QA flags suspicious edit patterns).

These metrics are how you turn GDP from policy into operational improvement and faster release. They also give you a defensible story in audits: not “we train GDP,” but “we monitor and improve evidence quality.”

15) Copy/paste demo script and selection scorecard

Use this demo script to test whether a vendor’s “GDP compliance” is real or marketing.

16) Selection pitfalls (how digital systems reintroduce paper problems)

• Editable-by-default fields. If users can overwrite values, you’ve recreated pencil-and-eraser in software.
• Audit trails without before/after. “Field changed” is not evidence.
• Meaningless e-signatures. Clicking “sign” without defining what is being signed is weak governance.
• Optional controls. If “audit mode” can be turned off, it will be turned off under pressure.
• Shared logins. If multiple people can act under one user, the record is not attributable.
• Version drift. If SOP/MMR version enforcement is weak, you’ll run batches on old instructions.
• GDP treated as training only. If controls aren’t designed into workflow, training won’t hold under stress.

17) How this maps to V5 by SG Systems Global

V5 supports Good Documentation Practices by combining controlled execution with audit-ready governance—so records are contemporaneous, attributable, and defensible without slowing production.

• Execution evidence: V5 MES supports step-level capture, scan-verified identity, and hard-gated controls that reduce manual entry risk.
• Quality governance: V5 QMS supports approvals, deviation/CAPA linkage, and exception-based review for corrections and late entries.
• Inventory enforcement: V5 WMS supports hold/quarantine enforcement and traceability so records remain linked to true material status.
• Integration: V5 Connect API supports structured connections (ERP, LIMS, devices) that reduce transcription and correction volume.
• Industry fit: Dietary Supplements Manufacturing.
• Platform view: V5 solution overview.

18) Extended FAQ

Q1. What are Good Documentation Practices (GDP)?
GDP are the rules that make records credible as evidence: attributable, contemporaneous, accurate, original, and protected from silent changes.

Q2. Is GDP only a paper concept?
No. GDP matters more in digital systems because bad platforms allow invisible edits. GDP requires audit trails, controlled corrections, and role-based approvals.

Q3. Are corrections allowed under GDP?
Yes. Corrections must preserve the original entry, capture reason-for-change, record who/when, and require approvals for high-risk changes.

Q4. What’s the fastest GDP improvement you can make?
Remove manual timestamps and silent overwrites. Enforce scan verification for identity and require reason-for-change for corrections.

Q5. Why does GDP improve release speed?
Because clean, defensible evidence enables review by exception: QA focuses on exceptions (corrections, late entries, overrides) instead of rechecking every line.

Related Reading
• Supplements Industry: Dietary Supplements Manufacturing
• Core Guides: eBMR for Supplements | Paperless Batch Records | Audit Trail Software | Electronic Signatures (Part 11) | Part 11 Readiness | Review by Exception
• Quality Workflows: Deviation Management | OOS Investigation | CAPA for Dietary Supplements
• Glossary: ALCOA+ | Data Integrity | Audit Trail (GxP) | Electronic Signatures | Document Control | Revision Control
• V5 Products: V5 Solution Overview | V5 MES | V5 QMS | V5 WMS | V5 Connect API