Excel spreadsheets under Part 11

Published 2026-10-23 14 min read

Chromolithograph of a long folded gridded printout beside a keyboard and an amber bottle — The printout and the file it came from: which of the two is the record decides what Part 11 asks.

Short answer

Is Excel Part 11 compliant?: No software is, by itself. Part 11 binds the people who use a system, and compliance is a claim you make, with evidence, for your own use of it.
When it matters: Only when the workbook holds a record a predicate rule requires, or a signature that stands in for a handwritten one. A scratch sheet nobody relies on is not a Part 11 record.
What a file alone lacks: A secure trail of every entry and change, a check of who is typing, and a signature tied to the record. A validated, locked template closes the calculation risk, not those three.
What usually works: Lock and validate the template and treat a signed printout as the record, or move the calculation into a validated system that has the trail and signatures built in.

Every regulated lab runs on spreadsheets: assay calculations, dilution sheets, stability trends, the standard curve someone built years ago that nobody has touched since. The question an auditor asks is not whether the program is compliant. It is whether this workbook holds a record a rule requires, and if so, who can change it, what it remembers, and whose signature is on the result.

This guide answers that for a small QC or R&D lab: when a spreadsheet becomes a Part 11 record, what a spreadsheet file can and cannot do against the rule, what FDA itself has written about spreadsheets, three workable options with their costs, and a checklist to run on each template. It sits in the validation pillar because most of the work is validating a small piece of software you wrote yourself.

When a spreadsheet becomes a Part 11 record

Part 11 does not reach every file on the lab's shared drive. It applies to records “in electronic form that are created, modified, maintained, archived, retrieved, or transmitted, under any records requirements set forth in agency regulations” (§ 11.1(b)). The requirement comes first, from a predicate rule: the drug CGMP laboratory records in § 211.194, GLP data recording in § 58.130(e), the device maker's record controls. If no rule asks for the record, the workbook is just a workbook.

FDA's 2003 guidance narrowed the scope further. When you use a computer to produce a paper printout, the paper meets every requirement of the predicate rule, and you rely on the paper to do the regulated work, FDA said “the use of computer systems in the generation of paper records would not trigger part 11.” The catch sits two paragraphs later. If you print the record but “nonetheless rely on the electronic record to perform regulated activities”, FDA “may consider you to be using the electronic record instead of the paper record”. It recommends deciding in advance, record by record, which one you rely on, and documenting that decision in an SOP or a specification.

Scroll sideways for every column

Five roles a spreadsheet plays in a GxP lab, and whether each makes it a Part 11 record
What the workbook does	Lab example	Part 11 record?
A tracker no rule asks for	Reagent reorder list, a project timeline, a training plan kept beside the real training records.	No. Not required by a predicate rule, so not a Part 11 record. Write that decision down.
A calculation template whose signed printout is the record	Assay result sheet: analyst types peak areas, the sheet calculates the result, the printout is signed and filed.	Usually not, if the printout meets the predicate rule and the lab relies on it. The calculation still has to be right: § 211.68(b) and the predicate rule's data rules apply in full.
A workbook kept and reviewed as the record	Stability data entered and trended in a workbook; the reviewer reads and approves the file on screen.	Yes. A required record kept in electronic form in place of paper. The § 11.10 controls apply to the file.
Paper and file both kept, the file relied on	The printout goes in the binder, but trending, investigations and the next calculation run from the file.	Likely yes. FDA says it may look at what you actually do, not what the SOP calls the record.
A sheet carrying sign-offs	Reviewer and approver type their names into cells, or paste a signature image.	If those names stand in for signatures a rule requires, they are electronic signatures, and §§ 11.50, 11.70 and 11.100 to 11.300 apply.

Our reading of § 11.1(b) and FDA's 2003 guidance. Your predicate rule decides the record; the answer can differ by site.

Most labs have all five on the same drive. The awkward ones are the second and fourth rows, because the same file moves between them depending on habit. A template that started as “print, sign, file” becomes the record the day someone trends last year's results from the saved copies instead of the binder.

Common misreading

“We print it, so Part 11 does not apply” is only half the rule. Printing takes the Part 11 controls off the file. It does not take the predicate rule off the calculation. Drug CGMP still says input to and output from a computer system “of formulas or other records or data shall be checked for accuracy” (§ 211.68(b)). Where computerization removes data such as “calculations performed in connection with laboratory analysis”, the same paragraph asks that “a written record of the program shall be maintained along with appropriate validation data.”

To walk one record through the test, use the Part 11 applicability checker: three questions about the record and its signatures, and an answer you can paste into the decision log. The wider picture of predicate rules is in what 21 CFR Part 11 is and who it binds.

Takeaway

List every spreadsheet that touches GxP data. Beside each, write the predicate rule behind its output, whether the paper or the file is the record you rely on, and who decided. That list is your first answer when an inspector points at a workbook.

What a spreadsheet alone gives you against the rule

Once a workbook is a Part 11 record, the closed-system controls in § 11.10 apply to it, and its sign-offs fall under the signature sections. The table reads a spreadsheet file the way most labs use one: a workbook on a shared drive, opened by whoever has the folder. Programs and versions differ, so test yours against the middle column rather than taking it on trust, ours or anyone's.

Scroll sideways for every column

Part 11 requirements against what a spreadsheet file gives you on its own, and what closes each gap
Requirement	What the rule says	A spreadsheet file on its own	What usually closes it
Audit trail § 11.10(e)	“secure, computer-generated, time-stamped audit trails” of entries and actions that create, modify or delete records; “Record changes shall not obscure previously recorded information.”	A saved file holds its last state. Overtyping a cell leaves no old value behind. Some versions, on some storage, show a change or version history; test whether the user can switch it off, clear it or edit it, and whether it lasts as long as the record.	Treat the signed printout as the record, or keep the data in a system whose trail is secure and kept for the retention period.
System access § 11.10(d)	“Limiting system access to authorized individuals.”	Access is the folder's permissions and the computer log-in around the file. The file can be copied out of the folder, and the copy carries no access control of its own.	A restricted folder with named users and an access list; better, a system with individual log-ins.
Authority checks § 11.10(g)	“authority checks to ensure that only authorized individuals can use the system, electronically sign a record, … alter a record”	Sheet and cell protection limit what can be typed where. It is a setting on the file, guarded at most by a password that is often shared. It does not tell analyst from reviewer.	Locked formula cells for accuracy; roles enforced by a system for who may enter, review and approve.
Signature manifestation § 11.50	Each signed record shows “The printed name of the signer”, “The date and time when the signature was executed” and “The meaning”, on screen and on any printout.	A typed name and a typed date are text anyone can type. A certificate signature over a whole file, where your version offers one, still has to show all three items and the meaning.	Handwritten signatures on the printout, or a system that applies all three items when the person signs.
Signature linking § 11.70	Signatures linked to their records “to ensure that the signatures cannot be excised, copied, or otherwise transferred to falsify an electronic record by ordinary means”	A name or a pasted image in a cell can be copied to another file in two keystrokes. That is the ordinary means § 11.70 has in mind.	Sign on paper, or sign in a system that binds the signature to the exact record version.
Validation § 11.10(a)	“Validation of systems to ensure accuracy, reliability, consistent intended performance, and the ability to discern invalid or altered records.”	The spreadsheet program is tested by its maker. Your formulas, ranges, rounding and lookups are not. They are your code.	Verify each template against known answers before release and after every change.

Rule text: eCFR 21 CFR Part 11, read 2026-10-22. Middle and right columns: our reading; check your own program and version.

The pattern is plain once it is laid out. A locked, verified template can make the arithmetic trustworthy. It cannot make the file record who changed a number, stop a copy leaving the folder, or tie a name to a result. Those three are what an audit trail, authority checks and linked signatures are for, and what an audit trail has to capture explains why the first one carries most of the weight in an inspection.

The 2003 discretion, and its limits

FDA's 2003 guidance said it does not intend to enforce the Part 11 validation, audit trail, record retention and record copying requirements as explained there. It kept enforcing access limits, authority checks and “requirements related to electronic signatures (e.g., §§ 11.50, 11.70, 11.100, 11.200, and 11.300)”, and it said: “We will enforce all predicate rule requirements.” A spreadsheet result still has to be accurate, complete and attributable under the predicate rule, trail or no trail.

Takeaway

Hold a workbook to the six rows above before anyone calls it compliant. If the file is the record and the answer to the first row is “no trail”, the honest fixes are to make paper the record or move the data, not to add another password.

What FDA has said about spreadsheets

FDA has no guidance on spreadsheets as such. It has used them as examples, and the examples are pointed. Below are the passages we found in FDA's own documents that name a spreadsheet or a worksheet, each quoted, with what it means for a lab template. We searched the 2003 Part 11 guidance, the 2018 data integrity Q&A, the 2024 clinical investigations Q&A and the 2026 computer software assurance guidance.

Data Integrity Q&A, 2018
1.d, static and dynamic records

“It also may allow the user to modify formulas or entries in a spreadsheet used to compute test results or other information such as calculated yield.”

FDA's own example of a dynamic record is a calculating spreadsheet. A printout is static; the formulas live in the file.

Data Integrity Q&A, 2018
Q6, blank forms

“If used, blank forms (e.g., electronic worksheets, laboratory notebooks, and MPCRs) should be controlled by the quality unit or by another document control method.”

A blank calculation template is a controlled form. Issue it from a master, not from last month's filled copy.

Data Integrity Q&A, 2018
Q3, validating each workflow

“If you validate the computer system but you do not validate it for its intended use, you cannot know if your workflow runs correctly.”

FDA's example is a record that must contain the correct calculations. Testing the program is not testing your template.

Data Integrity Q&A, 2018
Q4, restricting access

“FDA recommends that you restrict the ability to alter specifications, process parameters, data, or manufacturing or testing methods by technical means where possible”

Locked formula cells are a technical means for the method. Who may change the master is the other half.

CSA guidance, 2026
Intended use of features

“if a custom formula automatically calculates time and temperature statistics to monitor the performance and suitability of the curing process, then additional validation by the manufacturer might be necessary.”

Typing readings into the same program was low risk in FDA's example. The formula is what raised it.

Two more answers in the 2018 Q&A bear on how a template is used rather than how it is built. Shared log-ins break attribution: with them “a unique individual cannot be identified through the login” (Q5). And an electronic copy counts as a true copy only if it preserves “the content and meaning of the original record, which includes all metadata” and its “static or dynamic nature” (Q9). A PDF of a calculating sheet shows the numbers and drops the formulas.

The CSA guidance also gives a full worked record for a spreadsheet: one that collects and graphs nonconformance data, judged not high process risk, tested by exploratory testing. One check failed when text was typed into a numeric field. The fix was a validation rule on that field, then a retest that passed. It is a fair model of a light, honest spreadsheet validation record. Note the scope: CSA is written for device makers' production and quality system software (issued 2026-02-03). A drug or GLP lab may borrow its reasoning, but the guidance does not bind it.

Common misreading

The 2018 Q&A is drug CGMP guidance and the CSA guidance is for devices. Neither is a rule. The rules behind them are the predicate rules and Part 11, which apply whatever program does the arithmetic.

Takeaway

Treat a calculating template as two things at once: a controlled blank form, and a small piece of software you validated for one intended use. Keep the evidence for both.

Chromolithograph of a gridded sheet with a block of amber cells and a cobalt bottle on it — Formula cells locked, inputs left open: the template is a controlled form before it is anything else.

Three ways to handle it, with the trade-offs

There is no single right answer for every workbook. The choice is per spreadsheet, by what its output decides and how often it runs. Here are the three options most QA teams end up with, written down plainly.

Option 01

Lock and validate the template

One released master per calculation, with formula cells locked, inputs limited to the type and range they should hold, a template ID and version printed on every page, and a short validation record behind each version.

Fixes: Wrong or drifting formulas, typing text into a number, copies of copies, nobody knowing which version made a result.
Leaves: The audit trail, user identity and signatures. On its own this option is a calculation control, not Part 11 compliance.
Costs: Hours per template, not weeks. Every change goes back through verification.
Fits when: It is paired with option 02, or the file is not the record at all.

Option 02

Print, sign, and make paper the record

The validated template calculates; the analyst prints, signs and dates by hand; the reviewer checks the printout against the raw data and signs. The SOP names the signed paper as the record.

Fixes: Signatures and their meaning, in ink, on a paper record. FDA's 2003 guidance accepts paper and electronic parts co-existing in a hybrid as long as the predicate rule is met and the records keep their content and meaning.
Risks: The file still exists. Print until it passes, overwrite and reprint, a hidden row, a result trended from the file: none of that shows on the paper. A printout is static and the formulas are not on it. If the lab starts relying on the file, the 2003 guidance says FDA may treat it as the record.
Costs: Cheap to start, slow to run: reconciliation, filing, and a second check that reads every input.
Fits when: Volumes are low, the calculation is simple, and nobody needs the data electronically afterwards.

Option 03

Move the calculation into a validated system

Put the calculation where the data already lives, such as a LIMS, an ELN or the instrument's data system, configured and validated for that intended use, with its own trail, roles and signatures.

Fixes: All three gaps at once: a computer-generated trail (§ 11.10(e)), roles (§ 11.10(d), (g)) and signatures linked to the record (§ 11.50, § 11.70).
Leaves: Your own validation for the configured calculation, and the system's upkeep: change control, periodic review, access reviews.
Costs: The most up front: licences, configuration, a validation pack. The least per result once it runs.
Fits when: The result is reportable, high volume, trended, or used for release, and the file is the record in practice.

Inspection risk

The pattern that hurts is a template whose failing result can be recalculated, overwritten and reprinted with only the passing page filed. Paper hides it and a file with no trail cannot show it. If your template allows it, your SOP has to forbid it in words, and your review has to look for it: printout times, page counts, template versions.

Whichever option you choose, the validation follows the same lifecycle as any other system, sized down: a requirement per calculation, a risk call, test cases with known answers, and a release signature. The steps are in computer system validation for small labs, and the CSA validation scope planner helps decide how much testing each function earns. For option 03, the LIMS comparison and the ELN comparison read systems against the same Part 11 rows as the table above.

Takeaway

Sort your spreadsheets into three piles: retire or ignore the ones no rule needs, run the simple low-volume ones as options 01 plus 02, and put the reportable, trended or release-deciding ones on the list for option 03.

Spreadsheet control checklist

This is the list we would walk through for each GxP template before its first use and at each periodic review. It is our practice, not a list FDA publishes: the line under each row names the rule or FDA passage it leans on, so you can see which rows are law and which are good habit. Mark each row, then print the sheet for the template file or copy it into your review record.

Spreadsheet

Pick one word per row. Copy gives you the list as text; Print gives you the page with this sheet.

Decide

01
Every spreadsheet that touches GxP data is listed with its owner, intended use and the predicate rule behind its output.
FDA 2003 guidance
open
02
For each one, an SOP says whether the signed printout or the file is the record the lab relies on.
FDA 2003 guidance
open
03
Each has a written risk call: what its output decides, and what a wrong result would do.
FDA 2018 Q&A, Q3
open

Build

04
One released master per template, under document control, with a template ID and version printed on every page.
FDA 2018 Q&A, Q6
open
05
Formula and reference cells are locked; input cells accept only the type and range they should hold.
FDA 2018 Q&A, Q4
open
06
Each version is verified against known answers, including limits, rounding, blank cells and wrong-type entries, and the verification record is signed.
§ 211.68(b)
open
07
The master is read-only to users; only its named owner changes it, through change control and re-verification.
§ 11.10(d), FDA 2018 Q&A, Q4
open

Use

08
Every run starts from the released master, never from a filled copy of an earlier run.
FDA 2018 Q&A, Q6
open
09
The printout shows every input and result, the template ID and version, the file name and the print time; no hidden row or column carries data.
§ 211.194
open
10
Sign-offs are handwritten on the printout, or applied in a system that meets §§ 11.50 and 11.70. Never a typed name or a pasted image.
§ 11.50, § 11.70
open
11
No shared log-ins on the computers where templates are run or results reviewed.
FDA 2018 Q&A, Q5
open

Review

12
A second person checks inputs against the raw data and the template version, and looks for reprints of the same run.
§ 211.194(a)(8)
open
13
If the file is the record: individual log-ins, a named access list, a tested backup, and a trail that meets § 11.10(e). If it has no such trail, it should not be the record.
§ 11.10(d), (e), § 211.68(b)
open
14
Templates are reviewed on a set interval, and superseded versions are withdrawn from use.
Annex 11 § 11 (EU)
open

in place 0not yet 0not used 0open 14

Our practice, not a regulatory list. The line under each row names the rule or FDA passage it leans on.

The checklist covers the template. For one finished record, the ALCOA+ record check walks a single result sheet through attributable, legible, contemporaneous, original and accurate, plus the four attributes PIC/S lists. A sheet that passes this list but cannot say who typed a value fails the first of those.

Cheat sheet: spreadsheets under Part 11

Part 11 record?: Only if a predicate rule requires the record or the signature, and the electronic form is what you keep or rely on.
Printout relied on: Part 11 generally not triggered (FDA, 2003). The predicate rule still governs the calculation and the data.
File relied on: A Part 11 record, even if a printout sits in the binder. The § 11.10 controls apply to the file.
Calculations: Input and output of formulas checked for accuracy (§ 211.68(b)); each template version verified with known answers.
Templates: Blank electronic worksheets are controlled forms (FDA, 2018 Q&A, Q6): one released master, ID and version on every page.
Audit trail: A saved workbook keeps its last state. If the file is the record, it needs a trail that meets § 11.10(e).
Signatures: A typed name or pasted image is not a § 11.50 / § 11.70 signature. Sign the printout by hand or sign in a system.
Validation size: By risk: what the output decides. Readings typed in, little; a formula behind a reported result, real tests.

Questions

Is Excel 21 CFR Part 11 compliant?

No software is compliant on its own, and FDA certifies none. Part 11 places its controls on the people who use a system, and compliance is a claim the lab makes, with evidence, for its own use. A workbook can be part of a compliant process, usually as a validated calculation whose signed printout is the record. A workbook kept as the record has the audit trail, access and signature gaps set out in the table above.

Can we use Excel in a GMP or GLP lab at all?

Yes. No rule bans a program. What the rules ask is that the calculation is right and checked (§ 211.68(b) in drug CGMP), the blank template is a controlled form, the person entering data is identified (§ 58.130(e) in GLP), and the lab knows which version of the record it relies on.

Do we need to validate every spreadsheet?

No, but every one that calculates GxP results needs verification sized to its risk. FDA's 2018 Q&A says “The extent of validation studies should be commensurate with the risk posed by the automated system.” A tracker no rule needs gets none; a formula behind a reported result gets test cases with known answers. The method is in computer system validation for small labs.

Is a PDF of the spreadsheet enough to keep?

It can be the record when it holds every input and result and the lab relies on it. It is not a true copy of the workbook: FDA's 2018 Q&A says a true copy keeps “the static or dynamic nature” of the original, and a PDF drops the formulas. Keep the validated template version that produced it, so the calculation can be rebuilt.

Does a typed name in a cell count as an electronic signature?

If it is meant to stand in for a required signature, Part 11 treats it as one: an electronic signature is “a computer data compilation of any symbol or series of symbols executed, adopted, or authorized by an individual to be the legally binding equivalent of the individual's handwritten signature”. A typed name then fails §§ 11.50, 11.70 and 11.200. The signature manifestation check shows what § 11.50 expects.

Is a sheet protection password enough for access control?

No. It controls which cells can be typed in, not who is typing. FDA's 2018 Q&A asks labs to keep a list of the people authorised on each CGMP computer system and to give administrator rights to people “independent from those responsible for the record content”. A shared protection password does neither.

What about online spreadsheets that keep a version history?

Apply the same test. Is the history secure, computer-generated and time-stamped; can a user clear or edit it; does it keep old values readable; will it last as long as the record; does every user sign in as themselves? Then validate it and qualify the supplier like any hosted system: see Part 11 for cloud and SaaS systems.

Sources

Regulation text quoted from the eCFR; guidance quoted from FDA, the European Commission and PIC/S. General information, not regulatory advice.

21 CFR Part 11, Electronic Records; Electronic Signatures, eCFRread 2026-10-22
21 CFR 211.68, Automatic, mechanical, and electronic equipment (Part 211, CGMP for Finished Pharmaceuticals), eCFRread 2026-10-22
21 CFR Part 58, Good Laboratory Practice for Nonclinical Laboratory Studies, eCFRread 2026-10-22
Part 11, Electronic Records; Electronic Signatures: Scope and Application (2003), FDAread 2026-10-22
Data Integrity and Compliance With Drug CGMP: Questions and Answers (December 2018), PDF, FDAread 2026-10-22
Computer Software Assurance for Production and Quality Management System Software (2026-02-03), PDF, FDAread 2026-10-22
Electronic Systems, Electronic Records, and Electronic Signatures in Clinical Investigations: Q&A (2024), PDF, FDAread 2026-10-22
EudraLex Volume 4, Annex 11: Computerised Systems (2011), European Commissionread 2026-10-22
PI 041-1, Good Practices for Data Management and Integrity in Regulated GMP/GDP Environments (2021), PIC/Sread 2026-10-22