Skip to content
RiskTemplates · The Daily Brief Wednesday, November 18, 2026
Wire FINRA's Spartan Capital Case: The Private-Placement Diligence File That Wasn't There NOV 17

Feature Third-Party Risk

A Data Destruction Certificate Is Only Worth the Test Behind It

Vendor data destruction certificates fail audits in predictable ways. NIST SP 800-88r2 Appendix C fields, sampling rules, and the SEC's $35M lesson.

Table of Contents

TL;DR

  • Morgan Stanley paid the SEC $35 million in 2022 for, among other things, hiring a moving and storage company with no data destruction experience to retire servers, then not monitoring the work. Thousands of drives were resold with data on them.
  • A certificate proves a job happened to a device. It does not prove coverage, and NIST’s own documentation guidance says a certificate only shows specific devices were sanitized, not that the program is sanitizing everything.
  • NIST SP 800-88r2 Appendix C lists exactly what a certificate must carry. Most vendor templates are missing serials, property numbers, tool versions, and the validation decision.
  • Certification scope, remnant inspection, sampling, and read-back are the four tests that turn a PDF into evidence.

Morgan Stanley Smith Barney had a process for retiring hardware. The process produced paperwork. What it did not produce, according to the SEC, was destruction. The firm hired a moving and storage company with no experience or expertise in data destruction services to decommission thousands of hard drives and servers, then failed for years to properly monitor the work. The moving company sold the equipment to a third party, and drives ended up on an internet auction site with customer data still on them. Roughly 15 million customers’ personally identifying information was in scope. The penalty was $35 million, settled through an SEC order on September 20, 2022 that found violations of the Safeguards and Disposal Rules under Regulation S-P (SEC Release 2022-168).

Here is the part worth sitting with: somebody at that firm could almost certainly have produced a document with a serial number on it. That’s what a vendor data destruction certificate is — a piece of paper asserting a device was sanitized. Producing one is not the control. Having a test behind it is.

The certificate is not the control. Coverage is.

NIST writes the caveat into its own guidance. SP 800-88 Revision 2, published September 2025, says in section 4.6 that the value of a certification of media disposition depends on how the organization handles media across the entire life cycle, and that if tracking breaks down anywhere before the post-sanitization destination, “sanitization records will only show that specific ISM were sanitized and not whether the organization is effectively sanitizing all ISM that have been [processed]” (NIST SP 800-88r2, §4.6).

That single sentence is the whole audit problem. A certificate answered what happened to this drive. An examiner asks how do you know these are all the drives. Two different tests, and only one of them gets documented.

The requirement doesn’t leave room to fudge it. 12 CFR Part 30, Appendix B, paragraph III.C.4 requires a bank to “develop, implement, and maintain, as part of its information security program, appropriate measures to properly dispose of customer information and consumer information,” and III.D layers service-provider oversight on top. For SEC-covered entities, 17 CFR 248.30(b)(1) requires reasonable measures to protect against unauthorized access during disposal, and (b)(2) requires written policies, procedures, and records. Both Reg S-P compliance dates are past — December 3, 2025 for larger entities, June 3, 2026 for smaller ones.

And the definition is wider than shredding. 17 CFR 248.30(b)(7) says disposal means “the discarding or abandonment of” the information or “the sale, donation, or transfer of any medium, including computer equipment, on which” it is stored. Resale is disposal. Refurbishment is disposal. A pallet of laptops donated to a school district is disposal.

Four tests, run in order

The FTC’s Disposal Rule at 16 CFR 682.3 gets more specific than most bank policies about what reasonable measures look like. Section (b)(3) lists due diligence a firm can perform on a destruction vendor: “reviewing an independent audit of the disposal company’s operations,” “requiring that the disposal company be certified by a recognized trade association or similar third party,” and “reviewing and evaluating the disposal company’s information security policies or procedures.” Note the verb pairing throughout the rule: implementing and monitoring compliance. Contracting is half the obligation.

Practically, though, the audit question lands on the certificate. So audit the certificate against a fixed standard — NIST’s Appendix C form — and refuse the ones that fail. Four tests, cheapest first.

Test 1: Device identity

A certificate that says “12 pallets, hard drives, sanitized” is a shipping receipt. NIST’s sample form asks for manufacturer, model, serial number, property number, media type, and media source (Appendix C, p. 36). You should be able to match every serial to the asset record that says the device existed and the data classification that says what was on it.

The failure mode is easy to miss because it never looks like failure. The vendor sends a list of 400 serials, your asset system has 411 devices in the retired batch, and reconciliation takes a person who knows how to query the CMDB about three hours. Most programs don’t do those three hours. There is no prompt telling anyone they should.

Reject: certificates with a quantity instead of serials; serials that don’t match the retirement batch; property numbers from another cost center.

Test 2: Method and technique against media type

“Purge” on a certificate is a category, not a technique. Section 4.5.2 of SP 800-88r2 gives the example that catches most reviewers: a sanitization operation can degauss an SSD and complete successfully while no sensitive data is sanitized at all. The technique column is where that’s visible. Degauss on spinning rust is fine. Degauss on flash needs a second look. Cryptographic erase needs the key destroyed — which is a separate artifact from the drive being wiped, and which is worthless if encryption was never turned on in the first place. That is precisely what the SEC found at MSSB: local devices had encryption capability and the firm “had failed to activate the encryption software for years.”

Reject: a method box checked with an empty technique field; CE claimed on a device whose encryption status isn’t documented; shred certificates with no particle size, given that shredding an optical disc to 50 percent larger pieces than policy allows is a real validation failure in NIST’s own example list.

Test 3: Verification and the validation decision

This is the field most templates omit, and it is the one that shows whether anyone exercised judgment. Section 4.6 requires “verification method” and the name, title, date, location, contact information, and signature of the people performing verification and validation. Appendix C carries a separate concurrence signature block — a second human who isn’t the technician.

The r2 change log is blunt about why the language matters: “the issue of trusting the vendor’s implementation of sanitization techniques for clear and purge has been addressed,” and “sanitization validation is described and focuses on checks… to see whether the attempted sanitization was effective.” For destroy methods, verification means inspecting remnants and identifying the equipment used. So the certificate should name the machine, its blade configuration or screen specification, and the person who looked.

Reject: signature blocks with a typed name and no title or date; verification recorded as “Y”; validation filled in as “Pass” with no reference to the check that produced it.

Test 4: Destination

Appendix C asks where the media went — internal reuse, external reuse, recycling facility, manufacturer, other. This is the field that decides how hard the rest of the tests need to be. Drives headed for internal reuse in an encrypted fleet need less scrutiny than drives headed to a refurbisher who resells to consumers. Destroyed outright is the only path where destination stops mattering.

Reject: “recycling” with no named downstream facility; a certificate that names a broker rather than the site that touched the media.

Where certificates actually break: the quiet parts

Three structural gaps account for most of the real-world exposure, and none of them are the vendor’s fault.

The destruction vendor is paid for throughput, not evidence. A hauler who processes 900 drives a day has a per-device labor budget. Filling in 12 fields per serial, including tool version, blows it. So in practice you get batch sheets with one signature for 500 units, and the certificate technically exists. Contract language is the only fix: field-level data requirements in the MSA with a right to reject and re-run at the vendor’s cost.

Property tracking rots before the device retires. Half the devices reaching an ITAD dock have stale asset records, no property tag, or sit on a spreadsheet a business unit maintained by hand. Sanitizing an unknown device is easy. Reconciling retirements against an authoritative inventory is the hard part, and it happens 90 days upstream of the certificate.

Certification scope gets assumed instead of checked. i-SIGMA’s NAID AAA program audits facilities with scheduled and unannounced reviews, and e-Stewards requires NAID AAA as a precondition, so the certification is a genuine signal (i-SIGMA NAID AAA). But it is issued per site and per method. A vendor certified for off-site shred at a Georgia plant is not thereby certified for on-site purge in your New Jersey data hall. If your vetting file cites a certificate without the site and the method, you have validated a logo.

A sampling program you can defend

NIST r2 pulled back from prescribing sampling: “unless explicitly required by organizational policy, elaborate sampling of an ISM’s contents (e.g., full or representative) after clear or purge sanitization techniques is not necessary” (§4.5.1). That means the rate is yours to set, which means it needs to be written down and calibrated rather than guessed.

A starting shape, for calibration against your own three-to-six-month defect history:

Verify stageWhat happensRateSignature required
Per batch, destructiveRemnant inspection against your stated particle size, with photo of the blade rig and a caliper or sieve result100% of batches viewed, minimum one photoTechnician + second-line reviewer
Per batch, non-destructiveTool completion log plus anomaly and error review100% of batch reportsTechnician + reviewer
Independent read-backRandom drives pulled before release to refurbishment, mounted, examined for accessible dataStart at 5% of released units, adjust quarterlyInternal audit or outsourced forensic
Annual control testTwo cold batches pushed through by a backup operator with no vendor hand-holding1 test per vendor per yearVendor risk owner + audit

Two anti-gaming rules to pair with the table. First, read-back samples must be drawn by the institution, not supplied by the vendor — a vendor who knows which drives are inspected will make sure those are clean. Second, track the reopened-item rate. If a batch gets rejected and re-run, the re-run certificate should carry both the original and rework references, so a defect disappears into a fresh document with no trace.

Thresholds for escalation, framed as starter defaults until you have history: any single failed remnant inspection on a destruct batch holds the batch; two failures in a month triggers a vendor-site review; any read-back hit escalates to a vendor risk committee and opens an issue against the vendor record. Calibrate upward or downward after two quarters of real numbers, and say in the policy that you did.

The artifacts to keep

For each retirement event, the file should hold: the asset retirement authorization with classification, the reconciled serial list against inventory, the vendor certificate meeting every Appendix C field, verification photos, the validation decision with the accepting reviewer’s name, chain-of-custody transfer records, and read-back results where sampled. Retention follows your records schedule and, for SEC-covered entities, the recordkeeping expectations layered under 248.30(c).

An unbroken chain here does something a certificate alone can’t: it survives the departure of the vendor, the platform, and the analyst. That matters more than it sounds, because the SEC case ran across a five-year period and the failures were process failures, not single events.

If you’re rebuilding the offboarding side of this, the exit-process mechanics are covered in the vendor offboarding walkthrough that maps these steps to the OCC expectation, and the retention side — how long records have to exist before destruction is even legal — is in the retention schedule and legal hold guide. The critical vendor exit plan piece covers which relationships need this before termination instead of after.

So What?

Two things to do this week. Pull the last five destruction certificates your institution received and score each one against the four tests. Count how many have serial numbers, a named technique, a tool version, and a concurrence signature. Then pull the retirement batch lists behind them and reconcile to inventory. Whatever percentage you can’t tie back, that’s your coverage gap, and it exists right now, independent of any future event.

The second thing is boring and it’s the actual fix: put the Appendix C field list into the ITAD contract as a schedule, with a rejection and re-run clause. A certificate you can’t match to a device isn’t evidence. It’s a claim.

The Third-Party Risk Management (TPRM) Kit includes the offboarding tracking tab and the due-diligence evidence index that hold these artifacts per relationship, including a destruction-certificate check with the field-level review built in.

FAQ

What has to be on a vendor data destruction certificate for it to be defensible?

NIST SP 800-88 Revision 2, section 4.6, lists the minimum data points, and Appendix C turns them into a form: manufacturer, model, serial number, organizationally assigned property number, media type and source, sanitization method (clear, purge, or destroy), the specific technique, the tool used including version, the verification method, and a named person with title, date, location, contact information, and signature plus a separate concurrence signature block. A certificate missing serials, property numbers, or the tool version can’t be tied to a specific device, which is the whole point.

Is a certificate of destruction from a NAID AAA vendor enough?

Certification scopes the vendor, not the job. i-SIGMA’s NAID AAA program audits facilities with both scheduled and unannounced reviews, and e-Stewards requires NAID AAA as a precondition, so it is a real signal. But certification is issued per site and per method, so you still have to confirm the certificate covers the specific facility and the specific process — on-site shred versus off-site purge — that touched your media, for the dates your drives were in the building.

Does reselling or donating old hardware count as disposal under Reg S-P?

Yes. 17 CFR 248.30(b)(7) defines disposal to include not just discarding or abandoning information but the sale, donation, or transfer of any medium, including computer equipment, on which consumer information or customer information is stored. A refurbisher reselling drives on an auction site is inside the rule, which is why resale chains need the same evidence as shredding.

How much of a batch should be spot-checked?

NIST SP 800-88r2 says elaborate sampling after clear or purge isn’t necessary unless organizational policy requires it, so the rate is your call and it needs to be written down. A workable starting point for destructive methods is per-batch remnant inspection against your stated particle size, plus an independent read-back on a random sample of refurbishment-bound drives, calibrated against your own defect rate from the prior three to six months.

◆ Need the working template?

Start with the source guide.

These answer-first guides summarize the required fields, evidence, and implementation steps behind the templates practitioners search for.

◆ Immaterial Findings · Weekly

Sharp risk & compliance insights. No fluff.

◆ FAQ

Frequently asked questions.

What has to be on a vendor data destruction certificate for it to be defensible?
NIST SP 800-88 Revision 2, section 4.6, lists the minimum data points, and Appendix C turns them into a form: manufacturer, model, serial number, organizationally assigned property number, media type and source, sanitization method (clear, purge, or destroy), the specific technique, the tool used including version, the verification method, and a named person with title, date, location, contact information, and signature plus a separate concurrence signature block. A certificate missing serials, property numbers, or the tool version cannot be tied to a specific device, which is the whole point.
Is a certificate of destruction from a NAID AAA vendor enough?
Certification scopes the vendor, not the job. i-SIGMA's NAID AAA program audits facilities with both scheduled and unannounced reviews, and e-Stewards requires NAID AAA as a precondition, so it is a real signal. But certification is issued per site and per method, so you still have to confirm the certificate covers the specific facility and the specific process (on-site shred versus off-site purge) that touched your media, for the specific dates your drives were in the building.
Does reselling or donating old hardware count as disposal under Reg S-P?
Yes. 17 CFR 248.30(b)(7) defines disposal to include not just discarding or abandoning information but the sale, donation, or transfer of any medium, including computer equipment, on which consumer information or customer information is stored. A refurbisher reselling drives on an auction site is inside the rule, which is why resale chains need the same evidence as shredding.
How much of a batch of destroyed drives should be spot-checked?
NIST SP 800-88r2 says elaborate sampling after clear or purge is not necessary unless organizational policy requires it, so the sampling rate is your call and it needs to be written down. A workable starting point for destructive methods is a per-batch remnant inspection against your stated particle size plus a periodic independent read-back of a random sample from a refurbishment-bound batch, calibrated against the last three to six months of your own defect rate.
What is the difference between verification and validation in SP 800-88r2?
Section 4.5.1 defines verification as determining the outcome of the technique used, and section 4.5.2 defines validation as deciding whether the result is good enough given the sensitivity of the data. For destroy methods, verification means inspecting the remnants and identifying the equipment used. Validation is the approve-or-reject decision that records residual risk. A certificate with a status field and no validation decision skips the step where someone actually accepted the outcome.
When did the Reg S-P disposal requirements take effect?
The SEC adopted the amendments to Regulation S-P on May 15, 2024. Compliance dates were December 3, 2025 for larger entities and June 3, 2026 for smaller ones, so both are now past. Section 248.30(b)(1) requires reasonable measures to protect against unauthorized access during disposal, and (b)(2) requires written policies and procedures plus records.
Rebecca Leung

Author

Rebecca Leung

Rebecca Leung has 8+ years of risk and compliance experience across first and second line roles at commercial banks, asset managers, and fintechs. Former management consultant advising financial institutions on risk strategy. Founder of RiskTemplates.

◆ Related framework

Third-Party Risk Management (TPRM) Kit

Tier every vendor with one documented rule, run due diligence and AI vendor reviews, find contract gaps against the 2023 Interagency Guidance, and track SOC reports, issues, concentration and exits. Excel workbook, guide, Word policy and offboarding checklist.

Immaterial Findings · Newsletter

The brief, in your inbox.

Enforcement of the week, a framework breakdown, and the prompts that are actually worth running. Delivered to your inbox. Free.