AAllocator DeskDiligence Memory

All Insights

Diligence

Operational Due Diligence for LPs: A Practical ODD Framework That Catches Real Risk

Investment diligence tells you whether a strategy can win. Operational due diligence tells you whether the firm behind it can be trusted to hold your capital. Here is how leading LPs run ODD without turning it into a checklist exercise.

Matthew S. · Founder, Allocator Desk · June 2026 · 10 min read

Most blow-ups in private markets are not investment losses. They are operational failures dressed up as bad luck: a mispriced NAV, a side pocket nobody asked about, a CFO who left two quarters before the restatement, a fund administrator that quietly changed valuation policy, a key person who was never really key. Investment due diligence (IDD) decides whether the strategy is sound. Operational due diligence (ODD) decides whether the firm can actually deliver it without losing your capital to something stupid.

For LPs writing institutional checks, ODD is no longer a final-stage formality. It is a parallel workstream with its own evidence, its own opinions, and in many shops, its own veto. This guide lays out how to run it.

What ODD Is Actually Trying to Find

ODD is not a compliance scan. A compliance scan asks whether a manager checked the right boxes. ODD asks whether the firm, as a going concern, can be trusted to safeguard capital, value assets honestly, and behave the same way in year seven that it did in year one.

The questions worth asking sit in five buckets:

1. Control environment. Who can move money, who approves it, and how easy is it to bypass the policy? 2. Valuation integrity. Who marks the book, who reviews the marks, and what changes when results are bad? 3. Service provider quality. Are the auditor, administrator, custodian, and counsel actually independent and capable, or are they names on a page? 4. Key person and continuity. What happens to your fund if two named partners walk out on the same day? 5. Conflicts and governance. Where are the LP interests and the GP economics misaligned, and how is that resolved when it matters?

If your ODD process produces a green light without an opinion on each of these, it is theater.

When ODD Should Start

The classic mistake is to run ODD after the investment committee has fallen in love. By that point, the work is half-political. The institutions that catch real issues start ODD early, in parallel with IDD, and they let ODD findings actually pause a deal.

A reasonable rhythm:

  • **Screening stage:** A 30-minute ODD red-flag review. No deep work, just whether anything in the public record, prior fund DDQs, or service provider list looks off.
  • **Mid-diligence:** Full DDQ review, document requests, and a first ODD call with the COO or CFO.
  • **Pre-IC:** On-site (or video) visit, sample testing, reference calls with administrators and prior auditors.
  • **Post-IC, pre-close:** Final legal review of side letters, MFN, and any open items that became conditions of investment.

If you wait until the week before close, you are not doing ODD. You are documenting a decision you already made. We wrote more about pre-IC discipline in our guide on building a repeatable investment committee process.

The ODD Document Request, Without the Bloat

Most ODD checklists are inherited from a consulting template and run 200 questions deep. That tells you very little because GPs answer them in their sleep. The signal lives in a much smaller set of documents read carefully:

  • The last three audited financial statements, with footnotes
  • The most recent SOC 1 Type II for the fund administrator
  • The compliance manual and code of ethics, with the date last revised
  • The valuation policy, plus the last four quarters of valuation committee minutes
  • The business continuity and cyber incident policy, with evidence of the last test
  • The firm organization chart, with tenure and reporting lines
  • The complete service provider list with engagement dates
  • Any regulatory exams, deficiency letters, or 'no action' correspondence from the last five years
  • Litigation log, including settled and dismissed matters
  • A side-by-side of the current fund LPA against the prior fund LPA, with redline

Read those carefully and ask three follow-up questions per document. You will learn more than any 200-line DDQ produces.

The Five Areas That Catch Real Risk

1. Cash Controls and Wire Authority

Ask who can initiate a wire. Ask who approves it. Ask what the dollar thresholds are. Then ask for the actual signed wire policy, and compare it to the bank's signature card.

Red flags worth chasing:

  • A single person can initiate and approve wires above a meaningful threshold
  • The CFO is also the authorized signer on the management company operating account
  • The administrator does not receive a confirmed funds flow before capital calls go out
  • Founder-era 'we trust each other' language still sits in the policy

This is the single most common place where small firms get robbed, either by an outside attacker or, more often, an insider. It is also the easiest area to test. Capital call mechanics, including who controls the flow, are covered in our capital calls guide.

2. Valuation Practice

Every GP says they value assets quarterly with input from an independent third party. The interesting question is what happens when a portfolio company misses plan.

Ask to see the valuation committee minutes for the last four quarters. Look for:

  • Whether the committee actually met or just countersigned
  • Whether marks changed materially without a stated trigger
  • Whether the third-party valuation firm signed off on every Level 3 asset, or only the comfortable ones
  • Whether the policy distinguishes between hold value, exit value, and reference transaction value

Sophisticated LPs also ask for the rollforward of unrealized marks for the top five positions across three years. If the marks only move at exit, you are not looking at a valuation process. You are looking at a smoothing exercise.

This matters because every reported metric (TVPI, NAV, unrealized IRR) is downstream of valuation. We laid out which metrics actually mean something in our benchmarking guide.

3. Service Provider Quality and Independence

A brand-name auditor is reassuring until you find out the lead partner has audited the same firm for nine years, the firm pays the auditor a meaningful share of their PE practice revenue, and the audit committee is two of the founding partners.

For each major service provider, ask:

  • How long has the engagement been in place, and how long has the current lead partner been on the account?
  • What share of the firm's overall service spend goes to this provider?
  • Has the provider been changed in the last five years? If so, why?
  • Is there a written rotation policy?

The administrator is the single most important service provider an LP should care about. They are the independent record. If the administrator is small, related-party, or recently changed, your reported NAV is doing a lot of trusting.

4. Key Person and Succession

LPAs almost always include a key person clause. Most LPs read it once at close and never again. ODD is where you stress-test it.

Ask:

  • Who are the named key persons? Are they actually the people sourcing and decisioning deals?
  • What happens to the fund if two of them leave simultaneously? Read the actual mechanic, not the summary.
  • What is the equity ownership of the management company? Is it concentrated in the named partners or has it been broadened to retain the next generation?
  • What is the deferred compensation structure for non-partners? Without one, your fund is being run by people who can leave on any Tuesday.

Succession is the slow-motion ODD risk. It rarely kills a fund. It often kills the next one. We covered this dynamic in the re-up decision guide.

5. Conflicts, Allocation, and Side Letters

Conflicts are not a problem in themselves. Unmanaged conflicts are.

The questions that actually matter:

  • How are deals allocated across funds and SMAs? Is there a written policy, and does the allocation log support it?
  • Are GP commitments funded with cash or with management fee waivers? Waivers are not commitment in the same sense.
  • What co-invest economics flow back to the fund versus to the management company or named partners?
  • How are cross-fund transactions priced and disclosed?
  • What is the firm's MFN process for side letters, and have you read the side letter tier you are entitled to elect?

Side letter discipline is one of the most under-rated ODD areas. Most LPs sign one, file it, and never reconcile it again. We wrote a full guide on side letter management for exactly this reason.

Reference Calls That Actually Tell You Something

Reference calls with administrators, auditors, and prior LPs are where ODD goes from documentary to behavioral. The same trick applies as with manager references: ask open questions, listen for the pauses, and follow tone changes.

Useful prompts:

  • Tell me about a quarter where the firm pushed back on a mark you wanted to take.
  • Has anything about the firm's operations changed in the last 18 months?
  • Are there any open items between you and the firm right now?
  • If you had to flag one thing for a new LP, what would it be?

Our piece on running reference calls that surface real risk applies almost word-for-word to ODD references.

Writing the ODD Memo

A useful ODD memo is short, opinionated, and structured around decisions, not descriptions. The structure we like:

1. Recommendation. Approve, approve with conditions, decline, or pause for more work. One sentence. 2. Top three risks. Plain language, ranked, with the evidence behind each. 3. Mitigants and conditions. What the GP has agreed to, in writing, by close. 4. Open items. What is unresolved and what would change the recommendation. 5. Appendix. Document inventory, call log, and the DDQ.

If your ODD memo reads like a description of the firm, rewrite it as a set of opinions. A useful test: would an experienced reader who has never met this GP know whether to invest after reading the first page?

Ongoing ODD After You Commit

ODD does not end at close. The firm you underwrote in 2026 is not the firm you will hold in 2031. A light-touch annual ODD refresh should cover:

  • Changes in named key persons or senior team
  • Changes in service providers
  • Updates to the valuation policy or compliance manual
  • New regulatory matters
  • Any side letter MFN events triggered by later closes
  • Material litigation or counterparty defaults

A serviceable cadence is annual self-certification from the GP, with a deeper refresh every three years or at the start of any re-up evaluation. We described how to embed this into recurring portfolio reviews.

Common ODD Anti-Patterns

A short list of things experienced LPs have learned to push back on:

  • **Outsourced ODD to a single vendor with no internal opinion.** Useful as a complement, dangerous as a substitute.
  • **DDQ-driven ODD with no on-site or live conversation.** Documents do not show you culture or candor.
  • **ODD that only runs on new managers.** Re-ups are exactly where firm risk has had time to compound.
  • **Treating green ODD as a permanent state.** A clean ODD in 2024 is not a clean ODD in 2027.
  • **No written ODD policy.** If your process is in someone's head, you do not have a process.

What ODD Buys You

A well-run ODD program will rarely produce a dramatic veto. What it does, quietly, is shape the manager set you actually own. Over a decade, that is the difference between a portfolio that compounds and a portfolio that has one ugly footnote you spend a lot of time explaining.

The institutions that take ODD seriously do not have flashier returns than those that do not. They have fewer surprises. In private markets, fewer surprises is the same as better returns, just measured over a longer window.

Topics

  • ODD
  • Due Diligence
  • LP Operations
  • Governance
  • Risk

Continue reading