The Sampling Problem: Why PE Firms Review 10-20% of Contracts and What They Miss

Due diligence in private equity transactions involves systematically reviewing a target company's contracts, financial records, and operations to identify risks before closing. For contract review specifically, PE diligence teams face a persistent economic problem: comprehensive manual review is too expensive for most deal sizes, so they sample.

The standard practice is to review the top 10-20% of contracts by revenue contribution or perceived materiality. The remaining 80-90% receive, at best, a cursory scan or no review at all. This sampling approach is a rational response to the economics of manual review. It is also a known risk factor that produces post-closing surprises in a meaningful percentage of transactions.

The Sampling Logic

A mid-market PE target with $50 million in revenue might have 300 contracts in its data room: 40 customer agreements, 25 vendor contracts, 60 employment agreements, 15 real property leases, 20 IP-related agreements, and 140 miscellaneous contracts (NDAs, standard services agreements, equipment leases, software licenses).

The diligence team applies a materiality filter. They review:

• Customer agreements representing 80% of revenue (maybe 15 of the 40)
• Top vendor relationships (maybe 10 of the 25)
• Executive employment agreements (maybe 8 of the 60)
• All real property leases (15, because they are relatively few)
• Key IP agreements (maybe 10 of the 20)

Total: approximately 60 contracts out of 300, or 20%.

The other 240 contracts receive a spot check at best. An associate might scan titles and parties to confirm nothing obvious was missed, but no one reads the agreements.

This is not negligence. It is triage. Reviewing all 300 contracts manually would take 225+ attorney hours and cost over $112,000 just for the initial read-through. For a deal with a $3 million legal budget, spending $150,000-$200,000 on comprehensive contract review may not be justifiable, especially when the perceived risk of the unreviewed contracts appears low.

What Lives in the Other 80%

The problem with materiality-based sampling is that it optimizes for known risks and misses unknown ones. The 240 unreviewed contracts are unreviewed precisely because they seem immaterial. But "seems immaterial" is an assessment made without reading the contract.

In practice, the unreviewed 80% frequently contains:

Change of control provisions in mid-tier customer contracts. A customer representing 3% of revenue might not clear the materiality threshold for review. But if that customer's agreement includes a change of control termination right, and if 15 similar mid-tier contracts all have the same provision, the aggregate exposure is significant. You cannot see the pattern without reviewing the contracts.

Non-standard indemnification in vendor agreements. A $200,000 annual vendor contract might not seem worth reviewing. But if the vendor's standard form includes an uncapped indemnification obligation from the buyer for IP infringement claims, that $200,000 contract creates potentially unlimited liability exposure.

Non-compete restrictions in non-executive employment agreements. Diligence teams typically review executive employment agreements but skip the other 50. Mid-level managers may have non-compete or non-solicitation restrictions that, in aggregate, constrain the buyer's operational flexibility post-closing.

Assignment restrictions with consent requirements. Contracts below the review threshold may require counterparty consent for assignment. In an asset purchase, failing to obtain consent is a breach. The buyer discovers this post-closing when the counterparty asserts their termination right.

Exclusivity clauses that conflict with the buyer's portfolio. A vendor agreement with an exclusivity provision might conflict with another portfolio company's existing relationships. Without reviewing the contract, the conflict is invisible until it becomes a commercial problem.

These are not edge cases. In our experience working with M&A deal teams, at least one material issue is found in the unreviewed contract set in roughly 40% of transactions. The issue is not always deal-breaking, but it is always something the buyer would have wanted to know before closing.

The Economics of Full Coverage

The reason PE firms sample is economic, not strategic. No diligence team would choose to review 20% if 100% were feasible within the deal's time and budget constraints.

The math for manual full coverage:

• 300 contracts x 45 minutes reading = 225 hours
• Extraction and summarization = 75 hours
• Memo and deliverable preparation = 45 hours
• Total: approximately 345 hours at $500/hour = $172,500

For a deal with $5 million in total transaction costs, $172,500 for contract review alone may be difficult to justify, especially when the perceived risk of the unreviewed contracts appears manageable.

The math for AI-assisted full coverage:

• 300 contracts processed through structured extraction: system handles reading, classification, and extraction
• Attorney review of flagged findings: 40-60 hours
• Analytical review and deliverable preparation: 20-30 hours
• Total: approximately 60-90 hours at $500/hour = $30,000-$45,000

At $30,000-$45,000, 100% coverage costs less than the manual review of the top 20% alone. The economic argument for sampling disappears.

What 100% Coverage Looks Like

Full coverage does not mean attorneys read every one of 300 contracts. It means every contract is processed, and attorney time is allocated based on what the system finds rather than on pre-review materiality assumptions.

The workflow shifts from selection-based to exception-based:

Before (sampling): The team decides which contracts to review based on revenue thresholds and perceived materiality. They read those contracts in full. They skip the rest.

After (full coverage): The system processes every contract. It extracts deal-relevant provisions from all 300 agreements. It flags non-standard terms, unusual risk provisions, and deviations from expected patterns. Attorneys review the flagged items across the entire set, spending proportionally more time on complex issues and less on standard provisions.

The result is that the same 60-90 hours of attorney time that previously covered 60 contracts now covers 300. The attorney is not reading more. They are reading smarter, guided by structured extraction that has already identified what warrants attention.

For the PE deal team, this means three things:

Risk visibility. Every contract in the data room has been analyzed. There is no blind spot in the unreviewed 80%.

Pattern detection. When all 40 customer contracts are extracted, the team can see patterns: how many have change of control provisions, how many require consent for assignment, how many have non-standard termination rights. These patterns are invisible when you only review 15.

Negotiating leverage. Issues found before signing can be addressed through reps and warranties, indemnification provisions, or purchase price adjustments. Issues found after closing are expensive surprises with limited remedies.

The question for PE firms is not whether 100% coverage is better than sampling. Everyone agrees it is. The question is whether it is economically feasible. With AI-assisted review, the answer is now yes, at a cost that is typically lower than the sampling approach it replaces.

---