Several diligence engagements I've worked on came down to a single observation that wasn't in any one section of the CIM. The CIM was thorough. Each section was internally polished. The risk that mattered showed up in the inconsistency between sections, not in the content of any single one. CIMs are written to tell a story. Diligence is the work of finding the story the CIM didn't tell.
Each section of a CIM is internally consistent. The cracks live between them. Below is a stylized view of where the page-spanning arrows usually point.
The story is on the page. The risk is between the pages.
CIMs aren't dishonest. They're optimized. Each section is written to communicate that section's story, and the editing process smooths away anything that would surface tension with a different section. Four structural features of how CIMs get produced create the blind spots diligence has to find.
Polish per section. Each section is rewritten until it's internally consistent. The rough edges that would surface tension with other sections get smoothed away in the editing process, not because anyone is hiding anything, but because the editor's job is making each section read clean.
Selected emphasis. Each section emphasizes the metrics that flatter that section's story. The metrics that would complicate it are technically present but get less prominence. They live in a footnote, an appendix, or a column of a table that no one reads top to bottom.
Definitional choices. Each section uses the definitions that work best for that section. Customer counts, ARR, churn, growth: each can be defined in defensible ways that produce different numbers. The definitions don't always reconcile across the document.
Structural omission. Some risks live between functions and don't have a natural section in the CIM template. They get distributed across multiple sections in fragments, none of which are sufficient on their own. Diligence has to assemble them.
The CIM tells the truth. It just doesn't tell all of it in one place.
These are the inconsistencies I've learned to read for, in roughly the order I check them. None of them require information that isn't in the CIM. All of them require reading sections against each other rather than in sequence.
The growth chart shows expansion. The customer concentration table shows revenue concentrated in a small number of accounts. Each is internally fine. Read together, the picture changes.
The team section celebrates specific people by name. The org chart in the appendix shows those people reporting to someone who's leaving, or in roles that don't have the authority the narrative implies.
The churn cohort uses a definition that excludes the segments where the actual churn lives. The revenue trajectory looks better than the underlying customer health.
The TAM analysis assumes a customer profile. The actual customer logos imply a different market. The growth case depends on a market the company isn't currently selling into.
The unit economics work at the current scale. The headcount plan to support the growth case implies a cost structure that breaks the unit economics. This is rarely surfaced explicitly.
The product section describes what the team is building. The investment thesis depends on the team building something different. The two narratives never get cross-referenced.
A CIM crosses my desk for a growth-stage SaaS company. The pitch is an AI-forward growth story. Strong growth chart. Strong customer logos. Strong team narrative. The financial diligence comes back clean. The deal looks defensible.
DiscoveryI was reading the customer concentration appendix when I flagged the top three customers were 71 percent of revenue. The CIM didn't hide it. It was right there in the appendix. But the growth narrative on page 14 didn't acknowledge it. And two of those three customers had renewals coming up in nine months. The thesis assumed 35 percent annual growth. The customer concentration math said the thesis depended on retaining those two contracts and growing the rest of the book by 80 percent.
OutcomeThat wasn't the deal that was being sold. The CIM positioned this as an AI-forward growth story. The actual story was three customers and a renewal cliff.
Composite scenario based on common patterns in growth-stage SaaS CIMs.
Yes or no, asked of the deal team or the analyst who built the diligence file. Six cross-section reads that are easy to skip and expensive to skip. The questions are blunt by design. Hedging defeats the point. If the honest answer is yes-but, it's a no.
Systematic crack-finding isn't a list of nitpicks. It's a structured cross-reference that produces three artifacts the deal team can act on directly. Each one has to land in the IC memo, not in a separate appendix.
Each major claim in the CIM gets paired with the underlying table, footnote, or appendix entry. Where the claim and the data agree, the cross-reference confirms it. Where they don't, the gap gets named and dated.
Each crack is described in plain language, ranked by deal impact, and tagged with the question to send back to the seller. The list goes to the deal team in time for the LOI conversation, not after.
Some cracks are fatal to the thesis. Some change the price. Some are worth knowing but priced into the existing model. Each crack gets a tag, an estimated dollar impact, and a recommendation: walk, renegotiate, or proceed.
Diligence isn't reading the CIM. It's reading the document the CIM was written to obscure.
If you're running diligence and want a second set of eyes on the cracks, I'd be glad to talk. More on product due diligence for private equity.