The adjusted earnings figure in the book is a claim. The buyer hires someone to break it.

What is a CIM for?

To get a specific number out of a specific buyer, and then to hold that number through diligence. Everything else in the document is in service of one of those two things, and any page that serves neither is padding.

The first job is narrow. A buyer who has signed a non-disclosure agreement has already decided the business is worth an hour. The book has to give them enough to submit an indication of interest with a range and a set of assumptions attached, which means enough financial history to model, enough about the market to size it, and enough about the customer base to know whether their credit committee or investment committee will engage at all.

The second job is the one most books are not written for. Every material claim in the memorandum becomes a diligence workstream. An adjusted earnings figure becomes a quality of earnings scope. A pipeline chart becomes a customer reference call. A statement about contract renewal rates becomes a contract review. The book does not survive by being persuasive; it survives by being re-performable.

The standard the current market applies to that second job is concrete. Sellers should prepare at least thirty-six months of clean, normalized monthly financial statements and commission a quality of earnings report early rather than in response to a buyer's findings; firms that closed successfully were described as having well-prepared financial packages that helped minimize re-trading risk (Capstone Partners, Capital Markets Update, 4 June 2026). Thirty-six months of monthlies is also the dataset from which the book's own charts should be built, which is the second reason to have it before drafting starts.

What belongs in it, section by section?

Eight sections do the work, and each of them carries one judgement call that decides whether the section helps or hurts. The anatomy is in the table below; the reasoning is here.

The executive summary is not a summary. It is the investment thesis stated in the buyer's language, and the buyer's language differs by buyer type. A strategic acquirer is reading for what it cannot build. A platform company doing an add-on is reading for what integrates. An independent sponsor is reading for what a lender will finance. A book written in one dialect gets one class of bidder, and buyer composition has moved: 885 add-ons made up roughly three quarters of all US buyout transactions in the second quarter of 2026 against 289 platform deals (PitchBook, Q2 2026 US PE Breakdown, 6 July 2026).

The financial section is where books lose credibility. It should present statutory results, then the bridge to adjusted earnings, then the monthly detail behind it, in that order. The bridge is the document. A single table that walks from reported EBITDA to adjusted EBITDA with each adjustment named, quantified, dated and supported is worth more than any narrative, because it is the exact artefact a buyer's accountants will rebuild.

The market section is the one most commonly outsourced and most commonly wasted. Buyers do not need a market size they can look up. They need the definition of the market the business actually competes in, the share it holds of that narrower definition, and the evidence for both. A twelve-page industry primer with no company-specific inference tells a reader that the seller has nothing company-specific to say.

How much financial detail goes in before an NDA?

Enough to disqualify, and no more. The teaser exists to let a buyer decide it is not interested, which is the most valuable thing it can do for a seller with a limited number of hours. That means a size band rather than a number, a sector description rather than a name, a revenue model, a geography, and the transaction type contemplated.

The temptation to put the actual earnings figure in the teaser is strong and usually wrong at the lower end of the market, where the combination of sector, geography and size band identifies the company to anyone in it. Confidentiality is not an abstract concern for an owner whose staff, customers and lenders do not know the business is for sale.

Inside the book, the useful test for any disclosure is whether withholding it would cause the buyer to price risk they cannot see. A buyer who cannot see customer concentration assumes a worse concentration than the real one, because the buyers who read the most books have learned what an omission means. Withholding does not remove the discount; it converts a known discount into an unknown one, and unknown ones price worse.

How do you present concentration and adjusted earnings?

Concentration first, because it decides the buyer list rather than the price. Many institutional acquirers hold hard internal rules on customer concentration and decline above the threshold rather than bidding lower, which means the disclosure is not a negotiation over a discount, it is a filter over who is in the process at all. The right presentation is the full picture: share of revenue by customer for three years, contract length and renewal terms, the tenure of the relationship, the substitutability of the business to that customer, and any concentration already reducing. A book that discloses concentration alongside its mitigation gets a bid. A book that buries it gets a re-trade when diligence finds it.

Adjusted earnings second, and the discipline here is asymmetric. Every adjustment a seller proposes will be tested; every adjustment a seller omits stays omitted. The practical consequence is that the marginal add-back is expensive. One unsupportable adjustment in the bridge invites the buyer's accountants to treat the whole schedule as a negotiating position rather than a calculation, and the cost of that is not the rejected adjustment, it is the credibility of the ones that were correct.

The bridge should therefore be built to a standard the seller would accept if they were buying: each adjustment named, quantified by period, tied to a specific document, and categorized as non-recurring, owner-related, normalization or pro forma. Pro forma adjustments for cost savings the buyer will realize belong in a clearly separated section, not inside the historical bridge. They are an argument about the future, and mixing them into the history is what makes a whole schedule look like advocacy.

One more presentation point, because it costs nothing. Where a figure in the book differs from the statutory accounts, say so on the page it appears rather than in a footnote at the back. Buyers find the difference either way. Finding it in the book is a reconciliation. Finding it in diligence is a discrepancy.

What quietly kills a process?

Four things, and only one of them is about the business. The first is a forecast the company has never hit. A projection materially above the trailing trend, unsupported by a signed pipeline, does not raise the price; it tells a buyer that management's forecasting is unreliable, which then gets applied to every other number in the book.

The second is a workstream the book does not address at all. Fifty-one percent of senior investment bank executives now call technology diligence the single most burdensome element of the entire review and 84 percent anticipate increased cybersecurity scrutiny over the next twelve to twenty-four months, while 73 percent expect diligence to become more complex over the same window (SRS Acquiom and Mergermarket, M&A due diligence study 2026, published 23 February 2026, surveying 150 senior US investment bank executives, fielded in Q4 2025). A book with no systems section concedes that calendar to the buyer.

The third is inconsistency between the book and the people. The management section names individuals and describes what they run; the management presentation then puts those individuals in a room. If the book says the business has a chief operating officer running operations independently and the meeting shows the owner answering the operational questions, the book is what gets discounted.

The fourth is timing. The document expires. A book built on financials that are three months stale by the time bidders read it invites a request for updated numbers at exactly the moment the seller wants momentum, and each refresh resets the reader's clock. Publish on a monthly close cycle and keep the underlying pack current through the process.

There is one thing nobody can tell you, and it is worth saying plainly rather than filling with an estimate. No regulator, exchange or professional body publishes a page-length norm for a confidential information memorandum, and no dataset measures which disclosures cause buyers to withdraw. Any number offered for either comes from a firm that advises on these transactions. The anatomy below is ours.