Quality Of Earnings Model

A quality of earnings analysis is the diligence exercise that establishes what a target business actually earns, as distinct from what its accounts report. It normalises the reported result for owner-related, non-recurring and accounting items, annualises in-period changes onto a run-rate basis, tests whether the reported revenue and costs moved through the bank, and sets the working capital balance the purchase price will be adjusted against. It is performed before signing and it is about the past, not a forecast.

Because the seller and the buyer credit the adjustments differently. Every adjustment carries a diligence quality rating, and the rating selects a credit factor: full credit for fully documented items, half where the support rests on a management estimate, a quarter for pro-forma claims that cannot be verified from the records. The seller's run-rate EBITDA of 10,928 assumes every adjustment is credited in full; the buyer's 10,036 applies the factors. The 892 difference is the negotiating range.

Because a buyer does not discount bad news. If a diligence provider finds an unrecorded liability or a deferred revenue cut-off error, the buyer takes the full deduction regardless of how well it is documented; it is only the seller's proposed addbacks that get haircut for weak support. The model implements that asymmetry with a guard on the credited column, which is why the buyer's total falls by 892 rather than by half of everything.

They make different claims. A normalising adjustment says the reported number was wrong - the owner was overpaid, the ERP project was a one-off, the deferred revenue was cut off incorrectly. A run-rate normalisation concedes the reported number was right and asks the buyer to pay as though an in-period change had applied all year, such as a price increase that took effect in month eight. The model keeps them in separate blocks and separate subtotals, and most of the quality haircut correctly lands on the second.

As the twelve-month average of a normalised balance, not the reported one. Deferred revenue and the bonus accrual are lifted out of working capital because they are taken as debt-like items in the price bridge, and an aged receivable reserve and an unrecorded liability are pushed in. The peg lands at 7,538 against a closing normalised balance of 9,173, so the base case carries a 1,635 true-up in the seller's favour - a consequence of receivable days stretching from 58 to 72 across the period.

Because it is not supposed to. The reconciliation identity itself ties to zero in every month, and the model carries a check row proving it. What is left over after the receivable, payable, inventory, prepaid, accrual and deferred revenue movements are taken out is the unexplained variance, and on the shipped case that is 85 across the year - 0.14% of revenue - with 86 of it concentrated in the final month. A variance sitting in one month is a finding; one spread evenly is usually noise.

Yes. The entire reported trial balance is an input block on the assumptions sheet: twelve months of revenue by stream, cost of sales, five operating expense lines, six working capital balances, an aged receivable reserve, and monthly bank receipts and disbursements, plus real opening balances. Paste your target's figures over them and the bridge, the proof of cash, the peg and both price columns recompute without touching a formula.

An EBITDA bridge decomposes a movement in earnings between two periods into volume, price and mix. A quality of earnings review decomposes a single period's earnings between what was reported and what is defensible. Different axis entirely: one asks why this year differs from last year, the other asks whether this year's number is real.