Build a QoE bridge a buyer can challenge

A Quality of Earnings bridge explains the transition from accounting profit to a measurement defined for the transaction. I would start by asking what this measure should represent: comparable historical performance, normalized result, or future operating scenario. These readings can be useful, but they are not interchangeable. The presentation becomes misleading when acquirer synergies are mixed with accounting corrections indiscriminately.

The final figure is not enough. You need to understand revenue stability, cost structure, cash conversion and accounting choices. A business can have high adjusted EBITDA while relying on one customer, NWC funding, or weak assumptions. I would read the bridge with the operational and balance sheet analyses, not as stand-alone proof of the company's value.

I would reconcile the EBITDA presented with the accounts, general ledger and reporting, explaining the differences in scope, currency and classification. The EBITDA formula used must be written: it is not a subtotal defined universally by all benchmarks. Consolidation entries and central costs must remain identifiable. Without a reconciled foundation, a precise bridge can give an illusion of control.

Each adjustment must retain its gross amount, its already recorded effect and its net impact. A credit note or an accounting adjustment of value or provision for charges may modify the candidate. The signs must be explicit: adding an eliminated charge and deducting a missing recurring cost are two different operations. I would also control the links with debt-like and NWC to avoid the same bond affecting the price twice.

A cut-off error may call for a period correction; an exceptional cost may justify standardization; a future economy can belong to a scenario. I would keep separate categories even if the final presentation is compact. Management must explain the elements, but the qualification is based on their nature, their connection and their recurrence, not on the “one-off” label in reporting.

Symmetry matters. If a non-recurring expense is neutralized, exceptional income must also be examined. An unfunded recurring bonus may result in a reduction in profit. Searching only for add-backs mechanically pushes the bridge in one direction. I would ask a second reviewer to challenge the omissions and assumptions as well as the proposed lines.

Let's take an EBITDA of 5 M€. The supporting documents justify 180 k€ one-time transaction fees. A recurring bonus of 70 k€ was not recorded in the period. Under these assumptions, the normalized result is 5 M€ plus 180 k€ minus 70 k€, or 5,11 M€. Each line must remain connected to its proof and qualifying reasoning.

The buyer is also considering 400 k€ synergies. This amount does not enter into the historical QoE by simple addition. It can be included in a separate scenario with schedule, execution costs and implementation risks. Presenting 5,51 M€ as normalized historical EBITDA would confuse what the business has generated and what this acquirer hopes to obtain.

I would then test sensitivity to the contested lines. If the fees contain a recurring service, the candidate must be reduced. If the bonus concerns another period, the connection must be re-examined. The bridge is a discussion tool based on documented hypotheses, not an arithmetic operation that its graphic form makes indisputable.

I would give each line an identifier, a specific source, a period, a classification, an amount offered and a professional review status. Open items must be kept with the expected document. The management discussion enriches the file, but it does not replace the supporting documents. The reviewer must be able to understand why a line is accepted, rejected or presented with reservations.

A source update can modify multiple analyses. The file must identify the version retained and the dependencies. A reclassified cost can change the QoE bridge and the presentation of an obligation at closing. I would carry out a final cross-functional check before restitution, particularly when several teams are working on results, NWC and debt.

The AI can identify labels, extract amounts and suggest links between invoices and entries. I would measure rebinding errors and omissions, not just retrieval speed. A well-read invoice does not prove that a charge is exceptional. The file must allow calculations to be redone and contradictions between documents and explanations to be preserved.

To evaluate Solenor or another environment, I would ask for a candidate with an evolving source and a partially recurring cost. The value sought is a faster and more defensible professional review. The final conclusion remains that of the team, with a clear performance definition and visible limits.