The short answer
EBITDA normalisation strips one-off and non-operating effects out of earnings to show what a company sustainably earns. Because the purchase price in most transactions is a multiple of adjusted EBITDA, every adjustment is money: at a multiple of eight, an adjustment of 200,000 euros moves the price by 1.6 million.
The standard is therefore strict. What is accepted is what is one-off, non-operating or structurally changed – and can be evidenced.
A note for readers outside Germany
Two of the recurring adjustment items are specifically German. Owner-managed GmbH targets frequently pay the managing shareholder above or below market, so the adjustment runs in both directions and needs a market benchmark, not an assertion. And the German tax audit – the Betriebsprufung – arrives in multi-year cycles, so back payments and released provisions land in single years and have to be identified as out-of-period rather than operating.
What is accepted in practice
- Shareholder-related costs. Privately motivated expenses, vehicles, advisory contracts around the shareholders. Managing director remuneration is not struck but adjusted to market level – in both directions. Pay yourself too little and the difference has to be added as a cost.
- Genuine one-off effects. Severance, litigation, a relocation, a system implementation, aborted transaction costs. The condition: they demonstrably do not recur.
- Structural breaks with a contractual basis. A terminated lease, a renegotiated framework agreement, a price increase from mid-year. Such adjustments are strong when the contract is in the data room.
- Out-of-period income and expense. Released provisions, back payments from tax audits.
- Pandemic or subsidy effects, insofar as they are clearly separable and non-recurring.
What is routinely struck
The other half of the list is longer than sellers expect.
- Pro-forma effects from measures not yet implemented. Planned savings without a decision and without evidence of implementation count as planning, not as earnings. They belong in the plan, not in the bridge.
- Recurring “one-off” effects. Adjust for severance in three consecutive years and you do not have exceptional items, you have normal staff turnover costs.
- Capitalised internal work without proper records. A perennial issue in software: without time recording, the capitalisation is questioned – and with it the earnings.
- Deferred investment. Earnings that arise because maintenance or marketing was cut back are treated as unsustainable and reduced.
- Revenue from one-off large orders presented as recurring.
The bridge: structure and evidence
The way that survives review is a table running from audited or booked EBITDA to the adjusted figure – line by line, per financial year, with four columns: amount, rationale, evidence, recurrence.
Three rules decide whether it is accepted. First, consistency across all periods. Adjust only in the weakest year and you lose credibility for the whole bridge. Second, both directions. A bridge that only ever points upwards is read with blanket suspicion. Third, evidence. Every line needs a document in the data room – contract, board decision, invoice, account listing.
The same logic produces the quality of earnings analysis in the review report. Present your own bridge cleanly and you negotiate over individual lines. Present none and you negotiate over the buyer's bridge – which starts more conservatively.
The link to working capital and net debt
Adjustments rarely act in isolation. Strip out an expense as one-off and the buyer asks whether the related liability is still open – in which case it moves into net debt. Strike revenue as unsustainable and the reference for normalised working capital changes. Optimise EBITDA alone and you lose elsewhere what you gained.
Frequently asked questions
How large may the adjustments be in total?
There is no fixed limit, but a rule of thumb from practice: once the sum of adjustments reaches a double-digit percentage of starting EBITDA, the whole bridge is examined intensively. Many small, well-evidenced lines come through better than a few large ones without support.
When should normalisation be prepared?
At least twelve months before the intended process. The reason is practical: some adjustments cannot be evidenced retrospectively, only documented as you go – project hours for capitalised development work, for instance.
Who prepares the bridge?
The seller's finance function, often supported by an advisor with transaction experience. The tax adviser alone is rarely enough, because the logic of a transaction review differs from that of preparing accounts.
What is the difference between adjusted and pro-forma EBITDA?
Adjusted refers to results that actually occurred, with exceptional effects removed. Pro forma additionally assumes measures that are not yet effective. Buyers accept the first as a valuation basis, the second at best as a planning argument.
And if the buyer strikes lines?
Then the question is not right or wrong but evidenced or not. Lines with a contract and a decision survive the discussion; lines justified with “it only happened once” do not.
Read on
- Financial due diligence – where the bridge is tested.
- Purchase price mechanics – the multiple applied to the adjusted figure.
- The data room – where the evidence belongs.
- Exit readiness – the frame in which the bridge is built.
- Financial due diligence at nugrow.
Sources and status
The worked figures are model calculations for illustration. The remaining statements are based on nugrow's mandate practice in transaction processes. As of September 2026. This article is an overview and does not replace tax or legal advice.
