Add-back Tightening

When an owner-run business is marketed to a fund or a strategic buyer, its earnings are usually presented as adjusted EBITDA: the reported figure plus a list of items the seller argues a buyer should ignore. Some are legitimate and some are not. The list tends to grow as the sale approaches, because every item on it is multiplied by the price, and the people preparing it have every reason to find more. A buyer's quality-of-earnings team then spends weeks taking the list apart, and the multiple is applied to whatever survives, usually after a price has been agreed on the larger figure.

Add-back Tightening shrinks the gap between reported and adjusted earnings until what remains is a short list that a quality-of-earnings team accepts without a haircut. Every adjustment left on it is a past event with an invoice behind it, none of it asks the buyer to believe in the future, and reported and adjusted EBITDA end up close enough that a buyer barely notices the difference.

A small gap

Adjustments other than the owner's compensation should total no more than a modest share of EBITDA in each of the last three years and in the trailing twelve months.

That limit is tighter than the one set earlier on this bridge because the buyer has changed. A fund pricing a business on adjusted EBITDA wants the earnings it pays for to be the earnings the statements show. A large gap between the two leaves the business's real performance somewhere a reviewer has to argue about, and arguments about earnings end in lower prices.

Most of the gap is closed by a recurrence test. Every adjustment is checked across three years, and an item that keeps coming back, under whatever name, is a cost of doing business. It comes out of the schedule and stays in the expenses. The test catches the most familiar weakness in an adjustment schedule, the one-off cost that happens every year: legal fees for a dispute, emergency hiring, rework on a botched job, a write-off of bad stock. Each is presented as unusual. Seen across three years, some turn out to be the ordinary cost of running the business the way it is run.

Where the same kind of event keeps happening, the better answer is to fix its cause. A business that keeps paying for rework has a quality problem, and a business that keeps hiring in emergencies has a planning problem. Solving either removes a recurring cost from the expenses, which raises the earnings honestly, with nothing for a reviewer to strike.

The owner's adjustment, settled first

The limit leaves out one item: the adjustment for the owner's own pay, for the one seat the owner keeps. That adjustment is legitimate and often large. If the owner was paid more or less than the market rate for the seat, the difference is restated to market, because a buyer will pay whoever holds the seat the market rate.

It is also the line where other items hide most easily, which is why it is settled before anything else. A family car, a relative's salary, the owner's personal travel and a spouse's phone can all be folded into a generous figure labelled "owner compensation" and added back together, so that items which would never survive on their own lines pass unexamined inside one large and plausible one. Putting the owner on the payroll at the market rate for the retained seat, as Management Install does, fixes the owner's adjustment at a known figure from a named source. Anything else the owner took from the business then has to appear on its own line, where it is tested like every other item.

A buyer's reviewer reads the owner's adjustment first for exactly this reason. When it matches a documented market salary, the reviewer can see there is nothing hidden in it, and the rest of the review goes faster.

Nothing that looks forward

Every item on the schedule should be a historical cost already incurred. That rules out adjustments for savings not yet realized, prices not yet raised, staff not yet hired, and a buyer's own synergies.

Forward-looking adjustments are the easiest to argue for and the first to be struck. A seller who has just signed a cheaper supplier contract will want to show next year's lower costs in this year's earnings. A seller planning a price rise will want credit for it, and a seller who expects a buyer to merge the business with another will want a share of the savings. Earnings are a record of what happened, and none of these has happened yet. The saving the owner believes in belongs in the budget and is proven by next year's statements, and the price rise belongs in the forecast. The synergies belong to the buyer, who will pay for them, if at all, through the price it chooses to offer.

The outside accountant signs off the schedule each year as part of the review engagement, and any item the accountant would not accept is removed before the schedule is used anywhere. The sign-off is the cheapest version of the review a buyer will later commission, and it happens before anyone has a price in mind.

Every remaining item proved

Each adjustment that survives carries the invoice, contract or correspondence that proves its amount and date, and a one-line explanation of why it will not recur. Each gets a binder holding the source document, the ledger entries, the explanation and the accountant's note, filed in the order of the schedule, so a reviewer who asks about any item is handed its binder.

The illustration shows what is at stake. A business with $800,000 of reported EBITDA that presents $980,000 of adjusted EBITDA, including $60,000 of costs that recur under other names and $50,000 of forward-looking items, will find a review confirming $870,000. At an illustrative six times adjusted EBITDA, the $110,000 that did not survive is $660,000 of price, and it is lost at the stage of the sale where the seller can least afford to walk away.

What it does to value

In the framework's valuation, Add-back Tightening is a correction. Moving recurring costs back into the expenses and removing forward-looking items lowers the earnings power the valuation sees, and the honest figure is usually smaller than the one the owner had been quoting. The framework treats that smaller figure as the one a buyer's model would have reached anyway, so it records no loss. The gain is that the figure survives diligence: adjustments that would have been struck at the multiple are no longer there to strike. The framework records that as an effect on price, which is why the driver ranks low on value. It still has to be done, because a figure that fails a review casts doubt on everything else the business has shown.

What has to be in place first

The driver tightens a schedule that already exists: the ledger-tied schedule of adjustments built under Separation. It needs the owner's retained seat on the payroll at market, from Management Install, so that the owner's adjustment is settled first and cannot hide other items. And it needs three years of reviewed or compiled statements on the same basis, from Books, because recurrence can only be tested across years that were closed the same way.

Knowing that it works

Three numbers show the state of the schedule: adjustments other than the owner's pay as a share of EBITDA, for each year and the trailing twelve months; adjustments without a source document, which should be none; and forward-looking items still on the schedule, which should also be none.

A useful test is to show the schedule to an accountant or broker who has never seen the business, and to document on the spot, or remove, every item the stranger questions. The other is a habit: the owner quotes adjusted and reported EBITDA together, never one without the other, and can explain the difference in a sentence. A gap the owner can explain that briefly is usually one a buyer will accept.

Why it slips, and what holds it

Left alone, adjustments multiply. Each year brings a new item that seems one-off, and the schedule grows until a buyer stops believing it. Closing the schedule with each fiscal year and having the outside accountant sign it means a new item has to get past the accountant at the year end before it can appear in a sale document. Running the recurrence test every year across a rolling three years reclassifies a recurring item as an expense automatically, before the owner has grown attached to it.