Before a buyer forms a view about your market, your customers or your team, they form a view about your bookkeeping. It happens quickly, it happens from the first spreadsheet you send, and it is close to irreversible.
This article is deliberately narrow. It is not about how to prepare a business for sale in general, or about the wider seller-side diligence exercise that surrounds it. It is about five specific things that live inside the ledger, that a competent bookkeeper can fix in a few months, and that buyers test almost immediately because they determine whether the reported profit means anything at all.
The first thing a buyer's accountant does is test whether revenue landed in the right period. It is the fastest way to find out how the accounts are kept.
The common problems are simple and pervasive. Invoices raised when an order is received rather than when the work is done or the goods are shipped. Customer deposits and advance payments taken to revenue instead of sitting on the balance sheet as deferred income. Long jobs recognised on invoicing dates rather than on progress. Credit notes issued in the new year for work invoiced in the old one, with no accrual made. Annual contracts billed up front and recognised in full on day one.
The test is mechanical and you can run it yourself. Take the last ten invoices before each period end and the first ten after it, and match each one to the delivery note, timesheet, completion record or shipping document. If the two dates disagree systematically, your revenue line has a cut-off problem, and the profit for every period is misstated in a way the buyer will assume is deliberate.
The fix is a documented policy, applied the same way every month: revenue is recognised on a stated event, deposits sit in deferred income until that event, and a deferred income schedule reconciles month to month. Where the policy has been wrong historically, restate the comparatives rather than change quietly in the current year and hope nobody compares.
The second thing a buyer tests is whether the cost lines mean the same thing in every period.
This is where most small-company reporting quietly falls apart. Direct costs coded to overheads in one year and to cost of sales in the next. A new nominal code opened for something already captured elsewhere. Subcontractors sometimes in cost of sales, sometimes in administrative expenses. Delivery costs above the line in one entity and below it in another. Rebates netted against purchases in some months and taken to other income in others.
The consequence is that gross margin, the number buyers care most about in most sectors, becomes uninterpretable. If margin moves three points between years and nobody can say how much of the movement is pricing, how much is mix and how much is coding, the buyer will not give you the benefit of the doubt. They will take the worst plausible reading and price it.
The fix has three parts. Freeze the chart of accounts and write down what belongs in each line, particularly the boundary between cost of sales and overheads. Restate at least two prior years onto the same basis so the trend is comparable. Then stop opening new codes for one-off items; use the existing structure and explain the item in a note.
A clean, stable chart also produces a second benefit: it makes the gross margin analysis by product, customer or service line credible, and that analysis is often what supports the multiple rather than the EBITDA figure itself.
Owners think of the profit and loss account as the thing being sold. Buyers read the balance sheet to find out whether the profit and loss account is true.
Look specifically for what is missing. Holiday pay earned but not taken and not accrued. Bonuses and commissions committed but not booked. Warranty, claims or rework obligations with no provision. A bad debt policy that exists in principle but has not been applied to a ledger full of aged items. Supplier invoices held back at period end so that costs land in the following month. Capital expenditure capitalised when it was really repairs, or repairs expensed when they were really capital, in whichever direction flattered the year.
Then look at policies that have drifted away from reality. Depreciation rates that no longer reflect how long assets actually last. Development costs capitalised on optimistic assumptions. Provisions created in a good year and released in a bad one, which is the single fastest way to lose a buyer's trust in the numbers.
The fix is unglamorous: build a standing accruals and provisions schedule, apply the same policy every month, and document the basis for each estimate. If applying the correct policy reduces reported profit, that is worth knowing now rather than in diligence, and it is dealt with below.
Where a business carries stock or unfinished work, this is usually the largest single area of judgement in the accounts, and buyers know it.
Count it properly and on a schedule, not once a year in a rush. Reconcile the count to the ledger and investigate differences rather than posting a balancing entry. If book and physical stock differ materially, that gap is not a one-off adjustment; it is evidence of an ongoing loss the buyer will build into the earnings.
Value it on a documented basis. If standard costing is used, show when standards were last reviewed and how variances are treated. If overheads are absorbed into stock, show the absorption rate and the assumptions behind it, because an over-absorbed stock balance is a way of capitalising costs that should have hit the profit and loss account.
Provide for obsolescence with a written policy — an ageing profile with defined provision rates is far more persuasive than a judgemental figure that happens to move with the result. And check the housekeeping: negative stock lines, consignment goods included in your own balance sheet, and items that have been carried at cost since three managing directors ago.
Work in progress deserves the same discipline. On uncompleted jobs, WIP should reconcile to actual hours and materials, not to what was expected to be spent, and the recoverability of anything unbilled for more than a few months should be assessed openly.
The last fix is the one owners most often postpone, because it is personal.
Take an inventory of everything in the ledger that involves the owner or a connected party. The director's loan account and its movement. Personal expenses paid by the company. Family members on payroll and what they actually do. Rent paid to a property the owner owns, above or below market. Vehicles, phones, travel, subscriptions. Trading with another company the owner controls, on terms that were never negotiated.
Each of these is manageable if it is identified, evidenced and priced correctly. Each becomes a problem when a buyer's accountant finds it and asks how many more there are.
The fix has a sequence. Stop the items that should stop, ideally a full year before a sale so that a clean year exists. Formalise the ones that will continue: a written lease at market rent, a service agreement with the related company, a documented salary for a genuine role. Clear the director's account or agree how it will be settled at completion. And prepare a related-party schedule listing every item with its annual value and treatment, so the buyer receives a list rather than assembling one.
This matters beyond tidiness. Adjustments for owner costs are usually the largest add-backs in a small company's adjusted EBITDA, and add-backs are only accepted when they are identifiable, evidenced and demonstrably non-continuing. A well-documented related-party schedule is what converts an assertion into an accepted adjustment.
Sitting alongside these five, and touching three of them, is a document most small companies maintain badly: the fixed asset register.
Two problems recur. The first is ghost assets — items long since scrapped, sold or replaced that are still being depreciated, which overstates both the balance sheet and the depreciation charge. The second is the reverse: assets in daily use that were expensed years ago and appear nowhere, so the business looks less capitalised than it is and the buyer cannot see what they are acquiring.
The consequences reach further than the balance sheet. Depreciation is added back to reach EBITDA, so a wrong charge distorts the headline figure. Maintenance capital expenditure, which buyers deduct from EBITDA to understand real cash generation, cannot be estimated without knowing what exists and how old it is. And in an asset sale, the schedule of what is being transferred is built from this register, so errors become contractual rather than merely accounting problems.
The fix is a physical verification against the register, disposal of what no longer exists, capitalisation of what should have been capitalised, and a depreciation policy that reflects actual useful lives. In our experience at Conclave Partners this exercise takes a few days and routinely changes both the EBITDA figure and the capital expenditure discussion that follows it.
Underneath these five fixes sits one habit: reconcile everything, every month, and keep the evidence.
Bank accounts to the ledger, with no unexplained reconciling items carried forward. Indirect tax returns to the revenue and purchase ledgers. Payroll reports to the nominal ledger. Intercompany balances agreed both ways. Debtor and creditor ledgers to the control accounts. Fixed asset register to the balance sheet.
Then make sure there is one version of the numbers. Management accounts should reconcile to statutory accounts through a short, documented bridge, and everyone in the business should quote the same figures. Buyers accept differences between management and statutory reporting; what they do not accept is a difference nobody can explain.
In our experience at Conclave Partners, this reconciliation file is the single most reassuring document a seller can put in front of a buyer's accountant, because it answers the underlying question — can these numbers be relied on — before it is asked.
There is an honest trade-off here that most articles avoid, so it is worth stating plainly: doing this work properly sometimes lowers reported profit.
Accruing holiday pay that was never accrued, providing against stock that will never sell, recognising deferred income correctly and expensing costs that were wrongly capitalised all reduce the earnings figure in the year they are corrected. The instinct is to leave it alone.
That instinct is usually wrong, for three reasons. A buyer's accountant will make the same adjustments anyway, and will make them less generously than you would. Adjustments discovered in diligence carry a credibility cost as well as a value cost, because they raise the question of what else is unadjusted. And the correction is normally a level change rather than a trend change: once made consistently across the comparative years, the growth story survives, which is what the multiple is applied to.
The practical approach is to correct the policy, restate the comparatives on the same basis, and present the bridge from old to new clearly. A seller who says "we tightened this policy, here is the effect on all three years, here is why" is in a far stronger position than one who is caught mid-process.
There is a tax dimension too, since correcting accruals or provisions can change taxable profit, and the timing of a correction relative to the year end has consequences. That is a conversation to have with your accountant before the change, not after.
Start with revenue cut-off, because everything else is read through it and because it takes the longest to show a clean run of months.
Then fix the chart of accounts and restate the comparatives, so that the trend the buyer sees is the same trend on the same basis.
Then accruals, provisions and policies, then stock and work in progress, since both need a full cycle of consistent months to be persuasive.
Deal with the owner and related-party items in parallel, because the calendar matters: a clean year is only clean if it starts a year before.
Twelve months of consistent, reconciled monthly accounts on a fixed basis is the realistic target. Sellers who ask Conclave Partners how long the accounting work takes are given that number, because it is the shortest period that produces evidence rather than assertions.
None of these five fixes makes the business better. They make it legible, which in a transaction is close to the same thing.
The mechanism is straightforward. Adjusted EBITDA is what the multiple is applied to, and every one of these fixes affects either the adjusted figure itself or the buyer's willingness to accept it. Clean cut-off makes the revenue line believable. A stable chart makes the margin analysis usable. Complete accruals stop the buyer discovering costs later. Documented stock policy protects the largest judgement in the balance sheet. A related-party schedule converts owner add-backs from claims into accepted adjustments.
At Conclave Partners we would rather a seller spend a modest amount of time on these five items than on any amount of narrative about market opportunity, because the first thing that has to survive contact with a buyer is not the story. It is the ledger.
Usually three years of statutory accounts plus at least two years of monthly management accounts, and they will want the monthly detail because that is where seasonality, cut-off problems and one-off items become visible. Anything you correct should therefore be restated across the comparative years, not just fixed going forward.
Not in most lower middle market transactions. Buyers rely on their own quality of earnings work rather than on an audit opinion. What matters far more is that management and statutory accounts reconcile, that policies are consistent, and that the underlying records support the figures.
It can reduce reported profit in the short term, and it usually improves the price achieved anyway, because the corrected figure is the one that survives diligence. Corrections made by the seller in advance are level adjustments applied consistently; corrections forced during diligence come with a discount for uncertainty on top.
Revenue cut-off, followed closely by inconsistent classification of costs between periods. Both distort the trend rather than a single number, and the trend is what the buyer is pricing.
Stop them, ideally a year before going to market, and document historic ones item by item with dates and amounts so they can be evidenced as add-backs. Adjustments a buyer can verify are accepted; adjustments supported only by the owner's word cause the buyer to discount the whole earnings figure.
Around twelve months to produce a clean run of consistent, reconciled monthly accounts, less if the underlying records are already in good order. The restatement of prior years can be done more quickly, but the credibility comes from a period of consistency, not from a single tidy-up exercise.
Ildar Zakirov — Conclave Partners ildar@conclavepartners.com
Sergi Kosiakof — Conclave Partners sergi@conclavepartners.com