Most companies in this sector do not sell a product. They sell a bundle: equipment, embedded software, installation, a licence, and a service contract, often under one contract price and often with the software element carrying most of the value.
That structure is where the diligence work sits. The questions that decide the number are not about whether the technology works. They are about how the price is split, what has been capitalised, and which projects the reported margin actually depends on.
Revenue is a bundle, not a product
Where a contract contains hardware, software, installation and service, the allocation of the price across those elements determines both the amount recognised and the margin reported. Two companies with identical contracts can report materially different revenue and gross margin depending on the allocation, the timing of acceptance, and whether the software is treated as a separate performance obligation.
Read the original contract, not the summary, for three clauses:
- Acceptance. Whether revenue is contingent on commissioning, testing, or a customer sign-off — and how long that has historically taken.
- Software and licence terms. Whether the licence is perpetual, subscription, or bundled into the equipment price; and whether upgrades are included.
- Service and maintenance. Whether it is a separate revenue stream with its own margin, and whether it is contracted or discretionary year to year.
Then test the split: does the allocation follow the contract, or does it follow the reported margin? Where the software element has grown as a share of the price without a corresponding change in what is delivered, the margin is being moved rather than earned.
Project-based businesses also need project-level analysis rather than annual totals. Revenue by project, cost by project, and the stage of completion at period end will show whether the reported margin is being held up by projects that are still open — and whether any project has moved past its budgeted cost without a corresponding adjustment.
Development cost is an asset on the balance sheet
Capitalised development expenditure is common in this sector, and it is where reported earnings are most sensitive to policy. The test is consistency, not correctness in the abstract: is the same capitalisation threshold applied to every project, and are there projects that continue to be capitalised after the commercial case has weakened?
The useful output is a sensitivity rather than a verdict — what operating profit would have looked like under a more conservative policy. That number belongs in the valuation model, not in a footnote. It also has a second effect: capitalised cost reduces future expense only until amortisation starts, so a business that has been capitalising heavily is carrying a rising charge that the reported margin does not yet reflect.
Where the four diligence areas land
- Revenue — allocation of the bundle, acceptance timing, sell-in versus sell-through where distributors are used, and customer concentration in a market with a small number of large buyers.
- Cost of sales — imported components and currency exposure, subcontracting of assembly, warranty provisions, and whether outbound freight and installation sit in cost of sales or in selling expenses.
- Operating expenses — capitalised development, the four-way payroll reconciliation for a large engineering headcount, and shared costs provided by an affiliate.
- Related parties — founder-linked suppliers and customers, premises leased from a related holding company, and balances that represent value moving between the owner and the company.
What to request
- Contracts for the top twenty customers, in original form, with side letters
- Revenue and gross margin by project and by product line for three years
- Project cost records and stage-of-completion workings at each period end
- The capitalisation policy, and a list of development projects with amounts and status
- Supplier list by value, with purchase terms for imported components
- Warranty and after-sales provision analysis against historical claims
The binding constraint in these processes is usually not the analysis. It is whether project-level records exist in a form that ties to the ledger, and whether they can be produced before exclusivity expires. Establishing that in the first week is worth more than any single test.
This page describes a general approach. It is not accounting, tax, or legal advice for a specific transaction. Figures and examples are illustrative. Views are my own and do not represent any firm.