--- title: "Return on Investment: An Approval Instrument or a Management Record?" description: "In a valuation review, return on investment is assessed not through forward IRR forecasts but through a record showing the realized outcomes of completed investments against the assumptions approved at the time. Absent that record, a buyer tends to model growth capital expenditure as if it were maintenance spending, deducting the outlay from free cash flow while writing in no corresponding revenue lift." url: https://www.beirek.com/en/blog/return-on-investment-valuation-review canonical: https://www.beirek.com/en/blog/return-on-investment-valuation-review published: 2026-05-30 modified: 2026-05-30 category: "Financial Performance" category_url: https://www.beirek.com/en/blog/category/financial-performance language: en-US reading_time_minutes: 8 publisher: BEIREK LLC publisher_url: https://www.beirek.com license: "© BEIREK LLC — citation with attribution and link permitted" keywords: ["return on investment","capital allocation framework","growth capex versus maintenance capex","post-investment review","due diligence valuation discount"] topics: ["Investment readiness and valuation review","Capital allocation governance","Founder dependency and transferable capability"] alternate_language_url: https://www.beirek.com/tr/blog/return-on-investment-valuation-review --- # Return on Investment: An Approval Instrument or a Management Record? > **In short:** In a valuation review, return on investment is assessed not through forward IRR forecasts but through a record showing the realized outcomes of completed investments against the assumptions approved at the time. Absent that record, a buyer tends to model growth capital expenditure as if it were maintenance spending, deducting the outlay from free cash flow while writing in no corresponding revenue lift. *Most companies can produce a forward-looking return calculation for every new capital item, yet almost none can show the realized return of a single completed investment from a record. What the review table looks for is not the elegance of the forecast but the existence of an institutional memory that sets forecast against outcome, and the absence of that memory travels directly into valuation.* --- When a new capital expenditure item reaches an investment committee or a board session, the file almost invariably contains a payback period, an IRR estimate, and frequently a sensitivity table built around price and volume assumptions; what rarely surfaces in the same session is the realized return of an item approved two or three years earlier in the identical format, and where the question is raised, the answer arrives from the memory of the most senior person in the room rather than from a record. The magnitude of this asymmetry varies from company to company, but its direction is remarkably stable — forward-looking return calculations are plentiful, backward-looking return records are absent. Both calculations rely on the same formula, require the same data infrastructure, and could be produced by the same finance team; only one of them, however, has hardened into an institutional habit. At the review table the asymmetry surfaces through a differently framed question. Asked in the course of due diligence to set the approval-stage assumptions of the three largest capital items of the past three years alongside their present actual outcomes, a company typically responds with a narrative rather than a schedule: the line came on later than expected, a supplier delay pushed a season, capacity is nonetheless fully utilized today. The narrative may be accurate, and frequently is; it is not, however, verifiable, and a return claim that cannot be verified is not treated as existing for review purposes. The question the company has never put to itself is the first question the counterparty puts to it. The underlying mechanism is less an omission than a drift in function. In most organizations the return calculation is produced not to generate a decision but to secure a permission; its true addressee is not future performance but the threshold applied by the approving body. Framed that way, the calculation has discharged its purpose the moment approval is granted, and is left unmaintained as a matter of course — no one allocates resources to updating the documentation of a closed transaction. The same logic feeds the definitional looseness inside the calculation: which cost base enters the denominator, whether pre-commissioning indirect expenditure is counted, over what horizon and against which cash flow definition the return is measured, are questions left unwritten in most companies, because clearing an approval threshold does not require those definitions to be stable, only that they be persuasive once. It deserves acknowledgment that in a particular configuration this shortcut genuinely lowers cost. Where the person allocating capital is also the person operating the facility, speaking with customers, and negotiating with suppliers, the feedback loop is short enough not to require a record; whether the machine is performing as expected is visible on the production line, without reference to any schedule. The difficulty lies not in the shortcut itself but in its persistence after the conditions change. Once the number of sites multiplies, once several management layers interpose themselves between the investment decision and the operating outcome, and once part of the capital comes from outside, the intuitive feedback loop breaks; the break usually goes unnoticed, because no structure was ever built to succeed it, and its absence therefore registers as nothing at all rather than as a gap. Whether the structure exists at all is legible from a handful of concrete surfaces. Does the company maintain a written capital allocation framework; is there a defined hurdle rate that items must clear, and has that rate been tied to the cost of capital; above what size does an item require board approval; and where items characterized as strategic are exempted from the threshold, on what criterion does the exemption rest. On the documentation side, what is sought is not a deck but an approved policy text together with a board record carrying the reasoning behind the decision. On the implementation side, whether the framework is actually operative becomes visible through the exemption rate: the existence of a threshold and the bindingness of a threshold are different things, and where a meaningful share of items clears under an exception, the framework is decorative rather than determinative in practice. The first channel through which the deficiency reaches valuation is the classification of capital expenditure inside the model. A buyer-side financial model splits forward capex in two: maintenance spending that holds existing capacity in place, and growth spending that generates incremental cash flow. The former is deducted from free cash flow with no revenue increase expected in return; the latter enters the projection on the revenue side only to the extent that the return it will generate can be demonstrated. Where the realized return of past investments cannot be evidenced from a record, the predictable choice of a conservative modeler is to carry planned growth capex without its growth effect — that is, to write in the outlay and leave out the return. That single classification decision pulls down the base multiplied by the multiple more forcefully than the headline multiple negotiated over many hours. The second channel runs through ownership. Where no record of the capital allocation decision exists, its function is discharged by the memory of the founder or a single senior executive; why a given investment was made, why a given assumption shifted, and which lesson was carried into the following item reside with that person alone. This does not deny competence — a settled and well-calibrated allocation instinct is among the most valuable assets a mid-sized company holds — but from a review standpoint it is an asset that cannot be transferred, and capacity that cannot be transferred does not enter the transaction price. The consequence typically takes one of two forms: an earn-out structure tied to the founder's post-closing tenure, or the removal of the entire growth scenario from the base case into an optional upside case. The third channel is measurement, and it is generally the quietest. A company that does not track realized returns cannot separate the investment that worked from the investment that was masked by a favorable market; so long as aggregate profitability looks sound, both varieties count as successful, and the following period's allocation decision is built on that undifferentiated impression. In the first period in which prices soften or demand narrows, which portion of an intuitively constructed allocation logic reflected capability and which portion reflected the cycle becomes abruptly visible. The reviewing party cannot wait for that separation to occur, and precisely because it cannot wait, it prices every return whose attributability has not been demonstrated as though it were cyclical. What neutralizes this tendency is not individual discipline but a record architecture composed of four components. The first is a written capital allocation framework, in which the hurdle rate, the definition of return, the scope of the cost base, and the measurement horizon are fixed in a single text, so that comparison across items becomes meaningful. The second is opening the investment record at the moment of proposal rather than the moment of approval; the one element that cannot be calibrated retrospectively is the initial assumption, and an assumption recorded after approval inevitably conforms itself to the outcome. The third is a fixed-calendar realization review for every completed item — the first and second year after commissioning, for instance — conducted for calibration rather than for censure. The fourth is separating the role that prepares the proposal from the role that reviews the realization; where the same person both frames the assumption and judges the result, the record, even when kept, produces no independent information. BEIREK's intervention in this area typically begins not with drafting a policy text but with reconstructing the existing investment stock retrospectively: completed items from the past three years are set alongside their present actual outcomes together with their approval-stage assumptions — recovered from board records where available, and rebuilt from the budget file of the period where not — and the resulting variance is read not item by item but by the recurring direction of the variance. A volume assumption that proves systematically optimistic is more informative than the investment decision itself, since it supplies the correction coefficient for future projections. Once that backward-looking record is complete, the forward-looking line is established: a one-page assumption record opened at the proposal stage for every new item, a realization review rhythm anchored to the board calendar, and an allocation of authority in which the proposing and reviewing roles are separated organizationally rather than contractually. The principal gain from building this structure is not confined to answering the question during a review process. Once capital allocation decisions are recorded, the company learns the direction and magnitude of its own forecasting error; that learning improves not the hurdle rate applied to the next item but the quality of the assumptions brought to the hurdle, and after several cycles the improvement becomes visible in the realized return itself. At the review table, moreover, the record is not required to demonstrate that forecasts were accurate; the existence of a mechanism that observes the variance, records its cause, and carries it into the next decision constitutes a stronger signal of institutional maturity than accurate forecasts standing alone. What is ultimately sought under the heading of return on investment in a valuation review is not that the company has made profitable investments in the past — profitability on its own can equally be a product of the cycle. What is sought is whether the decision about where capital goes has detached itself from the founder's judgment and become a repeatable capacity of the company, and the sole observable evidence of that detachment is a record placing forecast and outcome face to face. A company that does not keep such a record today cannot manufacture one retrospectively in the weeks before closing either; the value of the record derives precisely from its having been written before the outcome was known. ## Key Points - In most companies the return calculation is produced as an approval instrument rather than a management tool; having cleared the threshold, it has served its purpose and is left unmaintained. - The absence of a realized-return record pushes an acquirer to treat growth capex as maintenance capex, which lowers projected free cash flow directly and quietly. - When the memory of capital allocation decisions sits in the founder's judgment rather than in a record, that capability is treated as non-transferable and widens the founder-dependency discount. - A company without a measurement layer cannot distinguish an investment that worked from an investment that was flattered by favorable market conditions. - Opening the investment record at the moment of proposal rather than the moment of approval fixes the one element that cannot be reconstructed later: the initial assumption set. ## Questions ### What exactly is examined when return on investment is assessed in due diligence? Not forward IRR or payback estimates, but a record permitting comparison between the approval-stage assumptions of completed investments and their actual outcomes. The reviewing party looks for a written hurdle rate, a defined method of calculating return, a board record carrying the reasoning behind the decision, and a realization review conducted after commissioning. Where these elements are missing, the return claim is not treated as verifiable. ### How does valuation change when realized returns on past investments are unrecorded? The effect appears most sharply in how capital expenditure is classified inside the model. Growth investments whose return cannot be demonstrated are handled by a conservative modeler as maintenance spending: the outlay is deducted from free cash flow while no corresponding revenue lift is written in. Beyond that, the growth scenario may be removed from the base case into an optional upside case, or an earn-out tied to the founder's continued tenure may be proposed. ### At what point should an internal investment return record be opened? At the moment of proposal, not the moment of approval. The one element that cannot be reconstructed afterward is the initial assumption set; written after approval, it inevitably conforms to the outcome already known, and the record then yields no independent information. Volume, price, commissioning date, and cost base assumptions fixed at the proposal stage constitute the only valid reference point for any later variance analysis. ### Does a small or mid-sized company genuinely need a separate capital allocation framework? Where the person allocating capital also runs the operation personally, the feedback loop is short enough that the marginal benefit of a written framework may remain limited. That loop breaks once the number of sites grows, once management layers interpose between decision and outcome, or once outside capital enters. The break usually escapes notice, because no succeeding structure was ever built and its absence therefore never registers as a gap. --- Source: https://www.beirek.com/en/blog/return-on-investment-valuation-review Publisher: BEIREK LLC — https://www.beirek.com