In a portfolio review meeting, the continuation of an existing product is seldom the matter under decision; what requires a decision is its termination. Last year's catalogue opens as this year's working baseline, discussion proceeds through the items proposed for addition, and the rationale for the items already there is not revisited. The observable consequence of this asymmetry is straightforward: in most companies a product leaves the catalogue not because a defined threshold has been tripped, but because a supply source has lapsed, a certification renewal has become expensive, or the person who carried that item alone has left the organization. Adding a product has a process — business case, budget, approval threshold. Removing one usually has only a cause.
The question posed at the diligence table targets this asymmetry directly. The investor side generally asks not which products have been developed, but which products have been discontinued over the past three years and against what rule that decision was taken, since the answer to the first is already contained in the roadmap deck while the answer to the second exposes the company's decision architecture in a single sentence. A response resting on the absence of any need, or on the portfolio already being narrow, tends to indicate not a narrow portfolio but a discontinuation threshold that was never defined at all. The quality of that answer determines whether product management is a discipline or an accumulation.
The mechanism underneath this behaviour operates in layers. The first is the asymmetry of decision cost: continuation requires no signature and attaches to no name, whereas termination produces a signature, an accountable owner, and a rationale that must be defended retrospectively. The second is the sunk cost fallacy — the tendency for expenditure already incurred to govern the decision still to be made — under which development time spent, tooling invested, or certification fees paid become arguments for continuation despite being wholly independent of future marginal contribution. The third is status quo bias, whereby maintaining the present configuration demands materially less justification than an economically equivalent alternative. None of these three layers is an error; each is a shortcut that lowers decision cost. The difficulty lies not in the shortcut itself but in its persistence as the portfolio widens.
An organizational layer sits on top of this. Product ownership roles are defined almost everywhere in the vocabulary of growth — new feature, new segment, new channel — while removal appears in no role's performance definition. The commercial function argues that a low-volume item carries a particular customer relationship, and taken individually that claim is frequently accurate; manufacturing notes that the tooling has long since been amortized, which is also accurate; engineering observes that the module's maintenance burden is marginal. Because each defence is reasonable on its own terms, the aggregate complexity cost of the portfolio never becomes an agenda item for any single meeting, and the tail grows as an accumulation of deferred decisions rather than as the outcome of a decision.
On the financial statements that accumulation typically resides not in cost of goods sold but in the working capital cycle and in the undistributed portion of overhead. On the physical product side, the tail slows inventory turnover, enlarges the slow-moving stock provision, extends setup and changeover time in production planning, and multiplies part numbers and supplier relationships in a manner that dilutes purchasing leverage. On the software side, the same tail appears as security patch maintenance, recurring compliance certification, support ticket resolution time, and the carrying cost of technical debt. What these items share is that they are allocated in proportion to revenue, with the result that a low-volume item is reported as bearing a cost burden materially below the complexity it generates.
The valuation channel opens precisely here. When the buy-side quality of earnings analysis reconstructs margin distribution by product on activity drivers rather than revenue-weighted allocation, the negative contribution of the bottom quartile to normalized EBITDA typically becomes visible; that finding converts into a direct price adjustment, into a downward revision of inventory valuation within the working capital peg negotiation, or into a discrete escrow item for obsolete stock and warranty exposure. The more expensive outcome is a different one. The buyer prices the discontinuation decisions it will have to take after closing using its own cost estimate, and holds that estimate conservatively in a predictable direction. Rationalization the seller never performed is ultimately priced as the buyer's discount.
The first two dimensions of the review are therefore formal, and discriminating nonetheless. Under existence, what is sought is whether the discontinuation criterion exists as a rule set with articulated thresholds rather than as unwritten commercial judgement: below which contribution margin level, within which volume band, and sustained over which period an item falls automatically into review. Under documentation, the requirement is that these thresholds sit in an approved and dated instrument whose most recent update falls within a plausible interval. A policy approved two years ago and never applied since produces a weaker signal than its outright absence, since it demonstrates simultaneously the absence of the rule and the absence of the discipline to observe it.
Implementation and measurement demand an evidentiary chain rather than a text. What is examined under implementation is not the wording of the policy but the decision record it generates: the list of items that breached the threshold, the decisions taken on those items, the rationale for exceptions granted, and the proportion those exceptions represent within the portfolio. A rule admitting no exceptions is not realistic; exceptions left unrecorded indicate that the rule is not in fact operating. Under measurement, what is sought is that the discontinuation decision itself is tied to a performance metric — how much of the discontinued item's revenue was retained within the portfolio, where the released manufacturing or development capacity was redirected, what movement was observed in complexity indicators. Where that measurement is absent, discontinuation reads solely as a loss, and taking the decision a second time becomes markedly harder.
The final two dimensions carry the valuation discount most directly. Under ownership, the question is at which role and at which threshold the discontinuation decision is taken; a tiered structure in which the product owner holds only a proposal right, the product manager decides below a defined revenue or customer-count threshold, and a committee decides above it, demonstrates that the decision has migrated from a person to a role. Under continuity, what is sought is that the process repeats independently of founder intuition: that the criterion runs on a calendar-bound cadence — typically synchronized with the budget cycle — without prompting, and that the last two cycles were completed without founder participation. That no portfolio review has ever taken place with the founder absent from the room constitutes a diligence finding in its own right.
The architecture that neutralizes this tendency has four components. The first is that the decision record is kept at the moment of proposal rather than the moment of approval: at every product launch, the conditions under which that item will be discontinued — volume threshold, contribution margin floor, first review date — are written into the launch document, so that discontinuation becomes a condition accepted at the outset rather than a judgement defended afterwards. The second is that review is bound to the calendar rather than to the agenda; the list of items breaching thresholds is generated automatically and enters the agenda without anyone requesting it. The third is that continuation also requires a signature, since the asymmetry in decision cost closes only when continuing likewise demands a rationale. The fourth is that the exception log is maintained separately, because the accumulation rate of exceptions is the most reliable indicator of whether the rule genuinely holds.
BEIREK's intervention in this area begins not with a portfolio screen but with the construction of the decision architecture. We map the portfolio's actual contribution through a cost allocation grounded in activity drivers, then translate that map into a criteria set composed of thresholds, tiered decision rights and automatic triggers, and embed it within the company's own governance instruments. The record we establish holds three things: items breaching thresholds, decisions taken, and the rationale for exceptions. The cadence we operate is tied to the budget cycle; the first two cycles are run jointly, the third observed only, because what determines value at the diligence table is not that the decision was once taken correctly, but that the same decision can be produced with the founder outside the room.
The quality of a portfolio is measured not by the sum of the products it contains but by which item the company is capable of releasing at which threshold, since the capacity to add can be purchased with capital while the discipline to remove can only be built institutionally. A company that has discontinued no product on a planned basis over three years demonstrates not the extraordinary accuracy of its portfolio but that accuracy has never been tested, and an untested claim to accuracy is priced at the diligence table not as an assumption but as an open position. The distinction usually resolves into a single question: among the items presently in this portfolio, which would be approved if it were proposed again today, on today's information?