In a year-end review session, the innovation unit typically opens with volume: how many ideas were collected, how many accelerator cohorts were completed, how many prototypes were built, how many pilots were signed with startups, how many employees attended the workshops. When someone in the same session asks which of these activities moved a line in the company income statement, the answer usually returns to another activity indicator — that the pilot advanced to a second phase, that the team was expanded, that scaling is planned for the next budget cycle. A question posed at the level of outcome and answered at the level of activity disturbs nobody in the room, the presentation having been constructed to perform exactly that translation. Within the same company, a production line or a sales region defended in the same manner would not survive the first follow-up question.
This asymmetry originates in the separate measurement regime that institutions grant to innovation activity. Given genuine uncertainty, the decision not to expect early results is a reasonable opening assumption; where the assumption carries no written expiry date, however, it converts into an open-ended exemption. Activity itself begins to generate legitimacy: the existence of an accelerator is presented to investors, to the board, and to the labor market as evidence that the company is prepared for what is coming, and the return on that presentation is real — talent attraction, institutional standing, positioning before regulators. Activity that produces no outcome persists precisely to the extent that it is rewarded despite producing none.
The name of this pattern is innovation theater — visible innovation activity generating institutional legitimacy independently of commercial result. Its mechanism is not incompetence but incentive alignment: the team running the initiative also produces the inputs to the assessment of whether the initiative should continue, and the collapse of those two roles into one location makes the outcome predictable. Closing a pilot dissolves the team's budget, its headcount, and its internal standing, whereas advancing the pilot to a further phase preserves all three; the reporting that emerges from such a structure will therefore be organized, quite naturally, to foreground continuation signals rather than closure signals. Nobody in this arrangement is behaving improperly, the conditions having been set that way.
A second layer arrives from where innovation spending is booked. These costs typically accumulate in a central overhead pool, never landing on the margin of any single business unit, which means that no budget owner experiences the expenditure as an obstacle standing between themselves and their own target. That an activity nobody bears the cost of is questioned by nobody is a structural consequence rather than a cultural failing; when accountability disperses, the reflex to interrogate disperses with it. Charged instead against the sales budget of a single region, the same amount would have produced an outcome question from that region's manager inside the first quarter.
Ignoring the phases in which this tendency is genuinely functional would render the diagnosis ideological. An institution making first contact with a new technology domain must build vocabulary, a supplier map, and internal competence before any commercial result becomes available to measure; during that window, measuring activity is the correct approach, since there is nothing else yet to measure. The shortcut remains rational as long as it lowers the cost of learning. The difficulty lies not in the shortcut but in the retention of the same measurement regime after the learning curve flattens — where the second year's indicators are identical to the first year's, the program is not learning, it is repeating.
Institutional cost accumulates first in management attention rather than in cash flow. The number of pilots running concurrently in a portfolio divides the review time senior management can allocate to each; beyond a certain threshold no initiative receives scrutiny deep enough to justify either a continuation or a closure decision, and the default decision becomes continuation. Because it never appears as a discrete line item, this is the cost recognized last. Add to it the senior technical personnel occupied by pilots that never close, and a causal link forms — one that is almost never drawn in the room — between delays in the core business line and the innovation program itself.
A second cost accumulates on the commercial side, in the form of a supplier relationship. Memoranda signed during the pilot phase are negotiated under an implied signal of scaling, which is why they are typically established with concessionary pricing and generous scope; where the pilot neither closes nor scales, those terms are carried forward through successive renewals for years, while the counterparty continues to exercise its reference rights in the absence of any real purchase commitment. On the institution's side, neither supplier leverage nor commercial outcome materializes; what materializes is a line in the contract inventory requiring administration.
A third cost becomes visible at the valuation table. In an acquisition or a capital raise, an innovation budget that cannot be matched to a defined revenue line, a measured cost saving, or transferable intellectual property is reclassified — not as research and development investment, but as a discretionary expense the buyer can trim comfortably after closing. That reclassification may appear to favor the seller, since normalized profitability rises, yet it carries two consequences: the buyer has established that the program is a cost item rather than a strategic asset, and that finding opens the question of whether the product roadmap is supported by in-house capability at all. What proves decisive in valuation is never the magnitude of the spend, but whether the spend can be tied to a repeatable outcome mechanism.
The neutralizing mechanism is decision architecture rather than individual awareness, and it has three separable components. The first is a written termination threshold defined at the inception of every initiative: the question of which observation triggers closure must be answered, and recorded, on the day the program begins. The second is the structural separation of termination authority from the team executing the initiative; the continuation decision belongs to a body whose own budget is unaffected by continuation. The third is the attachment of every initiative to one P&L line and one named budget owner, the central overhead pool being the quietest available method of dissolving accountability.
BEIREK's intervention in portfolios of this kind begins not by shortening the list of initiatives but by converting the list into a decision record. For each active initiative we open a single-page decision file containing the underlying assumption, the one criterion that must be validated, the date on which that criterion will be observed, the revenue or cost line to which the initiative attaches, and the closure threshold. What matters is that the file is maintained at the moment of proposal rather than at the moment of approval; a threshold written after the result is visible will, inevitably, be calibrated to the result. At portfolio level we run a quarterly review rhythm whose agenda is not which initiatives will continue but which will be closed — constructing closure rather than continuation as the default decision changes reporting behavior throughout the organization.
The second layer of that architecture keeps activity indicators and outcome indicators in the same table but in separate columns. Measuring activity is not prohibited, since in the early phase it is the only meaningful measurement available; what is fixed in advance is the month in which the initiative must migrate from the activity column to the outcome column, after which the activity indicator is treated as void. The same separation forms the spine of the narrative carried to the board: not how many pilots are running, but how many were closed, how many were transferred into the core business, and what the first-year margin contribution of the transferred ones amounted to. Once the closure count becomes a reportable measure of success, the incentive that produces the theater is inverted.
What indicates an institution's innovation capacity is not the number of initiatives it launches but the number it can close on schedule and without argument; in a portfolio lacking closure discipline, good ideas compete for resources against the persistence of poor ones. The question worth asking at the investment committee table is therefore not what the program has produced, but whether the conditions under which the program terminates itself have ever been written down.