When vision is owed

Belief priced forward. Numbers earned slow. Vision owed selectively.

Most strategy debate treats narrative and operations as substitutes or sequenced phases. They are neither. They are two surfaces a public company runs on. Perception, owned by the CEO. Economics, owned by the CFO. They run in parallel, on different clocks, with different mechanics. Strategy is the discipline of keeping them coupled. Stewardship is what remains when one of the surfaces is not movable.

The perception surface

Narrative moves are discrete, legible actions that change how the market understands a company's future. Acquisitions, equity placements, strategic stakes, partnerships, rights deals. Their function is not economic in the near term. It is to shift the discount rate by changing expectations of what the business could become. Done well, they pull valuation forward, re-anchor the company into a different peer set, attract a different investor base, and create strategic room that did not previously exist.

The mechanics are cheap relative to the impact. The NVIDIA-Nokia equity placement in 2025 cost roughly $1bn and produced a market cap delta several times that. No diligence, no integration, no carve-out. Pure perception change. That is the canonical specimen.

The risk is not in the mechanics. It is in the clock that opens the moment the announcement lands. The market grants provisional belief and then tests it. If proof arrives in time, the valuation holds and compounds. If not, the unwind is fast and often more damaging than if nothing had been attempted. Clock length depends on falsifiability. "We are now an AI company" buys years. "We will be number one in cloud by 2027" buys quarters. Boards self-select toward harder-to-disprove narratives. The discipline erodes accordingly.

Failed narrative moves carry a compounding credibility penalty. Multiple attempts in succession signal operational weakness. The market starts pricing the management team, not the business. The asymmetry should make boards more conservative about the front-stage move than its mechanics suggest.

The economics surface

The economics surface is the slow work of cost discipline, process, portfolio choice, and incremental product gain. It is fully within management control and compounds reliably. It rarely changes perception quickly enough to trigger a re-rating. Markets struggle to see it until it is already in the numbers. By then, much of the value is captured. Companies that rely only on operational grind stay stuck in low-multiple equilibria even as they improve.

Capital allocation events sit on the economics surface with a structural trigger attached. Spin-offs, divestitures, large buybacks. The arithmetic is the announcement. The substance is operational programme. Pre-deal carve-out, post-deal integration, transitional service agreements, customer reassignment. The re-rating arrives with the trigger. The work follows for years. Carve-outs and integrations are the highest-failure-rate part of the pattern. The org absorbs the lift while running the business. Capital allocation is a timing call as much as a structural one. Many of the cleanest re-ratings in mid-cap industrials over the last decade have come from this class of move. So have many of the most expensive failures.

How the surfaces interact

Narrative moves are financing instruments with three distinct uses.

The first is buying time for back-stage work on the same business. The narrative pulls valuation forward and creates room to do the integration, capability build, or product work that justifies the new story. Proof points are visible in the next two to three quarters under disciplined disclosure.

The second is providing currency for a structural move that reshapes the business. BT Sport in 2013 was rational triple-play defence at the time it was disclosed. By 2016 it had silently become the convergence currency that financed the EE acquisition and reshaped BT from fixed-line telco to converged operator. The narrative did not change the business. It financed the move that did.

The third is harvesting operational work already in flight. Nokia's Infinera acquisition is the structural move with proof loading into the numbers. The NVIDIA placement is the harvest mechanism, timed to coincide. Belief and arithmetic arrive together. Lower variance than narrative-first.

These three uses have different half-lives and different exit conditions. Boards conflate them and hold past purpose.

Disclosure debt

Every narrative move accumulates disclosure debt over time. The half-life of the public rationale is shorter than the half-life of the asset, and the gap widens silently.

Two patterns produce the gap. The first is oversell at announcement. A move financed by a story the board does not fully believe has no clean exit. Naming the real function retroactively invalidates the disclosed rationale. The position becomes structurally trapped. The second, more common, is disclosure drift. The move was honestly framed for the problem in front of the board. Then the problem changed. The rationale was never refreshed. There is no natural moment to re-disclose, and any attempt looks like retrofitting.

BT Sport is the canonical case of drift. Rational at entry against the 2013 triple-play problem. Useful in the middle as convergence currency for EE. Trapped at the end, when the original rationale was a fossil and the standalone economics never worked. The Warner Bros sale in 2024 was a forced unwind, more than a decade after the move had done its actual job. The error was not the purchase. It was the absence of a mechanism to retire the position as its function evolved.

The half-life of a narrative move is bounded by the half-life of its public rationale, not its strategic function. Boards that want optionality on retirement need to disclose narrowly enough at inception that the position can be wound down without contradicting the original case. Most do the opposite. They oversell to maximise the immediate re-rate, and inherit a position they cannot close.

Two functions, not two strategies

The two surfaces are typically owned by different functions. The CEO operates on the equity story. Inputs are perception, peer set, investor base, board conviction, optionality. Outputs are moves that change what the company is understood to be. The constraint is credibility budget and the clock that opens the moment a narrative move lands.

The CFO operates on the P&L. Inputs are cost, capital, portfolio, execution. Outputs are economics that compound. The constraint is what is actually controllable inside the firm.

The two run in parallel. The CEO's job is to keep perception ahead of operations by enough to expand the option set, but not so far that the gap becomes uncloseable. The CFO's job is to close the gap before the market does it for them. Healthy companies hold both functions live and in tension. The tension is the mechanism.

Failure modes fall out cleanly. CFO-led companies are operationally sound and perceptually stuck. They compound into a low multiple and get acquired or activist-targeted because the equity story never moves. CEO-led companies without a CFO counterweight do narrative moves the numbers cannot support, and unwind under forced disclosure. The pathology is not the move. It is the missing counterparty.

Roles blur on the org chart. CFOs run M&A. CEOs run capex. The split that matters is temperamental, not titular. Operating the perception surface requires comfort with deferred proof, asymmetric information, and managing belief under uncertainty. Operating the economics surface requires the opposite disposition. The skill is not interchangeable, regardless of who carries which title. This is why CFOs rarely make good CEOs, and why CEO-to-CEO transitions inside the same firm often fail.

Stewardship is not a failure mode

The framework distinguishes three states, not two.

The first is permanently non-movable. Regulated utilities, mature staples, deep cyclicals at trough, post-crisis stabilisation, integration digestion, late-cycle cash return. The equity story is structurally pinned by the underlying. The CEO who tries to run a perception surface here is not adding value. They are introducing risk. Stewardship is the correct configuration. It is what a competent PE operating partner or CFO-as-CEO delivers.

The second is movable in principle but with no window currently available. The sector has optionality, the peer set could be contested, the future state is undetermined. But the narrative architecture for a re-rate has not yet formed. The right discipline is operational rigour with the perception surface watched, not operated. The CEO who watches the window and steps through it when it opens is doing the job.

The window opens through observable signals. Customer values shift. Customer behaviour shifts. Adjacent markets show discontinuities. Significant innovations land. The CEO's job in state two is to track these continuously and step through when they converge. Microsoft pre-Nadella sat in state two for most of a decade. By the time Nadella took over, the signals were public: enterprise cloud-spend reallocation and the AWS adjacency. Most state-two CEOs miss them because they are not looking, or read them as noise. Nadella did not.

The third is window open. Sectors at transition, technology shifts in flight, peer sets reforming. The CEO who runs only operations here is leaving the strategic surface unattended.

State two is where most of the trouble lives. A CEO judged on perception in a phase that does not permit it is structurally forced toward oversell. There is no real story available, so the CEO invents one. The disclosure debt gets baked in at announcement because the move was forced rather than chosen. This is not bad discipline. It is what happens when the board demands narrative from a phase that cannot supply it.

The pathology is mismatch. The stewardship CEO is correct when the window is closed and wrong when it is open. The perception CEO is correct when the window is open and dangerous when it is not.

The earlier cases read more cleanly through this frame. Nokia's networking business sat in state two for years. Infinera was operational lift done during the windowless phase. The NVIDIA placement was the move when AI infrastructure demand created the adjacency discontinuity. State two to state three, executed in sequence. BT Sport at entry was a state-three move correctly read: triple-play architecture reforming, Sky the contested peer. EE was the structural follow-on the Sport currency financed. The error came after integration. BT had moved back into state two with convergence in the numbers and no new narrative architecture available. Sport could not be retired without invalidating the original disclosure. State two with a perception position that cannot be closed is the trap the framework predicts.

What the board hires for

The framework generates a hiring brief, not a generic CEO search.

State one wants a steward. The skill is execution discipline, capital return, and operational rigour. The profile is interchangeable with a PE operating partner or a CFO-as-CEO. The risk is hiring someone with a perception-operator disposition. They will manufacture a surface where none exists and create disclosure debt the phase cannot service.

State three wants a perception operator. The skill is reading the equity story, spending credibility deliberately, and managing belief under uncertainty. The risk is hiring someone who cannot tolerate the absence of immediate proof, or who runs the perception surface without a CFO counterweight strong enough to close the gap.

State two wants the rarest profile. A steward who reads window-opening signals continuously, holds operational discipline through the windowless phase, and is willing to step through when the signals converge. Most candidates are one disposition or the other. The profile that combines them is hard to identify in advance because the discipline shows only under conditions boards rarely assess.

This reframes succession. Boards typically run one search against a generic CEO profile. The framework says diagnose the state first and recruit against it second. The same firm at different points in its arc needs different CEOs. Internal succession from a long-serving operator into a window-opening phase often fails because the disposition does not transfer. Founder-CEOs often outlast their useful state because the board lacks the typology to name the mismatch.

The proxy is past behaviour at sector inflection points: how the candidate read signals while they were ambiguous, what they chose when the easy move was wrong, whether their prior bets retired cleanly or trapped them. State-two discipline is observable in the record. The board has to know where to look.

Synthesis

Perception changes what the market believes the business will become. Economics changes what the business actually is. Two surfaces, two dispositions, two clocks.

Strategy is the discipline of keeping them coupled when perception is movable and the window is open. Stewardship is the discipline of running economics alone when either condition fails. The errors are treating the surfaces as substitutes, treating the dispositions as interchangeable, holding a narrative move past the half-life of its disclosed rationale, or manufacturing a window the phase does not supply.

The framework resolves a long-running false binary. Vision-led strategy treats the future as the object. Diagnostic-led strategy treats the present as the object. The dispute collapses once the surfaces are separated. Diagnosis owns the economics surface. Vision operates the perception surface. Both are required when the window is open. Diagnosis alone is correct when the window is closed. Vision without diagnosis manufactures windows the phase cannot supply. The question was never which is right. It was always which is owed by which function in which state.