Topic 10.4 — Visibility for the CEO & Senior Leaders

In one sentence: Executives and boards do not need more reports — they need a trustworthy, layered line of sight from strategic outcome to delivery reality, built on a single source of truth and honest confidence assessments that surface bad news early enough to act.

Why this matters in health and care#

A chief executive of an NHS trust or the accountable officer of an integrated care board (ICB) is answerable — to a public board, to regulators, to Parliament, and ultimately to patients — for outcomes they cannot personally observe. Between the boardroom and the ward, or between the strategy and the codebase, sit many layers, each with an incentive to round up. If visibility is poor, leaders learn of a failing digital programme only when it has already consumed the money and missed the window; if visibility is dishonest, they learn even later.

The stakes are specific to this sector. Digital programmes here carry clinical-safety risk, statutory data-protection duties, and large sums of public money whose misdirection is a matter of public record. Boards operate within corporate governance frameworks that demand documented assurance, not vibes. Good executive visibility is therefore not a nicety of management style; it is how a board discharges its legal duty of oversight and how a system avoids the failures that follow when confident reporting masks quiet collapse.

Core concepts#

What leaders actually need to see. Not activity logs, but a small, stable set: strategic outcomes (are we achieving the change we promised — see Topic 10.0 — Objectives & Key Results (OKRs)); delivery confidence (will the things in flight land, and how sure are we); risks and issues (what could stop us, ranked by impact); spend and benefits (are we on budget and is the investment paying back); and safety and compliance (are we within our clinical-safety and information-governance obligations). Everything else is detail available on request.

Reporting layers. Visibility is layered, not flat. A board sees a portfolio-level view (a page per programme, exceptions only); an executive team sees a programme-level view; a delivery lead sees the team-level view. Each layer is a summarization of the one below, and the discipline is that they must reconcile — the board's green must be the same green the team would recognize. This is the reporting expression of enterprise portfolio and programme management (see Topic 5.3 — Enterprise Project Portfolio Management (EPPM)).

Single source of truth. The numbers a leader sees should derive from one authoritative dataset — a single source of truth — not from a slide deck someone re-typed. When finance, delivery, and benefits each maintain their own version, board discussions collapse into arguments about whose figure is right. A single source of truth means the dashboard, the finance system, and the risk register draw from the same underlying records.

Delivery Confidence Assessment (DCA). The UK Infrastructure and Projects Authority (IPA) uses a Delivery Confidence Assessment — a rated judgement (typically a red/amber/green (RAG) scale) of a project's likelihood of achieving its objectives at a point in time. Crucially, it is an assessment, not a calculation: a reviewer weighs multiple factors rather than averaging individual RAG scores. It is a professional opinion about the future, made explicit.

Watermelon reporting. The characteristic failure of RAG reporting: green on the outside, red on the inside. A status turns green at each layer as it rises, because reporting red is career-risky, until the board sees an all-green portfolio over a programme that is actually failing. Defeating the watermelon is the central problem of executive visibility.

Best practices#

1. Report exceptions, not everything. A board's attention is the scarcest resource in the organization. Give it a portfolio on a page where healthy programmes are a single green line and the discussion time goes to the exceptions — the ambers and reds — with a clear ask for each. Comprehensiveness is the enemy of attention.

2. Build on a single source of truth. Invest in one authoritative data layer from which every reporting view is generated, so finance, delivery, benefits, and risk cannot diverge. When a leader asks "where does this number come from?", there should be one answer. This also makes self-serve possible (see practice 8).

3. Use structured delivery confidence assessments, not just RAG dots. Adopt an IPA-style DCA for each major programme: a rated, narrated judgement of the likelihood of success, refreshed on a cadence and at gateways. Require the narrative — the why behind the colour — because a colour with no reasoning is exactly what the watermelon exploits.

4. Engineer honesty into the reporting culture. The watermelon is a cultural artefact, not a formatting problem. Make it safe — indeed expected — to report amber and red early; treat a programme that goes red as a request for help, not a confession of failure. Independent assurance reviews (the IPA model) and rotating challenge break the incentive to round up. A leader who shoots the messenger will be the last to know.

5. Anchor everything to a board assurance framework. NHS boards use a Board Assurance Framework (BAF) — a close cousin of the project risk register — to map principal risks to the controls and assurances that manage them, with RAG-rated confidence in each. Tie digital and delivery reporting into the BAF so the board sees strategic risk, the controls in place, and the gaps in one governance instrument, satisfying its assurance duty rather than running a parallel unofficial view.

6. Define escalation paths and thresholds in advance. Decide, before the crisis, what conditions force a programme onto the board's agenda between scheduled meetings (a red DCA, a budget breach beyond tolerance, a clinical-safety incident, a slipped critical milestone). Pre-agreed thresholds turn escalation from a judgement call — which the watermelon suppresses — into a rule.

7. Set a governance rhythm and hold it. Match reporting cadence to decision cadence: a weekly delivery signal, a monthly programme review, a bi-monthly or monthly board. Stale reporting is worse than none because it invites decisions on old reality. The rhythm should feed, and be fed by, the OKR check-in cadence of Topic 10.0.

8. Offer self-serve alongside curated views. Give leaders and non-executive directors a live dashboard they can interrogate themselves, so they are not wholly dependent on the story a report tells. Balance it with a curated narrative for the board pack — self-serve answers "what is happening?", curation answers "what does it mean and what should we do?". Both are needed; neither alone suffices.

9. Tell the story with the data, honestly. A number on a slide changes no decision; a leader needs the interpretation — trend, comparison, and the "so what". Use statistical process control (see Topic 10.0) rather than month-on-month arrows so the board reacts to signal, not noise, and pair every reassuring figure with the risk it does not capture. Storytelling with data is not spin; it is making a true situation legible.

Questions to discuss with your team#

1. Where in our reporting chain is a watermelon most likely forming right now — and what would make it safe for the layer below to tell us the truth earlier? The watermelon — green on the outside, red within, ripening upward as each layer softens the status because reporting red is career-risky — is the central failure of executive visibility, and the honest version of this question names the specific programme where you suspect it rather than discussing it in the abstract. In a health and care setting the cost is not just a missed deadline but public money already spent and a clinical or access benefit that never materialized, discovered too late to redirect. The tension to debate is that culture, not format, drives the watermelon: a leader who has ever shot the messenger has taught every layer beneath to round up, and no dashboard redesign will undo that lesson. Concrete angles include whether your recent red or amber statuses arrived early enough to act, whether independent assurance or rotating challenge exists to break the incentive to soften, and whether a team lead genuinely believes an early amber will be met with help rather than blame. A good answer is uncomfortable — it identifies a real candidate watermelon, admits where the culture punishes bad news, and commits to a specific change (a safe-to-escalate norm, an independent reviewer, a pre-agreed threshold) rather than a platitude about openness.

2. When a leader asks "where does this number come from?", can we give one answer — and where do our finance, delivery, and benefits figures currently diverge? A single source of truth is what stops board discussions collapsing into arguments about whose spreadsheet is right, yet most organizations discover on inspection that the "beds saved" or "savings realized" figure on the board slide was re-typed from a deck and no longer reconciles with the finance ledger or the delivery tracker. The exercise is to trace one or two of your most consequential board numbers back to their origin and see whether every reporting view genuinely draws from the same underlying records or from parallel, quietly diverging copies. The trade-off worth debating is the real cost of building that authoritative data layer against the recurring cost of boards litigating data provenance instead of making decisions — and whether "good enough for now" reconciliation is masking a figure that can be re-spun at will. Consider who owns each number today, how often the versions are checked against each other, and what a leader currently does when two sources disagree. A strong answer locates at least one number that cannot yet be traced to a single origin, names the divergence, and decides who is accountable for closing it so that self-serve and curated views can safely share one foundation.

3. Does every RAG rating we report carry the reasoning behind its colour, and have we agreed in advance what forces a programme onto the board's agenda between meetings? A colour with no narrative is unfalsifiable and is precisely the habitat the watermelon thrives in; an IPA-style Delivery Confidence Assessment is valuable because it is a rated professional judgement with its rationale made explicit, not an average of sub-scores that hides the reasoning. Test your current reporting: could a non-executive interrogate the "why" behind an amber, or does the colour arrive naked and unchallengeable? The second half matters just as much — deciding before the crisis what conditions (a red DCA, a budget breach beyond tolerance, a clinical-safety incident, a slipped critical milestone) trigger escalation turns raising the alarm from a judgement call the watermelon suppresses into a rule that fires automatically. The tension is that narrated assessments and pre-agreed thresholds feel like bureaucratic overhead until the month a programme fails quietly for want of them, and busy reviewers resent writing the rationale. Concrete evidence to bring includes a recent status that changed with no explanation, and any past escalation that happened too late because it depended on someone's discretion. A good answer commits to requiring the narrative behind every rating and to a written set of escalation thresholds the board has actually endorsed, so bad news reaches decision-makers while there is still time to act.

4. Do our reporting layers actually reconcile — would the board's green be recognized as green by the team doing the work — and is our board pack short enough to command attention? Visibility is layered rather than flat: the board sees a portfolio-on-a-page, the executive a programme view, the delivery lead a team view, and the whole edifice only works if each layer is a faithful summarization of the one beneath it. The discipline this question tests is reconciliation — whether the confident green a board reads is the same reality a ward manager or an engineer would recognize, or whether meaning has quietly leaked out at each hand-off. Alongside fidelity sits the scarcity of board attention: a wall of sixty undifferentiated metrics buries the two that matter, so exception-based reporting that shows healthy programmes as a single line and spends discussion time on the ambers and reds is not laziness but focus. The tension is real — summarizing for the board risks smoothing over the texture that would reveal trouble, yet refusing to summarize drowns the board in detail it cannot act on. Concrete angles include picking one recent green status and walking it down to the delivery team to see if it survives, and counting how many metrics your last board pack actually carried. A good answer commits to a layered structure that reconciles top to bottom and to a board pack disciplined enough that exceptions stand out rather than hiding in the noise.

5. Are our non-executives and senior leaders able to interrogate the data themselves, and do we present signal rather than month-on-month noise? A board that can only see the story a report chooses to tell is captive to the narrator, which is why live self-serve dashboards for non-executive directors matter — they let independent challenge probe beneath the curated narrative rather than accepting it whole. But self-serve alone is not enough: raw numbers without interpretation change no decision, and a leader confronted with a month-on-month arrow will react to noise as if it were a trend, chasing a single bad month or celebrating a single good one. This is where statistical process control earns its place, distinguishing a genuine shift from ordinary variation so the board reacts to signal, and where honest storytelling with data — trend, comparison, and the "so what" — makes a true situation legible without spinning it. The tension worth surfacing is that self-serve access can expose leaders to unmediated numbers they misread, while over-curation slides into spin; both a live view and a curated narrative are needed, and neither alone suffices. Consider whether your non-executives have ever found something the executive did not flag, and whether your board reacts to variation as if every wobble were meaningful. A good answer gives independent access real teeth while pairing it with interpretation that separates signal from noise.

6. Do the metrics we put in front of the board measure outcomes — including benefits realized, safety and compliance — or do they flatter us with activity counts? The vanity trap is reporting what is easy to count ("users onboarded", "systems deployed") instead of the strategic outcome the investment was meant to buy ("beds saved", "admissions avoided", "waits shortened"), and in health and care that gap is not cosmetic because public money is justified by patient benefit, not by activity. This question also forces the two dimensions boards most often under-report to the surface: whether the promised benefits are actually being realized against the business case, and whether the programme remains within its clinical-safety and information-governance obligations — a green delivery status over a live safety or data-protection risk is its own kind of watermelon. The trade-off to debate is that outcome and benefit measures are slower, noisier and harder to attribute than activity counts, so teams gravitate to the easy number precisely when the hard one matters most. Concrete angles include auditing your current board metrics to see how many are activity versus outcome, checking whether benefits are tracked past go-live or quietly abandoned, and confirming safety and compliance appear on the board's line of sight rather than only in a separate committee. A good answer re-anchors every board metric to a strategic outcome, keeps benefits realization visible after delivery, and ensures safety and compliance are reported to the board rather than assumed.

In practice: a health & care example#

An ICB is investing £14m over three years in a virtual-ward programme to move eligible patients from hospital beds to monitored care at home. The board has been receiving a monthly slide: three programmes, all green, a headline of "beds saved". For fourteen months it is reassured.

At month fifteen the programme misses its bed-reduction target badly, and it emerges the amber had been known at team level for two quarters — the community-nursing capacity to staff the wards was never secured, but each layer softened the status upward. Classic watermelon.

The new chief digital officer rebuilds visibility. A single source of truth links the delivery tracker, the finance ledger, and the benefits model, so "beds saved" is computed once and cannot be re-spun in a deck. Each programme carries an IPA-style Delivery Confidence Assessment with a written rationale, refreshed monthly and challenged quarterly by an independent reviewer. The board pack becomes a portfolio-on-a-page: two greens shown as one line each, the virtual-ward programme in amber with a specific ask — fund the nursing gap or rescope. Escalation thresholds are agreed: any red DCA or any staffing shortfall beyond 10% comes to the board within a week, not at the next monthly cycle. Benefits are plotted on an SPC chart so a single good month is not mistaken for recovery. Non-executive directors get a live self-serve dashboard for their own probing, and the curated narrative explains what the numbers mean. The programme does not instantly succeed — but the board now sees the truth early enough to redirect money and demand the staffing fix while it still matters.

Sector lenses#

Startup#

In a digital-health startup the "board" is a founder, a handful of investors, and a monthly update, so the visibility problem is inverted: the danger is not layers of watermelon but a single optimistic narrator with no independent check. Keep one honest dashboard — runway, active users, uptime, open safety issues — drawn from a single source of truth rather than a hand-built board deck, and report amber early to investors because a seed-stage company that hides a slipping metric simply runs out of cash in the dark. A lightweight delivery-confidence line ("will we hit the pilot by Q3, and how sure are we") is enough; a full IPA-style DCA would be theatre at this scale. The cultural work is making it safe for a five-person team to tell its backers bad news while there is still time to pivot.

Small business#

An established small provider — a GP practice, a community pharmacy, a care home or domiciliary-care service, a single-site clinic, or a small health-tech supplier past the startup stage — has real patients and ongoing operations but little or no dedicated IT capability and tight cash, so executive visibility falls to a practice manager or registered manager as one job among many. The point is not a layered board apparatus, which there is no one to staff, but a single honest weekly glance at the few things that can hurt patients or the business: open safety and information-governance issues, whether a supplier's system is actually delivering, and whether a funded digital change is on track or quietly stalling. Because the practice carries the same clinical-safety and data-protection duties as a large trust with none of the assurance machinery, the discipline is to write down a short escalation rule — what forces a problem to the partners or the owner this week rather than at the next quarterly review. A cheap dashboard drawn straight from the clinical or operational system, rather than a hand-typed spreadsheet, keeps the one source of truth honest without a reporting team. The cultural work is making it safe for a receptionist or a care worker to flag a slipping problem early, because in a small team the person who sees it first is often the only one who will.

Enterprise#

An NHS trust or integrated care board has exactly the layered structure the watermelon exploits — team to programme to portfolio to board — so enterprises need the full apparatus: a single authoritative dataset feeding every view, narrated DCAs per major programme, exception-based portfolio-on-a-page reporting, and independent assurance reviews that break the incentive to round up. The Board Assurance Framework is the statutory anchor, and digital delivery risk must map into it rather than run as a parallel unofficial view. Non-executive directors should get self-serve access so they are not captive to the executive's story. The reporting cadence must reconcile with the OKR and governance rhythm (see Topic 10.0 — Objectives & Key Results (OKRs)) so the board's green is the same green the ward or the delivery team would recognize.

Government#

A national or local public body — NHS England, DHSC, UKHSA, or a local authority — reports not only to a board but to Parliament, the National Audit Office, and the public, so visibility carries a formal accountability that the IPA's Delivery Confidence Assessment was designed to serve. Major programmes face published assurance reviews and gateway ratings, and a red DCA can become a matter of public record, which raises the stakes on honest reporting and lowers the temptation to bury bad news. Escalation thresholds and independent challenge are not optional management hygiene here but part of discharging a documented duty of oversight over public money. The discipline is transparency by default: reporting that would survive external audit, with every RAG rating carrying the reasoning a reviewer could interrogate.

Common failure modes#

• Watermelon reporting. Green-outside, red-inside status that ripens upward. Defeat it with narrated DCAs, independent assurance, and a culture that rewards early amber.
• Multiple sources of truth. Finance, delivery, and benefits each with their own figure, so board meetings argue about data instead of deciding. Fix with one authoritative dataset.
• Death by dashboard. A wall of 60 metrics with no hierarchy, so nothing stands out and the board disengages. Fix with exception-based, portfolio-on-a-page reporting.
• RAG without rationale. A colour with no "why" — unfalsifiable, and the watermelon's favourite habitat. Require the narrative behind every rating.
• Cadence drift. Reports that arrive weeks stale, driving decisions on old reality. Fix by matching reporting rhythm to decision rhythm and holding it.
• Vanity in the boardroom. Reporting activity ("users onboarded") instead of outcome ("beds saved, admissions avoided"). Anchor every board metric to a strategic outcome (Topic 10.0).
• No pre-agreed escalation. Relying on someone's judgement to raise the alarm — precisely the judgement the watermelon suppresses. Set thresholds in advance.

Maturity model#

Checklist#

• Board reporting is exception-based: a portfolio on a page, healthy programmes as one line, attention on the reds and ambers.
• Every reported number derives from a single authoritative source of truth.
• Each major programme carries a narrated Delivery Confidence Assessment, refreshed on a set cadence.
• Every RAG rating is accompanied by the reasoning behind the colour.
• Digital and delivery risk is mapped into the Board Assurance Framework.
• Escalation thresholds (red DCA, budget breach, safety incident, slipped milestone) are agreed in advance.
• Governance cadence matches decision cadence and is honoured; reports are current, not stale.
• Leaders and non-executives have a self-serve dashboard alongside the curated board narrative.
• Board metrics report outcomes (Topic 10.0), not activity or vanity counts.
• The culture makes reporting amber and red early safe and expected, backed by independent assurance.

Key sources#

• UK Infrastructure and Projects Authority (IPA), Project Assurance Reviews — Delivery Confidence Guide for Review Teams — the Delivery Confidence Assessment as a rated professional judgement, not a calculated average. https://www.gov.uk/government/organisations/infrastructure-and-projects-authority
• NHS England, NHS Oversight Framework 2025/26 — how NHS bodies are assessed and held to account. https://www.england.nhs.uk/long-read/nhs-oversight-framework-2025-26/
• Board Assurance Framework (BAF) — the NHS governance instrument mapping principal risks to controls and assurances (see e.g. NHS Counter Fraud Authority and trust BAFs). https://cfa.nhs.uk/about-nhscfa/NHSCFA-board-of-directors/board-assurance-framework
• Kaplan & Norton, "The Balanced Scorecard," Harvard Business Review (1992) — multi-perspective performance reporting.
• NHS England, Making Data Count — statistical process control for board and executive data, distinguishing signal from noise.
• Cole Nussbaumer Knaflic, Storytelling with Data — communicating a true situation so it drives a decision.
• Management of Portfolios (MoP) and Managing Successful Programmes (MSP) — layered portfolio/programme reporting and assurance (see Topic 5.3 — Enterprise Project Portfolio Management (EPPM)).

References#

1. Infrastructure and Projects Authority — Wikipedia — https://en.wikipedia.org/wiki/Infrastructure_and_Projects_Authority
2. Single source of truth — Wikipedia — https://en.wikipedia.org/wiki/Single_source_of_truth
3. Traffic light rating system — Wikipedia — https://en.wikipedia.org/wiki/Traffic_light_rating_system
4. Corporate governance — Wikipedia — https://en.wikipedia.org/wiki/Corporate_governance
5. Risk register — Wikipedia — https://en.wikipedia.org/wiki/Risk_register
6. Dashboard (computing) — Wikipedia — https://en.wikipedia.org/wiki/Dashboard_(computing)
7. Statistical process control — Wikipedia — https://en.wikipedia.org/wiki/Statistical_process_control
8. Project Assurance Reviews — Delivery Confidence Guide for Review Teams — UK Infrastructure and Projects Authority — https://www.gov.uk/government/organisations/infrastructure-and-projects-authority
9. NHS Oversight Framework 2025/26 — NHS England — https://www.england.nhs.uk/long-read/nhs-oversight-framework-2025-26/
10. Board Assurance Framework — NHS Counter Fraud Authority — https://cfa.nhs.uk/about-nhscfa/NHSCFA-board-of-directors/board-assurance-framework
11. The Balanced Scorecard — Kaplan & Norton, Harvard Business Review (1992) — https://hbr.org/1992/01/the-balanced-scorecard-measures-that-drive-performance-2
12. Making Data Count — NHS England — https://www.england.nhs.uk/publication/making-data-count/
13. Storytelling with Data — Cole Nussbaumer Knaflic — https://www.storytellingwithdata.com/books