Topic 10.0 — Objectives & Key Results (OKRs)

Objectives and Key Results give your organization a disciplined way to turn strategy into a small number of ambitious, outcome-focused goals — but in health and care their real value is forcing you to ask what better looks like for patients, not just what your teams are busy doing.

Why this matters in health and care#

Digital delivery in health and care rarely fails for lack of activity. Teams are busy: shipping features, closing tickets, running programs, attending boards. The harder question — the one that separates organizations that improve care from those that merely spend money on technology — is whether all that motion is pointed at outcomes that matter to patients, clinicians, and the public purse. Objectives and Key Results (OKRs) exist to make that alignment explicit, and to make it honest.

In a commercial company, a misaligned goal wastes money. In health and care, it can waste the one thing the system has least of — clinical time and patient trust. A digital team that hits every delivery milestone but never moves a single measure of access, safety, or equity has succeeded at the wrong thing. Because you steward public money and, indirectly, clinical outcomes, the discipline of naming what you are actually trying to achieve — and being willing to be graded honestly against it — is not a management fashion. It is part of how you justify your existence to boards, regulators, and the people you serve.

OKRs also carry a specific risk in this sector that deserves naming up front. Health and care is heavily target-driven, and staff are rightly wary of any new goal-setting language that smells like a performance-management stick. Used well, OKRs create focus, transparency, and permission to attempt ambitious change. Used badly, they become another top-down cascade of quotas that punish honesty and reward gaming. This topic is about getting the first version and avoiding the second.

Core concepts#

Objectives and Key Results is a goal-setting framework that pairs a single qualitative Objective — an ambitious, memorable statement of what you want to achieve — with two to five quantitative Key Results that measure whether you got there. The Objective answers "where do we want to go?"; the Key Results answer "how will we know we have arrived?" A well-formed set reads as a sentence: we will achieve this Objective, as measured by these Key Results.

The method has a clear lineage. It grew from management by objectives, the approach Peter Drucker popularized in the 1950s, and was sharpened into its modern form by Andrew Grove at Intel, who described it in his 1983 book High Output Management. It reached wide adoption when John Doerr, who had learned the method from Grove, introduced it to Google in 1999; Doerr later codified the approach for a general audience in his book Measure What Matters.

The single most important concept is the distinction between an OKR and a key performance indicator (KPI). OKRs drive focus and change — they name the handful of things you want to be different by the end of the quarter or year. KPIs monitor the ongoing health of business-as-usual — the vital signs you watch continuously to make sure nothing breaks. Topic 10.1 — Key Performance Indicators (KPIs) covers measurement in depth; this topic is about goal-setting. Confusing the two is the most common OKR mistake, and we return to it throughout.

Two further distinctions shape practice. First, committed versus aspirational (stretch) OKRs: committed OKRs are commitments you fully expect to deliver, while aspirational OKRs are deliberate stretches where landing at 60–70 per cent is a success. Second, outcomes versus tasks: a Key Result should measure a change in the world ("reduce the median wait for first digital triage from nine days to three"), not the completion of an activity ("launch the new triage form"). This connects directly to the SMART criteria for well-formed goals and, in health and care, to clinical and population-health outcomes.

The framework depends on alignment and transparency without a rigid top-down cascade. Everyone can see everyone else's OKRs, and team goals connect visibly to organizational strategy — but teams set their own OKRs rather than receiving them as quotas. Finally, stretch goals only work under psychological safety: if missing an aspirational target is punished, people set timid goals and the whole method collapses into theater. The failure of gamed targets is well described by Goodhart's law — when a measure becomes a target, it ceases to be a good measure.

Best practices#

1. Write the Objective to inspire and the Key Results to measure. The Objective should be a short, qualitative, motivating sentence a tired clinician could remember — "Make it genuinely easy to get community physiotherapy" — not a metric. The Key Results then make it unambiguous by attaching numbers: the change in wait time, the change in did-not-attend rate, the change in access for people who cannot use online tools. Keep to one Objective per team per quarter where you can, and never more than a few, so focus is real rather than nominal.

2. Express Key Results as outcomes, not tasks. "Ship the new referral template" is a task — you can complete it and change nothing. "Cut the proportion of referrals returned for missing information from 30 per cent to under 10 per cent" is an outcome. Outcome Key Results force you to have a theory about why your work matters and expose it to reality. When a Key Result can only be written as a deliverable, it usually belongs on a roadmap or delivery plan, not in your OKRs.

3. Limit the number ruthlessly. The power of OKRs comes from what they exclude. A team with one Objective and three Key Results has decided what matters most this quarter; a team with five Objectives and twenty Key Results has decided nothing and merely inventoried its backlog. Resist the organizational instinct to make every worthy activity an OKR. Most of what a good team does is business-as-usual, measured by KPIs, and should stay off the OKR list.

4. Distinguish committed from aspirational OKRs, and say which is which. Some goals you must hit — a regulatory deadline, a clinical-safety commitment — and these are committed OKRs where anything below 100 per cent is a problem. Others are deliberate stretches where 100 per cent would prove you set the bar too low. Label each type explicitly so teams and boards read the scores correctly, and so a 0.6 on a stretch goal is celebrated as ambition rather than mistaken for failure.

5. Align through transparency, not a rigid cascade. Publish every team's OKRs where colleagues can see them, and make sure each set connects visibly to the organization's strategy (see Topic 1.2 — Digital Strategy & Roadmapping). But let teams draft their own OKRs and negotiate them upward, rather than receiving decomposed quotas from above. A mix of top-down direction and bottom-up ownership produces goals people believe in; a pure cascade produces goals people comply with and quietly resent.

6. Run a deliberate cadence of setting, check-in, and grading. Set OKRs at the start of a cycle — quarterly for teams is common, annually for the organization. Check in frequently (weekly or fortnightly) on confidence and blockers, not just at the end. Grade honestly at the close, usually on a 0.0–1.0 scale, and treat the grade as a prompt for a learning conversation, not a verdict. The reflection at grading — what we learned, what we would change — is where most of the value lives.

7. Connect OKRs to strategy and to benefits. OKRs should be the near-term, measurable expression of your longer-term strategy and roadmap (Topic 1.2), and they should feed the benefits you have committed to in your business cases (see Topic 6.2 — Benefits Realization & Value Management). When an Objective cannot be traced to a strategic aim or a promised benefit, question why it exists. This traceability is also what makes your OKRs credible to a board and useful for Topic 10.4 — Visibility for the CEO & Senior Leaders.

8. Anchor Key Results in clinical, operational, and population-health outcomes where you can. A digital team's OKRs are stronger when at least some Key Results reach through to outcomes clinicians and patients recognize — waiting times, safety events avoided, uptake among under-served groups — rather than stopping at system metrics like uptime or logins. Use the analytics capability described in Topic 4.5 — Data, Analytics & Population Health Management to source these honestly. Where a true outcome is too slow or noisy to measure in a quarter, choose a well-evidenced leading indicator and be explicit that it is a proxy.

9. Protect psychological safety so stretch goals stay safe. Aspirational OKRs only work if people can miss them, say so plainly, and learn without penalty. Separate OKR grading from individual performance appraisal, pay, and ranking — the moment a Key Result determines someone's bonus, it stops measuring reality and starts measuring self-protection. Say this explicitly and repeatedly, because skeptical staff will assume the worst until the organization proves otherwise over several cycles.

Questions to discuss with your team#

1. Which of our current goals are genuinely OKRs, and which are actually KPIs or task lists in disguise? This is the foundational sorting exercise, and most teams get it wrong the first time. Walk through everything you are currently calling a goal and ask, for each, whether it names something you want to change (a candidate OKR) or something you want to keep healthy (a KPI, belonging in Topic 10.1 — Key Performance Indicators). Then check whether each supposed OKR is written as an outcome or merely as a task you could tick off without improving anything. The honest result is usually uncomfortable: a long list of activities, a few real KPIs, and only two or three genuine change-focused Objectives. That narrowing is the point. A good answer leaves you with a short, defensible set of OKRs, a separate dashboard of business-as-usual KPIs, and a delivery plan for the tasks — three different things that were previously tangled together.

2. Are our stretch goals safe to miss, and does everyone actually believe that? Aspirational OKRs are supposed to encourage ambition, but they only do so if the organization has genuinely decoupled missing a stretch from being punished. Discuss candidly whether your staff — especially clinicians and delivery leads scarred by target regimes — believe that landing at 0.6 on an aspirational Key Result will be read as healthy ambition or as failure to be explained to a committee. Look at where OKR grades actually flow: do they touch appraisals, pay, contract reporting, or public league tables? If they do, people will rationally set timid goals and game the numbers, exactly as Goodhart's law predicts. A good answer names concrete protections — grades kept out of individual performance management, stretch OKRs explicitly labeled, leaders modeling honest misses — and acknowledges that trust here is earned slowly, over several cycles, not declared in a policy.

3. Do our OKRs reach through to outcomes that patients and clinicians would recognize, or do they stop at things only our team cares about? It is easy to write Key Results about logins, releases, and uptime because they are inside your control and quick to measure. It is harder, and far more valuable, to connect at least some Key Results to waiting times, safety, health outcomes, or equity of access. Discuss which of your Objectives could be expressed in terms a ward sister or a patient would find meaningful, and where you are hiding behind system metrics because the real outcome is slow, noisy, or politically awkward to own. Consider the equity dimension explicitly: an OKR that improves an average while widening the gap for digitally excluded people has arguably made care worse, not better (see Topic 1.9 — Digital Inclusion & Accessibility). A good answer includes at least one outcome-level Key Result per Objective, is honest about which measures are proxies, and has a clear line back to strategy (Topic 1.2) and to promised benefits (Topic 6.2).

4. How few OKRs are we actually willing to commit to, and what are we consciously choosing not to pursue this cycle? The discipline of OKRs lives in subtraction, yet almost every team drifts towards more Objectives rather than fewer, because saying no to a worthy idea feels like neglect. Discuss what would happen if you allowed yourselves only one Objective and three Key Results for the quarter: which of your candidate goals survive, and who would be disappointed by the ones that fall away. In health and care this conversation is politically charged, because competing services, clinical directorates, and commissioners all press for their priority to be named, and a digital team can end up with an OKR list that is really a peace treaty between stakeholders rather than a statement of focus. Be honest about the cost of spreading thin: a team nominally pursuing eight Objectives is usually making real progress on none, while patients wait for the one improvement that mattered most. A good answer names the two or three things that genuinely matter this cycle, records the worthy ideas you are explicitly parking, and can defend the exclusions to the people who lobbied for them.

5. When an OKR score comes back low, what do we actually do with it — and what does that reveal about the culture we have built? The grade itself matters far less than the conversation it triggers, and the nature of that conversation is the truest test of whether OKRs are working. Talk through a concrete recent miss, or imagine one: does the discussion turn to what you learned and what you would change, or to who is accountable and how the number will be explained upward? In a sector shaped by decades of performance management, the reflex to treat a low score as a failure to be defended is strong, and it quietly teaches people to set safer goals next time. Consider who is in the room when grades are reviewed, whether the tone differs for committed versus aspirational OKRs, and whether anyone has ever been visibly rewarded for an honest, instructive miss. A good answer describes grading as a structured learning ritual, points to real examples where a low score changed the next cycle's plan, and can show that reporting a miss plainly has never cost anyone their standing.

6. How do our OKRs behave when the world changes mid-cycle — a winter crisis, a safety incident, a funding shift — and can we re-point them without abandoning the discipline? OKRs are usually set for a quarter or a year, but health and care rarely respects a planning calendar, and a rigid set of goals can become either an irrelevance teams ignore or a straitjacket that keeps them working on yesterday's priority while a crisis unfolds. Discuss what your team would do if, six weeks into a cycle, an urgent operational pressure made your headline Objective suddenly less important than keeping a service safe. The tension is real: change OKRs too readily and they lose all meaning as commitments; hold them too rigidly and you privilege a plan over patients. Explore who has the authority to re-open a set mid-cycle, how you would record the change and its rationale, and how you avoid using "circumstances changed" as a cover for quietly dropping goals you were going to miss anyway. A good answer establishes a clear, sparing mechanism for mid-cycle change, distinguishes genuine re-prioritization from face-saving, and keeps a visible trail so the shift is transparent to the board and to other teams that were aligned to the original goal.

In practice: a health & care example#

An integrated care board (ICB) has a digital team supporting community services across several boroughs. For years the team reported activity: features delivered, tickets closed, systems patched. The board could see it was busy but could not see whether it was helping. A new director introduces OKRs, starting with a single Objective for the community physiotherapy service, chosen because waits are long and complaints are rising.

The Objective is deliberately human: Make it genuinely easy and fair to get community physiotherapy. The team drafts three Key Results as outcomes, not tasks. First, reduce the median time from referral to first meaningful contact from twenty-one days to seven. Second, hold the did-not-attend rate at or below its current level, so speed is not bought with chaos. Third — the equity Key Result — ensure that first-contact waits for patients in the two most deprived neighbourhoods are no worse than the service average, disaggregated and published. They mark the first two as committed and the wait-time target as aspirational, since seven days is a genuine stretch.

Crucially, the director separates these OKRs from the team's KPIs. System uptime, form-completion rates, and support-ticket volumes stay on a business-as-usual dashboard, watched continuously but never confused with the quarter's goals. The OKRs are published on the intranet alongside every other team's, so the physiotherapy service leads, the analytics team, and the board can all see them and see how they connect to the ICB's access strategy.

The cadence does the work. At fortnightly check-ins the team reports confidence, not just progress, and surfaces blockers early — the biggest turns out to be triage sitting in a shared inbox, not the online form everyone assumed was the problem. By quarter-end the median wait has fallen to nine days: a 0.7 on the aspirational Key Result. Rather than treat that as failure, the grading conversation examines what the remaining bottleneck is and rolls a refined Objective into the next quarter. Because no one's appraisal hangs on the score, the team reports the miss plainly, and the honesty makes the next set of OKRs more ambitious, not less.

Sector lenses#

Startup#

A health-tech startup lives and dies by focus, which makes OKRs a natural fit — but the classic trap is turning them into a feature backlog. With a small team and investor pressure, every Key Result is tempted to become "ship X," measuring output rather than whether customers or patients are better off. Resist this: pick one or two Objectives per quarter and write at least one Key Result as an outcome your NHS buyers would recognize, such as adoption by real clinicians or a measurable change in a care pathway. Lightweight cadence — a shared doc, a weekly confidence check — is enough; heavy tooling is not the point. Above all, keep OKR grades away from a tiny team's sense of personal worth, so honest misses stay possible when you most need the truth.

Small business#

An established small provider — a GP practice, a community pharmacy, a care home, or a small supplier with real customers but no dedicated digital function — needs OKRs that fit around a working day, not a planning department. The discipline here is to pick a single Objective the whole team can hold in their heads, such as making it easier for patients to get repeat medication, and to express it with one or two outcome Key Results drawn from data you already collect, rather than building new dashboards you have no time to maintain. Cadence should be minimal but real: a ten-minute check at an existing team meeting beats a formal quarterly ceremony no one has capacity to run. The trap to avoid is letting the OKR become another compliance task layered on top of ongoing operations, because a small business carries the same regulatory and safety duties as a large trust with a fraction of the staff to meet them. Keep grading light and blameless, because in a team of a handful of people a punitive score poisons everything, and one honest, well-chosen Objective delivered is worth more than a suite of goals that quietly lapse.

Enterprise#

A large NHS trust, ICS, or established supplier has the opposite problem: OKRs proliferate and calcify into a governance ritual. Dozens of teams each produce dozens of OKRs, a cascade decomposes strategy into quotas, and the whole thing merges with performance management until it becomes a target regime in a new costume. The enterprise discipline is subtraction — a genuinely small number of organizational Objectives, teams drafting their own aligned OKRs rather than receiving them, and a firm firewall between OKR grades and appraisal, pay, or board judgment. Invest in the check-in and grading cadence, not just the annual set, and use transparency across teams to reduce duplication. Keep the vast body of business-as-usual firmly in KPI-land (Topic 10.1), so OKRs stay about change.

Government#

A national or regional body — NHS England, a devolved government, an arms-length body — shapes OKRs mainly through the goals it sets and the behaviors it models. The dominant risk is confusing an OKR with a national target: a stretch Objective published centrally and then enforced as a mandatory quota will be gamed exactly as decades of NHS target experience predict. Government's better role is to set a few clear strategic Objectives, give local organizations room to write their own Key Results, and resist ranking bodies on two-point movements. Where it does use OKRs internally, it should be scrupulous about psychological safety and about separating goal-setting from assurance and accountability regimes. Modeling honest grading at the center gives everyone below permission to do the same.

Common failure modes#

• OKRs as a task list. Writing Key Results as deliverables ("launch the app") rather than outcomes ("reduce waits"), so the team can score full marks while changing nothing for patients.
• Too many OKRs. Turning the whole backlog into Objectives, destroying the focus that is the method's entire point, and leaving everyone unable to say what matters most.
• Confusing OKRs with KPIs. Putting business-as-usual health metrics into OKRs, or vice versa, so neither the change goals nor the operational vital signs are managed properly (see Topic 10.1).
• Weaponizing OKRs for performance management. Linking grades to appraisal, pay, or league tables, which kills honesty, drives sandbagging, and triggers Goodhart's law — people optimize the number, not the outcome.
• Sandbagged goals. Setting only safe, easily achievable Objectives because the culture punishes misses, so OKRs stop driving any real ambition or change.
• Set and forget. Writing OKRs at quarter-start and ignoring them until grading, skipping the check-ins where learning and course-correction actually happen.
• Orphaned OKRs. Objectives that trace to no strategy (Topic 1.2) and no benefit (Topic 6.2), existing because a team wanted to look purposeful rather than because they matter.
• Output-only Key Results. Measuring logins, releases, and uptime because they are easy and controllable, never reaching through to clinical, operational, or equity outcomes.

Maturity model#

Checklist#

• Each Objective is a short, qualitative, inspiring sentence, and we have very few of them.
• Each Objective has two to five Key Results, written as outcomes rather than tasks or deliverables.
• We have sorted our goals into OKRs (change) and KPIs (business-as-usual), and keep them separate.
• Every committed and aspirational (stretch) OKR is explicitly labeled as such.
• Every Objective traces to our strategy (Topic 1.2) and, where relevant, to a promised benefit (Topic 6.2).
• At least one Key Result per Objective reaches a clinical, operational, or population-health outcome.
• We have an equity lens on our Key Results so we do not improve an average while widening a gap.
• OKRs are published transparently so teams can see and align with each other, not cascaded as quotas.
• We run a reliable cadence of setting, frequent check-ins, and honest end-of-cycle grading.
• OKR grades are firewalled from individual appraisal, pay, and any league table.
• The board sees OKRs framed as focus and change, distinct from the KPI health dashboard.

Key sources#

• John Doerr — Measure What Matters, the standard modern text on OKRs, with health and non-profit case studies.
• Andrew Grove — High Output Management, the original articulation of Objectives and Key Results at Intel.
• Christina Wodtke — Radical Focus, a practical guide to running OKRs with cadence and focus.
• NHS England — A plan for digital health and social care and related strategy, for connecting local OKRs to national aims.
• GOV.UK Service Manual (Government Digital Service) — guidance on setting measurable goals and performance metrics for digital services.
• Amy C. Edmondson — The Fearless Organization, on the psychological safety that stretch goals require.

References#

1. Objectives and key results — Wikipedia — https://en.wikipedia.org/wiki/Objectives_and_key_results
2. Management by objectives — Wikipedia — https://en.wikipedia.org/wiki/Management_by_objectives
3. Andrew Grove — Wikipedia — https://en.wikipedia.org/wiki/Andrew_Grove
4. High Output Management — Wikipedia — https://en.wikipedia.org/wiki/High_Output_Management
5. John Doerr — Wikipedia — https://en.wikipedia.org/wiki/John_Doerr
6. Performance indicator — Wikipedia — https://en.wikipedia.org/wiki/Performance_indicator
7. SMART criteria — Wikipedia — https://en.wikipedia.org/wiki/SMART_criteria
8. Psychological safety — Wikipedia — https://en.wikipedia.org/wiki/Psychological_safety
9. Goodhart's law — Wikipedia — https://en.wikipedia.org/wiki/Goodhart%27s_law
10. Measure What Matters — John Doerr — https://www.whatmatters.com/
11. High Output Management — Andrew S. Grove, Vintage Books — https://www.penguinrandomhouse.com/books/298206/high-output-management-by-andrew-s-grove/
12. Radical Focus — Christina Wodtke — https://cwodtke.com/books/
13. A plan for digital health and social care — NHS England / Department of Health and Social Care — https://www.gov.uk/government/publications/a-plan-for-digital-health-and-social-care
14. GOV.UK Service Manual: measuring success — Government Digital Service — https://www.gov.uk/service-manual/measuring-success
15. The Fearless Organization — Amy C. Edmondson, Wiley — https://fearlessorganization.com/