The 9 Civil Service Behaviours: Complete Guide With Examples

Every Civil Service behaviour has a public definition and a private scoring reality, and the gap between them is where interviews are lost. The definitions — available on GOV.UK, restated by every careers site — describe virtues. The descriptors panels actually score against describe actions at a grade level, and each behaviour has a characteristic trap: an answer that satisfies the definition while missing the marks entirely. This guide covers all nine through that scoring lens — what each behaviour actually rewards, the trap inside it, and how the bar moves from EO to G7 — with links to full worked fail-vs-pass rewrites for every one.

Two facts frame everything below. First, you'll face three to five of these, named in your advert — never all nine. Second, every behaviour is scored 1–7 against descriptors for the advertised grade, so the calibration sections in each linked guide matter as much as the behaviour itself. If the scoring system is new to you, start with how Civil Service interviews are scored.

Seeing the Big Picture

What it rewards: awareness that changed something — a strategic driver, policy shift or wider priority that visibly altered what you or your team did. The trap: awareness held in mind. Candidates recite the departmental context they knew, and knowing isn't evidence; the mark is in the decision the knowledge produced. By grade: EO connects daily work to team objectives; HEO/SEO aligns activity to strategic drivers; G7 brings a cross-government perspective and shapes strategy upward. Full worked example: Seeing the Big Picture.

Changing and Improving

What it rewards: a problem quantified, an improvement owned through implementation, a managed risk, a resistor brought along. The trap: the suggestion trap — "I suggested it to my manager, who took it forward." The moment ownership leaves the story, so does the score, and from HEO up this is the single most common cap in the framework. By grade: EO can score on suggesting and helping others adapt; HEO must identify and implement; SEO weighs impacts across users and takes managed risk; G7 creates the conditions for others to improve. Full worked example: Changing and Improving.

Making Effective Decisions

What it rewards: a decision that was actually yours, options genuinely weighed with costs and risks, and the confidence to hold it — including when unpopular. The trap: owning the research and outsourcing the decision. "I analysed the options and presented them to my manager" is HEO evidence of an EO answer; the panel wants "I decided, because." By grade: EO decides within guidance and explains clearly; HEO/SEO makes timely, evidenced recommendations and difficult calls; G7 decides on incomplete evidence and defends it. Full worked example: Making Effective Decisions.

Leadership

What it rewards: accountability that costs something, direction others followed, inclusion as concrete acts, and — the most under-used vein — defending your own or your team's decisions under pressure. The trap: answering with management. Rotas, 1:1s and coordination are administration; none of it is a leadership mark, and no line reports are required for a 6. By grade: EO scores on personal accountability and fairness; HEO/SEO on clear objectives, credit given, conflict managed; G7 on defending decisions despite pressure and inspiring engagement. Full worked example: Leadership.

Communicating and Influencing

What it rewards: a position moved — someone who thought one thing and, because of how you made the case, did another. Channel chosen deliberately for the audience, message landed, resistance met with reasons. The trap: communication as output. Updates sent, meetings held, stakeholders "kept informed" — activity without a changed mind is unscoreable. By grade: EO communicates clearly and handles difficult conversations; HEO/SEO chooses styles for impact and communicates honestly under pressure; G7 delivers difficult messages with sensitivity and monitors whether communication actually worked. Full worked example: Communicating and Influencing.

Working Together

What it rewards: a relationship you built across a gap — friction, distance or distrust overcome — with reciprocity in it, plus the wellbeing element almost nobody claims. The trap: the harmony trap. "We all pulled together" describes weather; panels can't score an atmosphere, only attributable acts. By grade: EO builds contacts beyond the team and asks for help; HEO/SEO builds stakeholder partnerships and puts wellbeing support in place; G7 maintains a network that delivers shared objectives. Full worked example: Working Together.

Developing Self and Others

What it rewards: someone else's capability changed by you — a gap diagnosed, an intervention matched to it, independence engineered rather than dependence created. Your own development scores only when targeted and applied. The trap: the course list. An answer that is entirely your own training caps at a 3 from HEO up; the marks live in the second half of the behaviour's name. By grade: EO shares knowledge and meets development objectives; HEO/SEO coaches, mentors and builds team capability for future work; G7 builds learning culture and equal development opportunity across teams. Full worked example: Developing Self and Others.

Managing a Quality Service

What it rewards: the system fixed, not the case handled — a pattern found behind an instance, prevention built into the process, a feedback loop established, quality weighed against cost. The trap: the one-case answer. A difficult customer handled beautifully proves service skills; the descriptors score service management, and one satisfied customer leaves the service unchanged. By grade: EO identifies common problems and proposes solutions; HEO/SEO develops and reviews the systems themselves; G7 shapes whole services against best practice and regulatory requirements. Full worked example: Managing a Quality Service.

Delivering at Pace

What it rewards: a trade-off named and owned when demands competed — something cut, deferred or renegotiated deliberately, with quality protected where it mattered most. The trap: heroic hours. "We worked evenings and got it all done" reads, from EO upward, as the absence of a plan told proudly; effort is what candidates offer when they can't show prioritisation. By grade: EO owns and organises its own workload; HEO makes the prioritisation call and renegotiates constraints; SEO re-plans a team, not a person; G7 builds pace into systems — triggers, review points, resources moved between teams. Full worked example: Delivering at Pace.

The pattern across all nine

Read the nine traps together and they're one trap wearing nine costumes: evidence pitched one grade below the descriptors being scored. The suggestion trap is HEO evidence told at EO pitch; heroic hours are EO evidence offered at HEO; the one-case answer is customer service offered where service management is being scored. This is why generic "good answers" fail — the answer isn't bad, it's mis-calibrated, and the panel's scoresheet has no line for effort at the wrong level.

The fix is the same for every behaviour: read the descriptors at your advertised grade, build each example around a decision or difficulty (frictionless stories carry no evidence from HEO up), attach a because to every action, and quantify the result. Our grade guides do this calibration behaviour-by-behaviour — EO, HEO, SEO and G7 — and the grade hubs (EO, HEO, SEO, G7) show how it plays out across a whole interview.

Preparing your example bank

For a typical campaign assessing four behaviours, prepare five examples: one per advertised behaviour, each built for its descriptors, plus one flexible story that can serve two or three behaviours if a probe opens a gap. Test each example against three questions: Whose decision was it? (yours, visibly), What was traded or difficult? (something, always), What changed, measurably? (a number, ideally). An example failing any of the three needs restructuring or replacing before interview day.

Then test the delivery, because a written example and a spoken one are different animals. The AI coach runs scored mock interviews against the same grade-level descriptors panels use, and tells you which behaviour is leaking marks and why — three sessions free, with all nine behaviours, every grade guide and the full 250+ question bank on membership.