Civil Service Behaviours: 9 Examples, One Per Behaviour

HEO: Seeing the Big Picture, 246 words (246 words)

I led a caseworker team at a Jobcentre in Bradford. We processed Universal Credit work capability referrals for around 1,400 claimants a month. In March 2025 the district was asked to cut the time claimants waited between referral and assessment. My task was to work out what our team could change locally without moving the target somewhere else.

I started by reading the department's published outcome measure rather than our internal dashboard, because the two counted different things. Our dashboard stopped the clock when we sent the referral. The national measure stopped it when the claimant was assessed. That gap explained why our figures looked good while complaints rose.

I mapped where referrals failed at the assessment provider end. Of those, 31 per cent came back for missing evidence of a mental health condition. I raised this at the district operations meeting and proposed a change to our checklist, adding two evidence prompts before submission. I checked with the provider's team leader that the prompts matched what their clinicians needed.

I also flagged the same pattern to the regional lead, because three other Jobcentres used the same checklist and would see the same effect.

Returned referrals from my team fell from 31 per cent to 12 per cent over the following quarter. The district adopted the revised checklist in July 2025 across all four offices. The claimant-facing wait, measured on the national definition rather than ours, fell by an average of nine days across the district.

HEO: Changing and Improving, 242 words (242 words)

I led a team of six at the Land Registry office in Plymouth. We handled first registration applications. Our requisition rate is the proportion of applications sent back to conveyancers for missing information. In September 2025 it stood at 24 per cent against a target of 15.

My task was to reduce it without adding a checking stage, because we had no capacity for one.

I asked each caseworker to log the reason for every requisition for three weeks. We used a shared sheet rather than the case management categories. That produced 211 logged reasons. Two causes covered more than half of them: unsigned plans and missing evidence of identity for corporate applicants.

I drafted a one-page note for the four conveyancing firms that generated most of our volume, showing the two errors with an example of each. I asked our customer team to send it rather than sending it myself, so it came through the established route. Internally I changed our acknowledgement email to name the two items explicitly.

I asked the team whether the change created extra work. Two said the logging sheet was a burden, so I stopped it after week three as agreed.

By December our requisition rate was 16 per cent. The four firms accounted for most of the improvement. The other two teams adopted the acknowledgement wording in January 2026. I wrote a short note for the improvement log on what had and had not worked.

HEO: Making Effective Decisions, 243 words (243 words)

I was an operations manager at a DVLA contact centre in Swansea. Each week I decided how 48 advisers were split across telephony and written correspondence. In January 2026 a system release caused a spike in driving licence queries. Our answer rate fell from 86 per cent to 61 per cent in four days.

My task was to decide whether to move staff off correspondence, knowing that correspondence had a published 10-day service standard.

I pulled three days of half-hourly call data and the correspondence ageing report before deciding anything. The call spike was concentrated between 8am and 11am, not spread through the day. Moving people for the whole day would have broken the correspondence standard for no extra benefit.

I considered three options: a full-day move of 12 advisers, a morning-only move of 18, and overtime. I costed the overtime at roughly £4,100 a week and rejected it as the first move rather than the last.

I consulted the correspondence team leader on which case types could safely age by two days, and she identified duplicate address changes.

I moved 18 advisers to telephony for the morning only, for two weeks, and set a review point after five days.

The answer rate recovered to 83 per cent within a week. Correspondence aged by an average of 1.4 days and stayed inside the 10-day standard. At the review I returned six advisers early, because the call data showed the morning peak had flattened.

HEO: Leadership, 241 words (241 words)

I managed nine caseworkers in a Home Office visa casework unit in Sheffield. Four of them joined within the same month in 2025. Morale dropped after a restructure moved two experienced staff out, and our internal engagement score fell to 52 per cent.

My task was to rebuild a functioning team, not to explain the restructure away.

I held one-to-ones with all nine in the first fortnight. I asked one question in each: what would you change if you could. Six named the same issue, which was that allocation felt arbitrary and the hardest cases always went to the same two people. I checked the allocation data and they were right.

I published the allocation rule, rotated complex cases weekly, and named who held each category. I said openly at team meeting that the previous arrangement had been unfair, rather than defending it.

One caseworker raised that our team meetings always ran over into her carer hours, so I moved them to 11am. Two colleagues were coached through their first complex refusal decisions by me rather than by the same senior pair.

I raised the two cases of inappropriate banter I witnessed with the individuals concerned on the day.

The engagement score rose to 71 per cent at the next survey. Two of the four new starters had passed probation early by March 2026. Team throughput returned to the unit average within two months, having been 14 per cent below it.

HEO: Communicating and Influencing, 249 words (249 words)

I worked in policy support at the Environment Agency in Bristol. In October 2025 I had to explain an abstraction licensing change to 60 farmers, over two evening meetings. The change removed an exemption that several of them relied on, and the first meeting was expected to be hostile.

My task was to make the change understood and to bring back what we heard, not to sell it.

I rewrote the technical note into two sides of plain English and had it checked by a colleague who had never worked on abstraction. I cut the acronyms and kept the legal reference in a footnote. The sentence that mattered went at the top: your existing exemption ends on 31 March 2026.

At the meeting I said that sentence first and then took questions for 40 minutes before presenting anything else. I did not defend the policy, because the decision was made and pretending otherwise would have cost me the room.

I wrote up the seven distinct concerns raised, including two that had not been considered, and sent them to the policy lead within 48 hours.

I offered a follow-up session for the eleven farmers with the most affected holdings and ran it in November.

The plain English note replaced the technical version in all subsequent communications. Two of the seven concerns led to a transitional arrangement for small holdings, announced in December. Attendance at the second meeting was higher than the first, which in that room counted as a result.

HEO: Working Together, 247 words (247 words)

I was a resettlement manager at a category C prison in Nottinghamshire. My job was to see that men leaving custody had accommodation arranged before release. In 2025 around a third of releases from my wing were recorded as homeless on the day of release.

My task was to build a relationship with the two local authority housing teams. They had no duty to prioritise our cohort and no reason to take my calls.

I asked for a standing monthly meeting with both teams and went to their offices rather than asking them to come to the prison. I brought data they could use, showing release dates twelve weeks ahead rather than two. I accepted their point that our referrals often arrived with no identification documents attached, which was a fair criticism.

I agreed a shared referral form with a named contact at each authority. Before agreeing anything I asked our own offender management staff what would make it workable.

I also brought in a housing charity that already worked with one of the authorities, so the relationship did not depend only on me.

Recorded homelessness on release from my wing fell from 33 per cent to 19 per cent over the following eight months, measured on the prison's own release records. The shared form was adopted by the other two wings in February 2026. When I was on leave in the summer, the monthly meeting still ran, which was the part I was most pleased about.

HEO: Developing Self and Others, 242 words (242 words)

I led seven examiners at Companies House in Cardiff. Two of them did all of the complex company restoration work. If either was absent, those cases stopped. My task was to spread that capability without slowing the team down while people learned.

I started with a capability grid listing the six case types against the seven staff, marked by who could do each unsupervised. It showed that five of seven could do only routine incorporations. I shared the grid with the team rather than keeping it to myself, and asked each person which gap they wanted to close first.

I paired each learner with one of the two experienced examiners for four cases, with the experienced examiner checking rather than doing. I protected two hours a week for this. I told the operations manager throughput would drop for six weeks, which it did, by about 8 per cent.

I also asked for feedback on my own management in those one-to-ones and was told that my written instructions were too long. I shortened them to bullet points, which I have kept.

Two people also took the department's internal legislation course. I had to argue for both places in the training budget round in October.

By March 2026, five of seven could handle restorations unsupervised. Throughput returned to its previous level in week eight and was 6 per cent above it by the quarter end. Neither original examiner has had to cancel leave since.

HEO: Managing a Quality Service, 240 words (240 words)

I ran an inbound line at an HMRC debt management office in Newcastle. It handled time to pay arrangements, around 9,000 calls a month across 32 advisers. Our quality assurance scores were good, but complaints about repeat contact were rising.

My task was to find out why customers were calling back, and to fix the cause rather than the score.

I listened to 40 calls that had been followed by a second call within seven days. In 27 of them the customer had not understood what would happen to their direct debit in the first month. Our script said the arrangement would start, but not when money would leave the account.

I rewrote that section of the script with the date made explicit and tested it with six advisers for a fortnight before rolling it out. I asked the two advisers with the highest repeat-contact rates to help write it, rather than treating it as a performance issue.

I also set up a monthly sample of ten calls, reviewed with a vulnerable customer specialist. Three of the 40 involved customers who should have been referred to the extra support team and were not.

Repeat contact within seven days fell from 14 per cent to 8 per cent over three months. Two customers were referred to the extra support team in the first monthly sample. The revised script section was adopted by the other two debt management sites in April 2026.

HEO: Delivering at Pace, 244 words (244 words)

I worked in a Department for Transport regional office in Leeds. In early 2026 I had six weeks to produce the evidence pack for a ministerial decision on a road scheme. The deadline was fixed by the announcement date and could not move. Two of my four analysts were part time.

My task was to deliver a defensible pack on time, with the gaps stated rather than hidden.

In week one I listed everything the pack could contain and cut it to the six items the decision actually required. I confirmed that list with the policy lead in writing, so scope could not drift later.

I set three milestones with dates and a short Tuesday check-in of fifteen minutes. When the traffic modelling slipped in week three, I reassigned the appraisal write-up to myself rather than letting one analyst carry two strands.

I told the policy lead at the point the model slipped, not at the deadline. We agreed to present the modelling with a stated confidence range instead of waiting for a rerun.

I also protected one analyst's non-working days, because working them would have cost us more in the final fortnight than it saved.

The pack went to the minister on the agreed date with all six items present and the modelling caveat stated on the front page. The decision was announced on schedule. The Tuesday check-in and the six-item scope note were reused by the team for the next two submissions.

EO: Making Effective Decisions, 244 words (244 words)

I worked as an examiner at HM Passport Office in Durham. I decided whether first-time adult applications could be granted on the documents supplied.

In October 2025 an application reached me with a birth certificate in one name and a driving licence in another. The customer was travelling in eleven days and had already paid for the fast track service. My task was to decide whether to grant, refuse or request further evidence, and to do it that day.

I checked the guidance on change of name before doing anything else. It allowed a deed poll or a marriage certificate as evidence, and neither was on file. I looked at the rest of the application and saw a previous passport in the second name, which the system confirmed.

I was not sure whether the earlier passport was enough on its own, so I asked my team leader rather than guessing. She confirmed it was not, because the name change itself still had to be evidenced.

I phoned the customer the same afternoon, explained in plain terms what was missing and why, and told her exactly which document would resolve it. I noted the call and the guidance paragraph I had relied on.

She emailed the deed poll that evening. I granted the application the next morning, five days inside her travel date. My team leader used the case in the following week's team meeting as an example of when to ask rather than assume.