A Nine-Point Engagement Drop in a 24/7 Contact Centre
A customer support operation employs 320 agents across four sites and runs 24/7. The annual engagement survey has just closed. Overall favourability fell to 54%, down 9 points on the previous year. "I would recommend this as a place to work" scored 38%; "my manager cares about my wellbeing" 47%; "I have the tools I need to do my job" 71%. Voluntary attrition is running at 38% annualised with average tenure of 11 months, and unplanned absence is 6.8%. Three changes landed in the same period: a new ticketing system introduced with a two-hour webinar and no follow-up, mandatory overtime averaging six hours per agent per week, and a quarterly forced ranking that places the bottom 10% on performance plans. Site directors report spending around 40% of their time on hiring. At the leadership review the proposed response is a monthly "fun Friday", an employee-of-the-month award and a refurbished break room.
1. Assess the leadership's proposed response. What will it achieve and what will it cost?
Read the worked answer
These are mood interventions aimed at a problem the survey describes as structural. The lowest scores are manager care (47%) and intent to stay (38%); the highest is tools (71%). Nothing in that pattern points to the break room.
Awards can backfire. In a site that ranks the bottom 10% into performance plans each quarter, employee-of-the-month signals that recognition is scarce and political. It reinforces the ranking logic agents already distrust and gives managers another discretionary judgement to defend.
Fun Friday ignores the schedule. A 24/7 operation has agents on nights and weekends; a Friday event either excludes them or becomes unpaid extra time. Perceived inequity across shifts is the predictable outcome, and the night shift is usually where attrition is worst.
Break room is fine but is not a lever. Fund it if the facility is genuinely poor, but do not book it as the engagement plan.
The real cost is credibility. Presenting these as the response to a 9-point decline tells staff their feedback was not read, which is worse than silence. The plausible drivers - six hours of mandatory overtime, an untrained system cutover, a forced ranking - are policy decisions leadership controls. Change one of them visibly within 30 days; that buys more goodwill than a year of events.
2. Design a 90-day plan. State what you stop, keep and start, and why the sequence matters.
Read the worked answer
Days 1–30 - stop the bleeding, test the diagnosis. Suspend the forced ranking for one cycle and announce it; the announcement is itself the intervention. Cap mandatory overtime at three hours per week and publish an overtime report by team so the cap is visible. Run eight focus groups - two per site, mixed shifts, facilitated by someone outside the reporting line - to confirm which drivers actually matter. Do not run a survey about the survey.
Days 31–60 - fix the two biggest drivers. Ticketing: 90-minute role-based training plus a floor coach for four weeks, measuring handle time and error rate, not just satisfaction. Staffing: model attrition backfill to reach zero mandatory overtime at target volume and secure the headcount approval. Manager capability: put 40 supervisors through a two-day programme on coaching, absence conversations and workload triage, since manager care is the lowest item.
Days 61–90 - rebuild the feedback loop. Monthly skip-levels, a ten-question pulse, and a visible "you said, we did" board refreshed fortnightly.
Keep: the recognition budget - redirected from individual awards to team-level, peer-nominated recognition that explicitly includes nights and weekends.
Sequence matters because credibility is the scarce resource: the first 30 days are for proving that feedback changes decisions, not for launching programmes.
3. Which metrics tell you within 90 days whether this is working, and what would make you change course?
Read the worked answer
Use leading indicators, because attrition takes a year to move and leadership will not wait that long.
Weekly: overtime hours per agent, schedule adherence, unplanned absence (baseline 6.8%), and 30/90-day voluntary attrition by team and by supervisor - supervisor-level splits are the most actionable cut.
Monthly: a three-item pulse only - manager care, workload manageable, would recommend - with a 60% response floor, or the numbers are not interpretable. Internal transfer applications are a useful leading signal: rising applications mean people want to stay but not in that job.
Qualitative: focus-group themes and exit-interview reasons coded against a fixed list so they can be trended.
Thresholds agreed in advance. If 90-day attrition has not fallen by a third within two quarters, the diagnosis was wrong rather than the execution - return to the focus-group data. If absence falls but the recommend score does not, the problem is managerial, not workload, and the coaching effort moves up from supervisors to site directors. If overtime drops but handle time collapses, the headcount model was too thin and the fix is staffing, not engagement.