The Dashboard the Director Loves
MetroHealth's shared IT service desk has 55 agents and handles 18,000 contacts a month for 31,000 clinical and corporate users. The monthly operations report to the CIO shows 96% of calls answered within 60 seconds, average handle time of 9 minutes 20 seconds, agent utilisation of 91%, schedule adherence of 94%, 22 tickets closed per agent per day, satisfaction of 3.6 out of 5, and first contact resolution of 74%. Eleven months ago a target was set to reduce average handle time to 10 minutes from 12. After it was introduced, handle time fell to 9:20 and satisfaction fell from 3.9 to 3.6. Two measures are absent from the report: 21% of contacts are repeat contacts within seven days, against 14% a year ago, and 1,100 tickets are open, of which 340 are more than 30 days old. Agents transferred 2.5 times as many contacts to second line as a comparable desk last quarter. The CIO's instruction for next year is 'average handle time to 8:00, no additional headcount', while contact volume is forecast to grow 6% per quarter. The desk's engagement survey score fell 12 points.
1. Diagnose what the current measurement set is causing, using the evidence in the report and the evidence missing from it.
Read the worked answer
The set is internally consistent and almost entirely about activity, so it rewards closing interactions rather than meeting needs.
Look at what the handle time target did. Handle time fell from 12:00 to 9:20 while satisfaction fell from 3.9 to 3.6 and repeat contacts rose from 14% to 21%. At 18,000 contacts a month, that is roughly 1,260 additional repeat contacts a month - several hundred hours of rework created to save talk time on the first contact, with the cost moved onto the user and onto the second-line queue, where transfers are 2.5 times a comparable desk. Twenty-two tickets per agent per day and 91% utilisation push in the same direction: no time to probe, no time to write the article, no capacity for a surge.
Two measurement failures make this invisible. First contact resolution at 74% is almost certainly defined as 'not reopened', which counts a user who gives up and phones back as a success. And there is no measure of backlog health while 340 tickets sit beyond 30 days.
The mechanism is familiar: handle time stopped being a diagnostic and became the target, so it stopped being a good measure. Analysts are behaving rationally against the measures they are given.
The point for leadership is not that handle time is bad. It is that the report shows improvement while the system degrades - engagement down 12 points, repeat work up half again year on year, backlog ageing - and the desk is being asked to accelerate the same behaviour.
2. Propose a balanced measurement set tied to critical success factors, keeping the number of key performance indicators small enough to be actionable.
Read the worked answer
Anchor the set to a handful of critical success factors, then pick one or two indicators for each.
CSF 1 - Users' needs are met at the first reasonable opportunity. KPIs: repeat contact rate within seven days on the same topic; first contact resolution defined against that same test rather than against reopens. Diagnostic: reopened tickets.
CSF 2 - Help is easy to obtain through the channel of choice. KPIs: satisfaction or customer effort score by channel and by resolver group; verified self-service containment. Diagnostic: percentage of demand arriving through shadow channels.
CSF 3 - Incidents are contained and communicated. KPIs: major incident detection and communication timeliness; percentage of priority 1 incidents with a named coordinator. Diagnostic: transfers to second line with reason codes.
CSF 4 - Capability and capacity are sustainable. KPIs: attrition and engagement; percentage of agents with current knowledge accreditation. Diagnostic: utilisation (never a target) and handle time.
CSF 5 - The backlog is healthy. KPI: percentage of open tickets breaching age thresholds by priority, plus staleness. Diagnostic: 30-day-plus count.
State the principles in the report itself: balance output with outcome and experience; include at least one leading measure per outcome; keep the KPI count small enough to act on; and never set a target on a measure that is a proxy for effort rather than result. Handle time stays in the pack as a capacity-planning and coaching diagnostic.
3. How would you challenge the instruction to cut average handle time to 8:00 without appearing to resist efficiency?
Read the worked answer
Challenge the number, not the person, and arrive with a counter-proposal and its cost.
- Present the trade-off as evidence. The last handle time target delivered 2 minutes 40 seconds of savings alongside a 0.3-point satisfaction drop, a seven-point rise in repeat contacts and 2.5 times the transfers. At current volume, 21% repeat contact is roughly 3,780 contacts a month - several hundred agent-hours - which is more work than an 80-second reduction would recover.
- Quantify the 8:00 option honestly. With volume growing 6% a quarter and no hiring, the target can only be met through more transfer-and-close, more deflection of things that should not be deflected, or overtime. Put those three on one page with their consequences.
- Offer a controlled test. Two teams for one quarter with repeat contact rate and satisfaction as the target and handle time as a guardrail, against a control team on current targets. That settles the disagreement cheaply and protects the CIO's credibility, because it is an experiment rather than a refusal.
- Ask which outcome the CIO actually wants - cost per resolved contact, agent availability, clinical-user satisfaction - then show that reducing repeat contacts serves it better than cutting 80 seconds.
- Record it in the continual improvement register with an owner and a review date, and use the service owner as a sponsor rather than escalating over the CIO's head.