Company: Tyco / JCI Period: 2013–2016 Role: Director, Global Technical Support Pillar: Process Improvement

Building a Support Scorecard
That Actually Changes Behaviour

Most support scorecards measure activity. This one was designed to change it — with weighted components, industry benchmarks, and a scoring scale specific enough that agents knew exactly where they stood and exactly what to do about it.

4Weighted components
35%Weight on quality
5-pointScoring scale
90%FCR target within 24h
Global4 regional centres

Why Most Scorecards Fail

The typical support team scorecard tracks calls handled, average handle time, and CSAT. It reports these numbers monthly. Agents see the report, note where they rank, and then continue doing what they were already doing — because the scorecard doesn’t tell them specifically what to change or give them a precise enough target to aim at.

The Tyco/JCI scorecard was designed differently. The goal wasn’t to measure performance — it was to create a shared language between agents and managers that made performance conversations precise, evidence-based, and actionable. An agent scoring a 2 on First Case Resolution knew exactly what that meant: they were below the 88% threshold. They knew what a 3 was (88–90%) and what it would take to get there. The path from a 2 to a 3 was a defined improvement, not a vague aspiration.

“A scorecard that tells you where you are without telling you how to get somewhere else isn’t a management tool — it’s a report. The design question is whether the person reading it knows exactly what to do next.”

Four Components, Deliberately Weighted

The scorecard weighted four distinct areas of performance. The weighting itself was a deliberate statement about what the organisation valued. At 35% — the single largest component — quality outweighed quantity. This was not obvious to agents who had come from environments where calls-per-hour was the primary metric.

10%
Administrative Results
Attendance, break exceptions, and lunch exceptions. Zero lates scored a 4; any late scored a 0. These weren’t punitive metrics — they were fairness metrics. An operation where some agents consistently exceeded break limits while others adhered creates resentment that undermines team cohesion.
20%
Quantity
Calls answered (within ±3 of team average), average call time (within ±20% of team average), and average after-call work time. Critically, the scorecard flagged that exceeding call volume targets carried a quality risk — an explicit design choice to prevent gaming.
35%
Technical Support Quality
First Case Resolution within 24 and 48 hours (target: 90% within 24h), call quality shadowing score, email quality score, and Salesforce ticket quality. FCR used a 5-point scale: 99%+ scored a 5, 90–98% scored a 4, dropping to 0 below 85%.
35%
Surveys & Customer Feedback
Customer satisfaction surveys, weighted equally to quality. An agent could have technically correct ticket handling but poor customer experience, or warm interactions but substandard technical resolution. Both matter, and the equal weighting made that explicit.

Specificity Over Labels

Each metric used a 0–4 or 0–5 point scale with specific thresholds for each score, not vague labels like “exceeds / meets / below expectations.” This specificity was the difference between a scorecard that changed behaviour and one that just generated arguments.

For First Case Resolution: 99%+ = 5, 90–98% = 4, 88–90% = 3, 86–88% = 2, 85–86% = 1, below 85% = 0. Every agent could calculate their own score before the manager ran the numbers. That self-scoring was not accidental — managers who present scores as conclusions tend to get defensiveness. Agents who arrive at their own score through data they can verify own it differently.

Giving Context Beyond the Team

The scorecard included an industry comparison column for each metric. Where benchmark data was available, agents could see not just how they performed against team averages but how those team averages compared to industry standards.

This served two purposes: it gave context to team performance conversations, and it gave individual agents a reference point beyond their immediate team. Agents who scored above team average but below industry benchmark had a different development conversation than agents who scored below team average. The benchmark column made that distinction visible without the manager having to manufacture it.

What Made This Scorecard Different

Quality outweighed quantity by design. Putting quality and surveys at 70% combined was a deliberate signal that resolution and customer experience were the priority, not throughput. That signal had to be backed up in calibration conversations.
The internal check between quantity and quality prevented gaming. An agent who inflated their call-answered count by cutting calls short would see the quality score fall. The best way to score well overall was to actually be good at the job.
Specificity in thresholds eliminates management subjectivity. “You need to improve your FCR” is a different conversation from “you’re at 87% and the 3 starts at 88% — here’s what we’re going to do in your next 10 coaching sessions.”
Self-scoring changes the conversation. When agents could calculate their own scorecard before the 1-on-1, the conversation started from shared data rather than a defence of the manager’s assessment. This reduced friction significantly.
Performance ManagementScorecard DesignFCR Quality FrameworkCSATCoachingContact Centre Ops