How to Build a Support Team Scorecard: The Metrics That Matter

Here is how I build a support team scorecard, metric by metric. A support team scorecard is a one-page set of metrics, each with a formula, a target and an owner, that the team reviews on the same day every week. Mine has 8 metrics: incoming volume, backlog age, first response SLA attainment, median time to resolution, first contact resolution, escalation rate, CSAT and QA score. Together they answer one question: is the team keeping up, and is quality holding?

Most of these numbers have a free calculator, starting with the first response time and SLA calculator. In Part 6 we built a ticket deflection strategy. This part brings the earlier plays together in one weekly scorecard, so you can see whether any of them are working.

Which metrics belong on a support team scorecard?

The scorecard should cover four things: demand, speed, quality and customer outcome. Two metrics per area is enough.

#	Metric	Area	Owner
1	Incoming volume	Demand	Support ops
2	Backlog age	Demand	Team leads
3	First response SLA attainment	Speed	Team leads
4	Median time to resolution	Speed	Support manager
5	First contact resolution	Quality	Support manager
6	Escalation rate	Quality	Tier 2 lead
7	CSAT	Outcome	Support manager
8	QA score	Outcome	QA lead

How do you calculate each scorecard metric?

Incoming volume     = Tickets created in the week (also per 100 active accounts)
Backlog age (%)     = Open tickets older than 72 hours ÷ All open tickets × 100
SLA attainment (%)  = Tickets answered within SLA ÷ Tickets due in the week × 100
Median resolution   = Middle value of hours from creation to solved
FCR (%)             = Resolved on first contact ÷ Total resolved × 100
Escalation rate (%) = Escalated tickets ÷ Tickets handled × 100
CSAT (%)            = Satisfied responses (4 or 5 of 5) ÷ Total responses × 100
QA score (%)        = Points earned ÷ Points possible × 100

A few choices matter more than the formulas:

• Use the median for time metrics. A few tickets open for weeks will distort an average. I explain why in how I calculate first response time.
• Measure SLA against tickets that were due in the period, not tickets created. Write the SLA definition down, as you would in any service level agreement.
• Keep FCR on a fixed reopen window, as in how I calculate first contact resolution rate, and the QA score on the rubric from your support QA program.

How do you read one week of the scorecard?

Say a team of 13 agents received 2,080 tickets this week, against a four-week average of 1,850.

Metric	This week	Target	Status
Volume vs average	+12%	within ±10%	Red
SLA attainment	1,892 ÷ 2,080 = 91%	95%	Amber
Backlog age	97 ÷ 540 = 18%	10% max	Red
Median resolution	19 hours	16 hours	Red
FCR	71%	72%	Green
Escalation rate	11%	11%	Green
CSAT	361 ÷ 410 = 88%	90%	Green
QA score	84%	85%	Green

This week

Volume jumped, speed slipped, quality held

Clarivoxx analysis

Incoming volume vs average+12%

First response SLA attainment91%

Open tickets older than 72h18%

CSAT88%

The formula

SLA attainment = tickets answered within SLA ÷ tickets due in the period × 100. Here 1,892 ÷ 2,080 = 91%, against a 95% target. Volume change = (2,080 − 1,850) ÷ 1,850 × 100 = 12%.

Read it as a story. Volume jumped, speed slipped and quality held: a demand problem. Find what drove the extra 230 tickets, such as a release or a billing run, instead of asking agents to work faster. To check the SLA line on your own data, use the SLA compliance calculator.

Then check the trend. One red week can be noise. Backlog age rising six weeks in a row is a pattern.

Backlog trend

Aged backlog has doubled in six weeks

Clarivoxx analysis

0%9%18%

Week 1
9%

Week 2
10%

Week 3
8%

Week 4
11%

Week 5
14%

Week 6
18%

+9 points in six weeks, against a 10% ceilingBacklog age = open tickets older than 72 hours ÷ all open tickets × 100

Three rising weeks in a row is a pattern worth an owner, even before the metric turns red.

Which scorecard metrics should be read in pairs?

Each metric can be improved at the cost of another, so I never read one alone:

• Resolution time with CSAT: faster closures with lower satisfaction are not progress.
• FCR with reopen rate: both rising means tickets are closed too early.
• Escalation rate with resolution time: fewer escalations with slower resolution means tickets are held too long.
• Volume with deflection: if deflection works, tickets per 100 accounts should fall even as total volume grows.

The scorecard as a weekly dashboard

This is how the week above looks as the table we open at the start of the review.

Metric	Target	This week	Last week	Change
Incoming volume	1,850 avg	2,080	1,870	+210
First response SLA	95%	91%	94%	−3 pts
Tickets older than 72h	10% max	18%	14%	+4 pts
Median resolution	16 hours	19 hours	17 hours	+2 hours
First contact resolution	72%	71%	71%	0 pts
CSAT	90%	88%	89%	−1 pt
QA score	85	84	84	0

Read it top to bottom: volume up, then speed down, while FCR, CSAT and QA hold close to target. Escalation rate, flat at 11%, stays on the full page but off this view to keep it short.

How do you run a 30-minute weekly review?

Same morning each week, with the support manager, team leads and an operations owner.

1. 5 minutes: volume and what drove it.
2. 10 minutes: red and amber metrics, and the likely cause of each.
3. 10 minutes: one owner and one due date per red metric.
4. 5 minutes: last week's actions.

Agree the red and amber lines in advance, so the meeting is about causes, not about whether 91% is bad.

Status vs target

How close each metric is to target

Clarivoxx analysis

CSAT (88% vs 90%)

On track

SLA attainment (91% vs 95%)

Watch

Time to resolution (19h vs 16h)

Off track

Aged backlog (18% vs 10% ceiling)

Off track

What I've seen in practice

On one team, we built a dashboard with more than thirty charts, and after two months nobody opened it. When we cut it down to eight numbers on one page and reviewed them together every Monday, people started asking questions about the data again.

The most useful item in the meeting turned out to be the last one, last week's actions. Without it, the same red metric came back every week with a new excuse. I also keep individual agent results out of this meeting and coach people in one-to-ones instead. Named agents on a shared screen turn a review into a defence.

What should stay off the weekly scorecard?

Review these monthly instead:

• Tickets per agent and utilization. Staffing does not change week to week. The Erlang calculator fits a monthly capacity check.
• Cost per ticket, which needs a full month of costs.
• NPS, which moves slowly and reflects the whole product.

What to do this week

Write your eight metrics on one page with a formula, target and owner for each. Fill in last week's numbers by hand if needed, and hold the first review.

Frequently asked questions

How many metrics should a support scorecard have?

Six to ten. Fewer misses a whole area, such as quality. More and nobody reads the page.

Should the scorecard be weekly or monthly?

Weekly for demand, speed and quality, because they move fast. Monthly for cost, staffing and NPS.

How do I set the first targets?

Use your last eight to twelve weeks as a baseline and set each target a little better than the median week. Outside benchmarks are a sanity check, not a target. For FCR, the SQM Group call center average of about 71% is one such check.

Should individual agents see the team scorecard?

Yes, the team view. Keep individual results for one-to-ones, where they can be discussed with context.

Sources

• What is an SLA? Service level agreements explained, Atlassian
• First Call Resolution Benchmark, SQM Group