Helpdesk metrics worth tracking

Helpdesk reporting tends to fill up with numbers that are easy to produce and hard to act on. Tickets raised, tickets closed, average handling time. They describe activity accurately and tell you almost nothing about whether the service is any good.

A small number of measures actually predict whether users are satisfied. Most reports do not include them.

The four that predict satisfaction

Time to first meaningful contact

Not an automated acknowledgement. The point at which a person engages with the problem: asks a question, starts work, or explains what happens next.

This matters more than resolution time for user perception. People tolerate a problem taking time to fix. What they do not tolerate is silence, because silence is indistinguishable from being ignored.

Measure it as a percentile, and measure it from when the user raised the ticket rather than from when it reached the right queue. Routing delay is invisible to the user and entirely visible in their experience.

Reopen rate

The proportion of tickets closed and subsequently reopened, or raised again by the same user for the same issue within a short window.

This is the single best indicator of whether resolutions are real. A team under pressure on closure targets will close tickets optimistically, and the reopen rate is where that shows.

It is also difficult to game. Pushing closures up tends to push reopens up with them.

Age of the oldest open ticket

Not the average age. The oldest.

Every helpdesk has a backlog of difficult tickets that nobody wants: the ones requiring a vendor, a decision, or a change nobody will approve. They sit, and the user attached to each one becomes progressively more dissatisfied.

Averages hide these completely. A single number (the age of the oldest open ticket, and the count older than an agreed threshold) makes them impossible to ignore.

Ticket volume per user, by category

Not total volume, which reflects headcount and reporting culture more than anything else. Volume per user, broken down by category.

A category with high volume per user is a systemic problem being handled as a series of individual incidents. Password resets, a particular application, one office's network. Each ticket gets resolved and the cause persists.

This is the measure that turns a helpdesk from reactive to preventive, because it identifies what to fix rather than what to handle faster.

The ones that mislead

Average resolution time. Averages hide the tail, and the tail is what damages the relationship. Use percentiles.

First-contact resolution, unqualified. Genuinely useful when it reflects capable first-line staff and good documentation. Easily inflated by closing tickets prematurely or reclassifying anything complex as a new request. Only meaningful alongside the reopen rate.

Tickets closed per agent. Rewards volume over quality, and encourages cherry-picking the easy ones. The difficult tickets, which are the ones that matter, get avoided.

Total ticket volume. Falling volume can mean a more stable estate or a helpdesk people have given up on. It does not distinguish, and the second is common.

SLA compliance percentage as a headline. Aggregated across all priorities and categories, it can sit comfortably in the nineties while the specific service one department depends on is consistently poor.

Satisfaction surveys, used carefully

A short survey on ticket closure is the only measure that reflects experience rather than process. It is worth having, with caveats.

Response rates are low and skewed towards strong opinions. It measures the interaction, not the outcome, so a polite engineer who did not fix the problem can score well. And it responds to things outside the team's control, such as a vendor delay.

Use it as a trend rather than an absolute, and read the comments. The free-text field usually contains more useful information than the score.

A persistent gap between good operational metrics and poor satisfaction is the most important signal in the whole report, and it should trigger investigation rather than explanation.

Reading them together

The measures are most useful in combination.

• High first-contact resolution and high reopen rate means tickets are being closed, not fixed.
• Good average resolution time and a growing oldest-ticket age means the easy work is being done and the difficult work is accumulating.
• Falling volume and falling satisfaction means people have stopped reporting problems.
• High volume per user in one category means a systemic issue that faster handling will never resolve.

What to actually report

A useful monthly report is short:

• Ninetieth percentile time to first meaningful contact, by priority.
• Ninetieth percentile time to resolution, by priority.
• Reopen rate, trending.
• Oldest open ticket, and the count over the agreed threshold.
• Top three categories by volume per user, with what is being done about each.
• Satisfaction trend, with a note on any significant comments.

Six items, each of which someone can act on. The value of a report is not its completeness; it is whether reading it changes what anyone does next week.