Here is a scene that plays out in support teams more often than anyone admits.
The monthly review opens with the dashboard. First response SLA: 98%. Average handle time: down. Tickets closed per agent: up. Every tile is green. Then someone from account management mentions that a key customer has just asked about their contract exit clause.
Nobody broke a rule. Nobody lied. The dashboard was accurate. It was simply measuring the wrong thing.
That gap has a name: a watermelon metric. Green on the outside, red on the inside.
Why support KPIs turn into watermelons
Support teams measure what is easy to count. Replies are easy to count. Handle time is easy to count. Closed tickets are easy to count. Whether the customer’s problem actually went away is harder, so it often goes unmeasured.
The trouble starts when an easy-to-count number becomes a target. People are smart. If the team is judged on first response time, the team gets very good at first responses. A quick “Thanks, we’re looking into this” stops the SLA clock just as well as a real answer does.
Helpdesk mechanics make this easier than most managers realise. In many tools, Zendesk included, some SLA targets pause while a ticket sits in a Pending or On-hold status. Move the ticket, pause the clock. The report stays green while the customer keeps waiting.
None of this needs bad intent. It is the natural result of rewarding activity instead of outcome.
The fix: pair every speed metric with an outcome metric
You don’t need to throw away your speed KPIs. Speed matters. Customers do care how fast you reply.
What you need is a partner for each one: a second number that shows whether the speed actually helped the customer. Report them side by side, never alone.
| If you track this (activity) | Watch out for | Pair it with (outcome) |
|---|---|---|
| First response time | Holding replies sent just to stop the clock | Full resolution time |
| Average handle time | Rushed calls, early transfers | Repeat contacts within 7 days |
| Tickets closed per agent | Tickets closed before the fix is confirmed | Reopen rate |
| SLA met % | Clock paused through Pending or On-hold | Total ticket age, including paused time |
| CSAT score | Surveys only reaching happy customers | Survey response rate, plus complaint volume |
The pairing works because one number is hard to game without the other moving. If agents rush calls to cut handle time, repeat contacts rise. If tickets get closed early, the reopen rate climbs. The pair keeps both honest.
A 30-minute check you can run this week
You can test your own dashboard for watermelons without new tools. Pull these four numbers for the last 30 days:
- Median first reply time vs median full resolution time. If the first is minutes and the second is days, your SLA is measuring the greeting, not the fix.
- Reopen rate. Above roughly one ticket in ten deserves a closer look at how “solved” is being decided. (Treat that as a rule of thumb, not an industry benchmark. Your own trend over time matters more than any outside number.)
- Time spent in Pending or On-hold. If a large share of ticket life sits in paused statuses, your SLA report is flattering you.
- The ten lowest CSAT comments. Read them in full. Then look at the dashboard again. If the comments describe problems the dashboard can’t see, you have found your watermelon.
Example: a team with 98% first response SLA, a 2-day median resolution time and 14% reopens doesn’t have a speed problem. It has a resolution problem that the headline KPI is hiding.
Where this approach can go wrong
Two cautions, because pairing metrics is not free.
First, more numbers can mean less focus. If every KPI gets a partner and every partner gets a sub-metric, the review turns into a spreadsheet nobody reads. Pick your three or four most important KPIs and pair those. Leave the rest as background.
Second, outcome metrics move slower. Resolution time and reopen rate lag behind daily activity, so they are poor tools for managing a single shift. Keep the speed metrics for daily operations and use the outcome partners for weekly and monthly reviews.
The takeaway
If a KPI can be gamed, it eventually will be, usually by good people trying to hit the number they were given. The answer is not stricter rules. It is a second number that shows whether the first one is telling the truth.
This week: pick your single most-watched support KPI, write down its outcome partner from the table above, and put both on the same slide at your next review. One change, one meeting, and you’ll know whether your dashboard is green or just painted green.
I also share a short daily lesson on KPIs and processes on Instagram at @vijitlabs. Comment GUIDE on any reel for the free 2-page guide that goes with it.
Not sure what your Zendesk is hiding?
Book a free 30-minute Zendesk review. We look at your SLAs, ticket flow and reporting together, and you leave with at least one thing worth fixing.
Book a free review