How to Measure Hiver SLA Performance Without Guessing

November 2, 2026

Hiver reports SLA breaches from the status field and the assignment record, so the numbers are only as honest as your team’s labelling. Three monthly metrics — first response median, breach count, and reopened rate — plus two audit rules will tell you whether support is improving or just closing tickets faster.

Bad SLA reporting is worse than none. It creates arguments about the number instead of the work, and teaches the team to game the field rather than serve the customer.

How does Hiver calculate SLA breaches?

Each SLA has a target (say, 4 hours to first response). The clock starts when the email arrives in the mailbox and stops on the first outbound reply from an agent. If it stops after the target, the conversation is marked breached.

Three things break this in practice: replies from a personal Gmail not connected to Hiver, templates sent but never marked as reply, and conversations reassigned after a reply. Only the third is detectable in reporting, and only if you look.

Step 1: define the status discipline in writing

SLA accuracy starts with three plain rules on the wall:

  1. A conversation moves to Pending the moment you ask the customer for something.
  2. A conversation moves to Closed only when the customer is done, not when you are.
  3. Anything waiting on another internal team leaves a note naming who, and stays Open.

Rule 2 is the one teams resist, and the one that makes reopened rate meaningful.

Step 2: measure first response as a median, not a mean

One 3-day-old ticket can drag a mean first response from 2 hours to 14. The median tells you what a typical customer experiences; the 90th percentile tells you how bad the tail is.

Hiver’s SLA report can export to CSV. Do the maths yourself for the first quarter — median and p90 per weekday, because Saturday volume is usually low and fast while Monday is high and slow, and a combined average hides both.

Step 3: track reappearance and reopening, not reply count

Reply count is a productivity theatre metric. A ticket that takes six replies usually means the first answer didn’t land.

Track instead:

Metric Definition Healthy range
First response median arrival → first agent reply under SLA target × 0.5
p90 first response 90th percentile of the same under SLA target
Reopened rate closed then reopened within 7 days under 8%
Collision incidents double replies to one customer near zero

Collision incidents should be near zero given Hiver’s collision detection. If they aren’t, agents are working with notifications off or from the mobile app in offline mode.

Step 4: audit 20 conversations a month by hand

Reports can’t tell you whether the status field is honest. Twenty sampled conversations a month can.

Pick 20 closed threads across the team. For each, check: did the customer reply after the close, did the agent close without confirming, and was the thread pending while someone was actively working it? Score each thread pass or fail. A failure rate above 20% means your SLA numbers are fiction and the fix is coaching, not tooling.

What does a good month look like?

For a 5-agent B2B support team handling around 900 conversations a month:

A good month: first response median under 90 minutes, p90 under 5 hours, reopened rate 6%, and an audit pass rate of 85%+. A bad month isn’t one with slow responses — it’s one with a 4% reopened rate and an audit pass rate of 40%. The second is a team closing tickets rather than solving them, and the dashboard looks great while CSAT slides.

How does this change when you add agents?

Above six agents, weekly review beats monthly. Status drift sets in fast once nobody can see everyone else’s inbox in a glance, and a month of bad labelling is a month of unusable data.

Additionally, split SLAs by view rather than by mailbox. A billing question and a technical bug will not have the same target, and forcing both into a 4-hour first response creates pressure to send something, anything, within four hours. That “something” is where reopened rates come from.

When Hiver reporting is the wrong tool

If you need to cross-reference SLA data with CSAT, product usage, or account value, Hiver’s in-product reports won’t get you there — export to a spreadsheet or a BI tool. Teams at 1,500+ conversations a month usually outgrow the built-in reporting within two quarters and start piping the CSV export somewhere else.

That’s not a flaw exactly. Hiver is a shared-inbox tool with reporting attached, not a reporting platform with an inbox attached. Pick it knowing which half you need.

Quick Summary

  • SLA accuracy depends on status discipline — Pending when waiting on the customer, Closed only when they’re done.
  • Use median and p90 first response, never the mean; one old ticket distorts everything.
  • Track reopened rate and collision incidents, not reply counts.
  • Audit 20 closed conversations monthly; under 80% pass rate means your numbers are wrong.
  • Weekly reviews above six agents, and split SLA targets by ticket type.
Tags: