The Response Time You Can Promise

Most SLA targets are chosen by someone picking a round number. This works backwards from your actual capacity.

What first response time can you actually promise?

Utilisation
97%
Peak hour headroom
-37 tickets/hr
Realistic first response promise
24 hours

Queues do not degrade smoothly. Past roughly 80% utilisation, waiting times climb sharply, which is why a target set on average load breaks every time volume spikes.

Stop paying for duplicates

Why targets break at peak, not on average

Queues do not degrade smoothly. At 60% utilisation, tickets are answered almost immediately. At 85%, waiting times climb steeply. At 95%, a queue that looked fine yesterday has a backlog this morning.

That non-linearity is why a target set on your average week fails every time volume spikes, and why the honest question is not "what is our average capacity" but "what can we still hit on the busiest hour of the busiest day".

Filling it in honestly

Productive hours per agent per day. Six is realistic. Eight is the number people put in and then wonder why the model was wrong.
Handle time. Use your median, not your mean. A handful of very long tickets distorts the average badly.
Peak multiplier. Take the busiest hour against the average hour from your reporting. Most queues sit between 2x and 3x, and consumer support runs higher.

What to do when the number is worse than your promise

Three levers, in order of how quickly they work.

Reduce volume. Removing duplicates cuts real volume by roughly a tenth with no change to service, and it works this week rather than next quarter.
Reduce handle time. Better macros, better internal documentation, order and account data in the ticket rather than in another tab.
Add capacity. The slowest and most expensive lever, and the one most teams reach for first.

Also worth remembering: a slow first response manufactures its own extra volume, because people who wait contact you again on another channel. That second contact is a duplicate, and it makes the queue you are already behind on longer.

Frequently asked questions

What is a good first response time?+

Under an hour is strong for business-hours support, under four hours is acceptable in most B2B contexts. Consistency matters more than the number: an occasional eight-hour reply damages trust more than a steady four.

Does an auto-acknowledgement count?+

No. Counting it makes your dashboard disagree with your customers, which is the worst kind of metric.

Why does my utilisation look fine but replies are slow?+

Because averages hide peaks. Check the busiest hour rather than the daily total.

The fastest lever on response time

Cutting duplicate tickets reduces volume immediately, with no new headcount and no change to what customers experience.

Start free trial

14-day free trial. No credit card required.