First response time (FRT) is the time between a customer asking for help and a person on your team replying. In email it is easy to compute because every message is a ticket. In Slack Connect it is not: messages arrive in shared channels, in threads and out of them, from several people at the customer, and your reply might be a quick emoji or a full answer. Before you put a number in an SLA, settle four definitions.
When does the clock start?
Start at the timestamp of the customer’s first message in a new request. Not when someone on your side noticed it, and not when a ticket was created by hand an hour later.
“New request” needs a rule. A practical one: a new top-level message in the shared channel from someone at the customer starts a new request, unless it is clearly a follow-up to a thread that is still open. Messages inside an existing thread belong to that thread.
What stops the clock?
The first reply from a person on your team that addresses the request. That excludes:
- Emoji reactions.
- Automatic acknowledgements (“Thanks, we’ve got this”).
- Replies from your side to someone else in the channel.
If you use an AI agent that answers customers directly, decide whether its replies count. If they do, report human and AI first responses separately so the number stays meaningful.
Which hours count?
If you promise “within one hour during business hours”, measure business hours. A message at 5:40 pm on Friday with a one-hour target and 9 am to 6 pm hours is due at 9:40 am on Monday, after 20 minutes on Friday and 40 on Monday. The first response time calculator works these deadlines out for any hours and days.
Write the time zone into the SLA. For global customers, either use the support team’s time zone or define coverage windows per region.
How do you report it?
Report two numbers per period and per customer tier:
- Attainment: the share of requests that got a first response within target.
- 90th percentile FRT: how long the slowest tenth of customers waited.
Averages are easy to improve while still failing customers. One reply in two minutes and one in eight hours average out to a respectable-looking number, and one of those customers is unhappy.
Making the numbers trustworthy
The hard part in Slack Connect is capturing every request. If your measurement only sees the messages someone remembered to log, your FRT will look great and be wrong. Track from the channel itself: every top-level customer message becomes a request with a start time, automatically.
Weft turns every customer message in a shared channel into a thread in one queue, and your reply posts back into Slack under your own name. SLAs and escalations are part of the Growth plan; see pricing.