How should we handle calls at night and on weekends?
Stop treating after-hours as one bucket. Late calls split into true emergencies, work that is happily bookable for tomorrow, price shoppers, and existing customers checking status — and each deserves a different response. The design question is not whether to answer. It is what promise you are willing to make at eleven at night, and whether the person who has to keep that promise finds out in time.
Four call types wearing the same timestamp
Pull a week of after-hours calls and sort them by what the caller actually wanted. The mix is consistent enough to design around.
- True emergency. Water, gas, no heat in a hard freeze, no cooling in extreme heat, electrical burning smell. Needs a human now.
- Bookable tomorrow. The caller is awake, motivated, and perfectly happy with a morning slot. This is the largest group and the easiest to capture.
- Shopping. Calling three companies at nine at night. Whoever answers gets the shot.
- Existing customer. Wants a status, a reschedule, or reassurance. Cheap to satisfy, expensive to ignore.
The promise problem
Most after-hours damage is not a missed call. It is a promise made at night that nobody kept in the morning. An answering service takes a message that lands in an inbox nobody opens until nine. An automated system says a technician will call back and no ticket reaches dispatch.
So the first thing to build is not the greeting. It is the write path: the after-hours conversation has to land as a real record in the system your morning team already works from. If it lands in a separate inbox, it will be missed. That write-back is ordinary integration plumbing, and it matters more than the voice on the phone.
Deciding when to wake someone
The escalation rule is an economic judgment you should make explicitly and write down, not one the software should improvise. Waking an on-call technician has a real cost in fatigue and next-day capacity. Not waking them has a cost in lost work and in a customer situation getting worse.
Set the rule by symptom, not by the caller's own urgency rating — callers overstate and understate roughly equally. Then bias the rule toward escalation, because the error of waking someone unnecessarily is recoverable and the other error sometimes is not.
Write the rule as a short list of observable conditions and review it seasonally. What counts as an emergency in July for a cooling company is not what counts in January, and a rule written once in spring will be wrong by the first hard freeze.
What to measure the next morning
Three numbers tell you if after-hours coverage is working: how many after-hours contacts became a scheduled job, how many escalations were genuinely emergencies, and how long it took the morning team to touch each captured record. Track them in your morning brief and the gaps surface within a week.
Topics: after hours · on call · emergency · coverage
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.