Maintenance emergencies vs routine requests
Almost every maintenance call falls into one of two buckets and the cost of confusing them is asymmetric. A genuine emergency — water escaping, no heat in winter, no power, a gas smell, a door that will not lock — costs more the longer it waits, and the damage compounds. A dripping tap, a slow drain or a broken cupboard door costs the same on Monday as it does at midnight.
So the routing rule should be explicit and short, and it should err toward the small number of unambiguous emergency triggers rather than trying to grade severity. A contractor called out at 2am for a dripping tap is expensive twice: the callout, and the next time they are slower to answer.
What separates a useful call from a message is the detail collected before anyone travels — exact unit, access arrangements, whether the tenant is on site, whether the stop tap or breaker has been found. Call routing is only as good as what it has to route on.
Tenant vs owner calls
These are different jobs sharing a number. A tenant is usually reporting a problem and wants to know when somebody is coming. An owner is asking about arrears, a vacancy, a renewal or an invoice, and expects a different register entirely — they are the client, and the tenant is not.
Identifying which is calling, early, is most of the work. Get it wrong and an owner gets a maintenance script or a tenant gets asked about a portfolio they do not have. It is also where the confidentiality line sits: an agent that will discuss an owner’s arrears position with whoever calls about that unit is a problem, and caller identity should gate what can be said rather than only what gets logged.
After-hours cover across a portfolio
Portfolio scale changes the shape of this. One building generates few enough out-of-hours calls that a manager’s mobile works. Two hundred units across several sites generate a steady trickle that is too small to justify a night desk and too constant to keep landing on one person — which is how out-of-hours duty becomes the reason property managers leave.
That trickle is exactly the shape a per-minute agent handles well: idle costs nothing, and simultaneous calls during a burst-pipe night have no queue. If you want to size the staffed alternative honestly first, the Erlang C calculator gives the seat count your actual call pattern implies — for most portfolios the answer is an uncomfortable fraction of a person.
Escalation rules
Write the rule before choosing a vendor. It needs four things: the list of triggers that count as an emergency, who each one reaches, what happens when that person does not pick up, and what the tenant is told in the meantime. The fourth is the one most often left out, and it is the one the tenant judges you on.
The failure worth designing against is the silent one — an escalation that reaches a contractor’s voicemail and is recorded as delivered. Confirm a human answered before the call is treated as handled, and make the fallback explicit rather than hoping the first number always works.
