Setting a callback promise for emergency calls

Why emergency calls need a promise

When a caller has no heat, water through the ceiling or a garage door stuck open, the worst part is not knowing when anyone will call back. A clear callback promise, such as “a technician will call you within 15 minutes,” turns that uncertainty into a wait the caller can plan around. It also gives your team a target and gives the owner a number to check.

Without a promise, callbacks drift. Research on web leads found that among companies that responded within 30 days, the average response time was 42 hours^[1]. Emergency callbacks are usually faster than that, but the pattern is the same: without a defined time and someone measuring it, the slow callbacks get slower. And slow callbacks lose people. Contacting a lead within an hour made firms nearly seven times as likely to qualify it as waiting even one hour longer, and over 60 times as likely as waiting 24 hours or more^[2]. A homeowner with water on the floor is a far more impatient lead than a web form.

Choosing 10, 15 or 30 minutes

Choose the shortest promise you can keep on nearly every call, not the one that sounds best. A 10-minute promise kept half the time does more damage than a 30-minute promise kept every time. Many shops start at 15 minutes for the primary callback, with a 30-minute outer limit by which the caller must have spoken to someone qualified.

Demand makes the promise matter more. On average 53% of surveyed US homeowners reported difficulty finding pros in 2025, rising to 70% among millennials^[3]. A homeowner who has been let down before will call two or three companies at once. The first one to call back with a clear plan usually gets the job.

Define what the clock measures

• Start: when the caller first reaches your company, whether a person, answering service or AI system takes the call. Not when the page reaches the tech.
• Stop: when a qualified person, usually the on-call tech, speaks with the caller. A text saying “we'll call soon” does not stop the clock.
• Qualified call types: only calls that meet your emergency definition. Routine calls follow the morning callback promise instead.
• Unreachable callers: if the tech calls and texts and gets no answer, the attempt time counts as a stop, provided a voicemail and a text were left.

Write the definitions down. Disagreements about whether a promise was kept are almost always disagreements about when the clock started.

The escalation ladder

1. 01

   Minute 0: page the primary

   The answerer sends the call details to the primary on-call tech by call and text at the same time, with the caller's name, number, address, problem and any safety notes.

2. 02

   Minute 10: second attempt

   If the primary has not acknowledged, the answerer pages again and calls the tech's phone directly.

3. 03

   Minute 15: go to the backup

   No acknowledgment means the call goes to the backup tech automatically. The primary is not waited for.

4. 04

   Minute 25: go to the manager

   If neither tech has called the customer, the manager on call is paged and calls the customer personally to explain and set a new time.

5. 05

   Update the caller

   If the promise will be missed, the answerer or manager calls or texts the customer before the deadline, not after, with the new expected time.

What to tell the caller

The promise should be stated the same way on every emergency call, followed by what to do while waiting. For example: “I've sent this to our on-call technician. He or she will call you from our office number within 15 minutes. If you smell gas or a CO alarm goes off before then, leave the house and call 911. If you don't hear from us in 15 minutes, call this number again and we'll escalate.”

Mention the number they will see. Eight in ten Americans say they don't generally answer their cellphone when an unknown number calls^[4], and a promised callback that goes unanswered because it looked like spam wastes the whole effort. Tell callers which number to expect, and have techs send a text before dialing.

Track it every week

Pull a weekly list of every call that met the emergency definition and record four numbers for each one: first contact time, first callback time, who called back, and whether a truck rolled. Then report three things.

1. The share of emergency calls called back within the promise.
2. The slowest three callbacks of the week, each with a one-line reason.
3. How often the backup or manager had to step in.

Emergencies are also valuable work. Millennial homeowners led emergency home spending in 2025 at an average of $1,519^[5]. Missing the promise on even a few of those calls a month is a measurable loss, which makes the weekly review worth its fifteen minutes.

Example: a month of callback data

Example. A five-tech HVAC company sets a 15-minute promise in January. The first weekly report shows most emergency calls called back on time, but the slowest three took 41, 38 and 33 minutes. All three were cases where the primary tech was already on a night job and did not see the page. The backup step existed on paper, but the answering service had been told to keep trying the primary.

The fix is a rule change, not a reprimand: if the primary is on an active job, calls go straight to the backup. The answering service script is updated, and the next three weekly reports show every emergency callback inside 20 minutes, with most inside 15.

A separate promise for routine messages

Emergency calls need a promise in minutes; routine after-hours messages need one too, measured from the start of the next business day. A simple version: every message left overnight is returned by 9 a.m., urgent ones first. Keep the two promises separate in your reports, because mixing them hides problems. A great emergency callback time can sit alongside a morning queue that does not get cleared until noon.

The routine promise should be in the greeting as well, so callers who are not in an emergency know exactly when to expect a call and are less likely to call a competitor overnight.

Tools that make the clock visible

• Paging that sends call details by text and push notification at the same time, with a one-tap acknowledgment.
• Automatic escalation timers in the answering service, phone system or AI answering rules, not in someone's memory.
• A shared log where first contact, page, acknowledgment and callback times are stamped automatically.
• Call recordings or summaries so the morning review can see what was said, not just when.
• A weekly report sent to the owner without anyone having to build it by hand.

None of these need to be expensive. What matters is that the times are captured by the system rather than typed in later, when memories have softened.

When the promise keeps slipping

• Look for the pattern first: a particular tech, time of night, day of week or kind of call.
• Check the tools: dead phones, silent modes, app notifications switched off, poor coverage at home.
• Check the handoff: whether the page had enough detail for the tech to call back without hunting.
• Lengthen the promise if you truly cannot meet it, and tell callers the new time honestly.
• Add capacity in peak weeks: a second on-call tech, a night dispatcher, or an answering layer that triages and books routine calls so the techs only see emergencies.

The greeting and the callback promise should always match; see after-hours phone greetings that keep callers for wording that sets the expectation honestly.