In a heat run the job is triage, and nobody has written the script for it
When the phone rings twice as often as your board can absorb, every booking becomes a judgment call, and untrained judgment is where your best jobs get lost.
Left alone, a phone under pressure books whoever called first and whoever was easiest to talk to. Neither is the job you most want, and one of them is fifteen minutes past the far edge of your service area.
Write the triage rules before the season, then hand them to whoever answers. Which conditions get today. What an existing maintenance customer is owed. What happens with an elderly customer with no cooling, which is a safety question and deserves an explicit rule rather than a shrug. Which calls get an honest offer of tomorrow instead of a soft promise for tonight.
Give the phone permission to say no clearly. Turning down a job outside your zone in thirty seconds is better for everybody than booking it, running it late, and eating the review. Log every one of those declines, because that list is your real service area.
Then measure what triage did. Booked jobs per hundred calls during the peak week, compared with a normal week. If the gap is large, the loss is on the phone, not in the field.
The person you are calling back at eleven in the morning is asleep
A large share of this valley works nights, which means your standard callback habits are timed to reach people at the worst possible hour of their day.
Ask on the form and on the call when someone is actually reachable, and offer real windows including early morning and late evening. Then honor them. A callback attempted once at ten in the morning, followed by a voicemail, is not a follow up.
Track attempts by hour and look at which ones connect. Most shops discover a pattern within a month and have simply never looked at it.
Text is often the right channel here, with consent and with the caller's phone number. Someone who cannot answer at eleven can read and reply to a message at eleven, and a text with an arrival window and a technician's name gets responses that a call never gets.
Set an internal rule for speed to first contact on a new inquiry and put a number on it. Minutes matter far more than anything you might change on the website.
A gate code you did not ask for costs you the afternoon
Gated communities are common across the valley, and an intake that skips access details converts a booked job into a second trip and a wasted slot.
Build access into the intake script and the form rather than into the technician's problem. Community name and gate code or call box entry, whether the equipment is on the roof, whether a side yard is locked, whether a dog is loose, whether anyone will be home.
Add the authorization question early. Is the person meeting the technician able to approve the work and pay for it. In a rental, is there a property manager, and what is the standing dollar limit before someone has to call an owner. Establishing that at intake removes a full day of waiting later.
For anything visible from outside, ask whether an association is involved. Approval steps are common here for exterior work, they vary by community, and finding out at the quote stage rather than the install stage protects your schedule. Tell the customer it is worth confirming with the association before you order equipment.
Count second trips as their own metric. Every one is a slot you sold twice and were paid for once, and the cause is nearly always a field the intake form did not have.
The quote nobody signed in July is the install you book in October
Replacement proposals go quiet in the busiest month because nobody in the office has time to chase them, and most shops then never look at them again.
Keep every unsigned replacement quote in one list with the date, the equipment, the area and the reason it stalled. Not in an inbox. In a list somebody owns.
Follow up on a schedule that survives the season. A same week call, a call the following week, then a deliberate revisit when the shoulder months arrive and your install crews have room. The customer who limped through August on a failing system is a genuinely warm prospect in October.
Log the reason each one stalled in plain words: price, a spouse, waiting for a rebate, a landlord, an association, wanted to try one more repair. Read the list monthly. Half your objections have a fixable cause in how you present a proposal.
Do not repeat a price you can no longer honor. Equipment costs move, so revisit with a fresh proposal and say why it changed. Honesty at that moment is worth more than the small awkwardness of it.
Cancellations are a metric, and in August they are the biggest one you do not track
A booking is not revenue, and the gap between what you booked and what you ran is where a busy summer quietly loses money.
Count four numbers separately: calls answered live, jobs booked, jobs actually run, invoices collected. Most shops track the first and the last and are blind in the middle, which is exactly where the losses sit.
Cancellations and reschedules deserve their own line, with a reason attached. A customer who waited past the window and called somebody else is a different problem from one whose landlord declined the repair, and only one of the two is your fault.
Reconcile the dispatch board against lead source at least monthly. Without that reconciliation you cannot say which marketing produced which revenue, and every conversation about budget becomes an argument about opinions.
Watch the collected number as closely as the booked number. Work run and not paid for is a conversion failure at the end of the funnel, and in a trade with a lot of same day work it hides easily.
Call your own after hours line from a Henderson number at midnight
Mystery shop your own intake before you spend a month arguing about a button, because one metro will never produce enough traffic to settle small questions on a page.
Run the test yourself. Call at midnight, call during the peak hour on the hottest afternoon, submit the web form and time the response, and ask a question a nervous homeowner would ask. Write down what happened. The results are usually uncomfortable and always more valuable than a heat map.
Understand the statistical limit honestly. A single valley does not generate enough inquiries for a small change in button color or headline wording to show a real effect in any sensible amount of time, and heat makes the comparison worse by moving demand week to week.
So change big things and hold them. The after hours answer, the diagnostic fee presentation, the arrival window promise, the intake script, whether a financing option appears with a replacement quote. Give each one long enough to accumulate real volume, and compare stretches of weather that actually resembled each other.
Then listen to calls. Fifty recorded calls, with your notice practices confirmed by your own counsel, will tell you more about why inquiries leak than a year of page level testing. The phone is the product here.
Questions we actually get
- Where do most HVAC inquiries actually get lost?
- Usually on the phone rather than on the website. The common leaks are calls that go unanswered during the busiest hours, callbacks timed for people who work nights, access details nobody collected so the visit fails, and replacement quotes that go quiet because the office was buried in July. Audit those four before spending anything on redesigns or additional advertising.
- Is an answering service worth it?
- It depends entirely on the script and on what happens next. A service that takes a name and a number, then hands you a message at nine the next morning, adds little. A service with your triage rules, your service area and a real callback commitment can be worth a great deal in a market where a lot of people are awake at three. Test it by calling your own line at midnight, several times, before you renew.
- How do we run split tests when we only serve one metro?
- You mostly do not, at least not for small changes. One valley will not produce enough inquiries for a headline or a button variant to show a reliable difference in a reasonable window, and swings in temperature move demand enough to swamp small effects anyway. Change substantial things instead, hold each long enough to accumulate volume, and compare stretches of comparable weather rather than adjacent months.
- What should we measure every week?
- Five things: calls answered live as a share of calls received, jobs booked per hundred inquiries, jobs run against jobs booked, average time to first contact on a new inquiry, and unsigned replacement quotes outstanding. Add cancellations with reasons attached during the cooling season. If a report does not reconcile back to work actually run and collected, it is measuring activity rather than results.
- Should we ask for reviews right after the job?
- Yes, and ask on the spot rather than by email three days later. The technician who just made a house comfortable in August has more goodwill than any follow up message will ever have. Ask every customer, never offer anything of value in exchange, and never write or edit the words for them. Platforms remove incentivized reviews and the damage to your reputation lasts longer than the gain.
