If you call an HVAC lead back in 90 seconds, you close it. If you wait until tomorrow, you don't.

You spend $80 to get a homeowner to fill out the "request service" form on your site. Your tech's on a roof. Nobody calls the customer back for four hours. By then, they've called two other companies.

You're eating the cost of that $80 lead, and the only thing that mattered was who dialed first.

Speed-to-lead is the cheapest revenue lever an HVAC shop has. You don't need more marketing budget, better SEO, a bigger sales team, or a single new truck. You need one phone call inside 90 seconds.

The decay curve is the most-replicated finding in lead-response research, and it hits harder for emergency service than for almost anything else.

Hatch ran the numbers across 132,188 HVAC speed-to-lead campaigns: 88% of contractors take longer than 5 minutes to reply. 3% reply inside a minute. We see the same pattern on our own test-call line: most shops don't pick up live. Whoever you're racing, they're slow.

That's relative qualification rate, not close rate. The original MIT study found shops dialing at 5 minutes are 21 times more likely to qualify a lead than shops dialing at 30. Twenty-one times.

HBR's larger 1.25M-lead replication was less dramatic but pointed the same direction: call inside the hour and you're 7x more likely to qualify, 60x more likely than waiting until tomorrow.

Speed doesn't magically close the deal. What it does is get you on the phone with a live human before your competitor does. On a no-cool call in July, that's the entire ballgame.

For an HVAC contractor, the asymmetry is sharper because the decision the customer is making isn't "do I want this." It's "who's available now." A homeowner whose AC compressor quit at 4pm in July isn't comparison-shopping. They're dialing whoever picks up. Vantal picks up.

The 3-question intake script that books in 90 seconds

You don't need a script for the friendly part. You need one for the triage part. Three questions, in this order, every time.

First, what's the symptom? Specifically: no cold air, no air at all, a strange noise, or a smell. The answer routes the call.

Then ask when it started. Today routes differently than two weeks ago. A 4pm-today no-cool in July with an upstairs nursery is the call you drop everything for. A two-week intermittent rattle is a paid diagnostic, scheduled for tomorrow.

Last, the address. You need it before you can offer a window. Distance is the silent killer of accurate scheduling.

ServiceTitan's own data puts a number on the gap: HVAC shops book 38% of inbound calls on average. Shops with fewer than 5 techs book just 24%, less than half what a 25+ truck shop hits. We built Vantal for the 38% shop trying to reach 60%. Same script the 25-truck shops run, no 25th tech needed.

The 2-tier on-call rotation: who answers, who actually rolls

The fastest callback in the world is useless if your dispatcher can't find someone to actually run the call. The fix is to publish an on-call rotation the intake person can see in real time and warm-transfer to.

Two tiers covers most shops.

Tier 1 is the intake desk. Takes the call, asks the three triage questions, books non-emergencies live, transfers true emergencies. Open 24/7. Can be a person, a service, or an AI agent. The only thing it can't be is voicemail.

Tier 2 is your on-call tech. Takes the warm-transfer for emergencies and runs the job. Rotates weekly. Gets paid for the rotation, not just the jobs that come in.

What breaks this in practice is tier 1 waking up tier 2 for non-emergencies. The triage rules in step one are what prevent that. If the intake person follows the rules, tier 2 only takes 3 to 8 calls a week, and only at hours where a paying job pays for the wake-up.

Your phone tree, drawn once, set up in an afternoon

The setup is a one-time afternoon of work with your phone provider. Most shops are on RingCentral, Vonage, or a softphone bundled with their FSM (Housecall Pro, ServiceTitan, Jobber). All of them support time-of-day routing.

The rule, drawn once:

The whole point is that the homeowner never lands on a recording. The moment they do, they hang up and dial the next listing.

This is not a fringe idea. Alex Hormozi profiled a business that pays one full-time person to do nothing but dial inbound leads inside 60 seconds. Their close rate on inbound is 55%. Sales went up 391%. The only differentiated input was speed.

Hormozi's framing: success loves speed. You can run the same play without the full-time hire, because the AI never sleeps.

2 problems speed-to-lead won't solve (and what to fix first)

Speed-to-lead isn't a script for closing harder. It's the opposite. The intake person closes nothing. They qualify, they book, they transfer. The actual sale happens at the door, with the tech, with the diagnostic already paid for and the trust already there. The faster the booking, the less your tech has to sell when they show up.

This also isn't a system you can layer on top of a broken dispatch. If your trucks are always behind, a faster intake just creates more disappointed customers. Fix the throughput first, then turn on the speed.

What to do tomorrow

Four moves that take less than a day. None of them need a developer.

• Audit last month's web leads. Pull the response times from your CRM. Count how many went unanswered past 5 minutes. That gap is your competitive opening.
• Write your 3-question intake script. Symptom, when it started, address. One page. Hand it to whoever picks up tonight.
• Set up time-of-day routing on your phone provider. RingCentral, Vonage, or your FSM softphone. One afternoon, no developer needed.
• Run a test call on your own line. Listen for the moment you'd hang up if you were a homeowner with no AC. Fix that.

Hear Vantal answer your shop's line in 30 seconds

Vantal is the AI intake for tier 1. It picks up in three rings 24/7, runs the triage questions in your own rules, books into your calendar, and warm-transfers true emergencies to your on-call cell. English and Spanish on the first ring.

If you want to hear what it'd sound like on your own shop's line, try it on your shop's number. The test takes 30 seconds. We email you the recording and the transcript.