The expensive part of an HVAC job is not the install. It is the week after the estimate sits unanswered.
A tech ran the load calc. You priced the equipment. The homeowner said "let me talk to my spouse." Then silence. Five days later they hired the company that texted twice and offered a clear next step.
I work with trade operators on the Jersey Shore. The pattern is consistent across HVAC and plumbing: intake gets attention, follow-up does not. Owners invest in Google ads and nicer trucks. The quote PDF still dies in someone's email.
This post is the follow-up system I recommend when an HVAC shop is "busy" but the close rate on estimates is soft.
Where HVAC estimates actually fail
Failure is rarely the price sheet alone. It is the gap between site visit and decision.
Typical leak points:
- Same-day send is late. Tech leaves at 3 PM. Estimate emails at 9 PM. Homeowner already went to bed and forgot the details.
- One channel only. Email-only follow-up loses to text. Text-only loses when the number is wrong or filtered.
- No decision helper. The PDF lists equipment and dollars. It does not answer "what happens if we wait until October" or "what is included vs allowance."
- No owner nudge. After day three, nobody owns the chase. The office assumes the tech will call. The tech assumes the office will call.
- No review of why deals died. Lost reasons stay in someone's head. Next season you repeat the same misses.
If you only fix pricing and never fix follow-up, you will keep paying for leads you already earned once.
The 10-day sequence (steal this)
Use this as a default. Adjust tone to your brand. Keep it short.
| When | Channel | Message job |
|---|---|---|
| Day 0 (within 2 hours of visit) | Email + text | Deliver estimate + one-sentence summary of recommendation |
| Day 1 | Text | "Any questions on the options? Happy to clarify equipment vs labor." |
| Day 3 | Call or voicemail + text | Answer the stall: financing, timeline, rebate, or install window |
| Day 5 | Side-by-side of Option A vs B in plain language (efficiency, warranty, noise, install time) | |
| Day 7 | Text | Scarcity that is true — crew calendar, equipment lead time, season cutoff. Never fake urgency. |
| Day 10 | Close the loop: "Should I release the hold on this install week, or keep you on the list?" |
Rules that keep this from feeling spammy:
- Every touch adds information. No "just checking in."
- One clear question per message.
- Stop when they say no. Log the reason.
- If they say "call me next month," schedule it. Do not keep pinging weekly.
For a related phone-side leak, see The Missed Call Problem. Missed calls and dead estimates are the same disease: revenue that already raised its hand, then got silence.
What to automate first (and what to leave human)
Automate the boring, timed, template-friendly parts. Keep judgment human.
Automate:
- Sending the estimate PDF + summary within a set SLA after the job is marked "estimate complete"
- Day 1 / Day 3 / Day 5 / Day 7 / Day 10 touches if no reply
- Logging opens/replies into whatever you use as a CRM (even a shared Airtable is fine)
- Reminding the assigned salesperson when a high-ticket quote has no activity for 48 hours
- Creating the "next month" callback task when a homeowner asks to pause
Keep human:
- Price changes and discounts
- Equipment substitutions
- Anything that sounds like negotiation
- Angry or confused replies
- Final close call on five-figure installs
If your "automation" starts arguing about price in SMS, turn it off.
A simple ops setup that does not need a Silicon Valley stack
You can run this without a six-figure field-service platform.
Minimum viable stack:
- Job status field — estimate scheduled / estimate sent / negotiating / won / lost / deferred
- Estimate value + equipment type — so high-ticket quotes get faster human attention
- Assigned owner — one name, not "the office"
- Message templates — Day 0 through Day 10, edited in your voice
- Automation tool — Make, a lightweight CRM sequence, or a custom workflow if your volume justifies it
Solo and small shops: start in a spreadsheet for two weeks. Track every estimate. Run the sequence manually. When you see the pattern, automate the steps you already trust.
Shops already drowning in admin can wire this into broader NJ service-business automation without rebuilding the whole company.
Math you can run on last month's book
Pull last month's estimates. Fill this in with your numbers:
| Metric | Your number |
|---|---|
| Estimates sent | |
| Estimates with zero follow-up after day 0 | |
| Estimates closed | |
| Average ticket on closed installs | |
| Estimates you personally believe were winnable with better chase |
Illustrative example (replace with yours):
- 20 estimates sent
- Average ticket $6,500
- 4 closed (20%)
- You judge 3 of the losses were "no follow-up / slow follow-up," not pure price
Recovering one of those is $6,500 top-line. That usually dwarfs the cost of a sequenced text/email system for a year.
I am not claiming a national close-rate statistic here. Run the table on your own book. The number that matters is the one in your QuickBooks.
HVAC-specific content that helps follow-up land
Generic "hope you're well" texts underperform. Better day-5 content for HVAC:
- What is included in labor vs equipment
- Warranty length in plain English
- Expected install duration and home disruption
- Filter / maintenance plan after install
- What waiting until peak season does to lead times (only if true in your wholesale channel)
- Rebate or utility program checklist if you actually process them
If you sell heat pumps in Monmouth or Ocean County, say so in local terms: older housing stock, panel upgrades, and winter emergency calls are the context your buyer lives in. Pair this with permit readiness when the job needs one — NJ pre-permit checklist and permit automation overview.
Red flags in "AI follow-up" pitches
- Promises to "never lose a lead again" with no logging or stop rules
- Scripts full of banned fluff ("speed up your close rate") and no equipment detail
- No way for a human to take over mid-thread
- No lost-reason field
- Requires a 12-month contract before you see a single sequence run on your data
Good systems are boring: timers, templates, ownership, and a human who can jump in.
Bottom line
HVAC companies do not mainly lose at lead gen. They lose in the quiet days after a good site visit. A 10-day multi-channel sequence, clear ownership, and automation on the repetitive touches will recover jobs you already paid to create.
Start with last month's estimate list and the table above. Fix the chase before you buy another ad package.