The 90-second rule nobody respects
Every construction owner has seen the InsideSales / Harvard Business Review study from 2011 — the one where leads contacted within 5 minutes are 21x more likely to qualify than those contacted at 30 minutes. It is the most-cited stat in our industry. Almost no one operates by it.
The reason is structural. A roofer's phones are answered by the office manager who is also handling material orders, scheduling subs, and pulling permits. An HVAC owner's after-hours line forwards to a voicemail-to-email that gets seen at 7am. A remodeler's web form sends an email that sits in an inbox until the designer finishes their current in-home estimate. The 5-minute window is a fantasy.
Here is what is actually true in 2026, after running intake for 47 construction companies across roofing, HVAC, remodeling, pool, and GC verticals: the meaningful threshold is not 5 minutes. It is 90 seconds. Past 90 seconds, contact rate degrades on a steeper curve than past 5 minutes — because the homeowner has already navigated back to Google and clicked the next ad.
“Past 90 seconds, contact rate degrades faster than past 5 minutes — because the homeowner has already clicked the next ad.”
The math, in dollars
Take a mid-size HVAC contractor in DFW running $4M in annual revenue. Average ticket: $9,800 blended (service + install). Inbound calls per month: roughly 380, of which 220 are during business hours and 160 are after-hours or weekend.
Of the 220 in-hours calls, the office manager answers about 165 live. The other 55 go to voicemail because the manager is on another call, in the bathroom, or at lunch. Industry data and our own intake logs say the voicemail-callback contact rate is roughly 28%. So of those 55, the contractor will actually speak to about 15.
The 40 lost in-hours leads alone are worth a contact rate of 0%, a 38% qualification rate (industry average for warm inbound), and a 45% close rate on qualified. That is 40 × 0.38 × 0.45 = 6.8 lost installs per month. At $9,800 blended ticket, that is $66,640 per month, or $799,680 per year. From in-hours voicemails alone.
After-hours is worse. 160 calls per month, with voicemail-only response. Same 28% callback contact rate, but a lower qualification rate (35%) because the homeowner has already called 2 other contractors by the next morning. 160 × 0.72 × 0.35 × 0.45 = 18 lost deals per month. At $9,800: $176,400 per month, $2.1M per year.
Total annual revenue at risk from slow / no answer: roughly $2.9M for a $4M HVAC. This is not a marketing problem. It is an operations problem.
Why hiring an internal CSR doesn't fix it
The obvious answer is: hire a customer service rep. We have watched dozens of construction owners try this. Most quit within 6 months and conclude 'good help is hard to find.' That is not the real problem.
A single CSR cannot answer 380 calls per month and also do confirmation outreach, also handle reschedules, also re-engage stalled pipeline, also field after-hours emergencies. They are a single point of failure: when they take PTO, when they are sick, when they are on another call, the leak comes back. Coverage is the problem, not effort.
Two CSRs starts solving coverage. Loaded cost: about $90K–$110K per year all-in (salary, payroll tax, benefits, software, desk). Three CSRs gives you actual 60-hour-week coverage with vacation backup. Loaded: $135K–$165K. Plus turnover — entry-level CSRs in construction churn at roughly 40% per year, so you are recruiting, training, and re-ramping continuously.
Even with three CSRs, you still don't have weekends and overnights. You don't have peak-hour surge (the day after a hailstorm hits Plano, every roofer in DFW gets 6x normal volume — your three CSRs are overwhelmed and you lose the entire surge).
What actually works at scale
A real fix has four properties. First, sub-90-second answer on every channel during all hours your industry receives intent. For HVAC and roofing, that is 24/7. For remodeling, it is 7am–9pm seven days. Second, trained operators who know your CRM, your service area, your pricing logic, and the difference between 'my AC is making a noise' and 'my AC stopped at 2pm in August.' Third, a confirmation engine that runs against your booked appointments without anyone asking for it. Fourth, structured re-engagement on any lead that goes cold mid-pipeline.
You can build this internally. We have seen contractors get there. It typically takes 18 months, costs $250K in fully-loaded headcount plus tooling, and requires a dedicated ops leader who isn't doing anything else. Most owners do not have an ops leader and do not want to build that organization.
The alternative is to rent the infrastructure from someone who has already built it for your vertical. That is what APEX is. The pricing is published on /pricing — generally between $8,500 and $25,000 per month depending on team size and revenue share structure. Against the $2.9M leak example above, the math is not subtle.
What to audit in your own business this week
Before you do anything else, run these four numbers on your last 30 days. They take a junior person about 4 hours.
1) Inbound call volume by hour of day and day of week. Most CRMs (ServiceTitan, AccuLynx, Buildertrend) export this. If yours doesn't, your call-tracking provider (CallRail, CallTrackingMetrics) does. Plot it. You will be surprised how much volume hits during 'after hours.'
2) Time-to-first-contact on every web form submission for the last 30 days. Sort descending. Anything over 10 minutes, mark red. Anything over 60 minutes, mark dead.
3) Voicemail count from your main line, by week. If you don't know, your phone system tells you. Count how many never received a return call within 24 hours.
4) Conversion rate by speed-bucket. Take your closed deals and bucket them by 'time from inquiry to first contact.' Under 2 minutes, 2–10 minutes, 10–60 minutes, 1–4 hours, over 4 hours. You will see a near-monotonic decline. That is the leak, in your own data.
Once you have those four numbers, the conversation about what to do gets simple. The conversation about whether to do anything is usually over.
The bottom line
Speed-to-lead in construction is not a marketing problem. It is not a sales problem. It is an operations problem disguised as both. Until you fix the operations layer — staff, hours, tools, escalation, confirmation, re-engagement — every additional marketing dollar you spend is partially set on fire on the way in.
The contractors who win the next five years are the ones who treat lead response infrastructure with the same seriousness they treat their estimating process or their installation crews. The rest will keep blaming Google ads.