Every unanswered ring is a click you already paid for, twice
Missed calls are the largest and most invisible loss in a contracting business, and they cluster exactly when demand peaks.
Pull the call log by hour and day. Almost every contractor finds the same shape: a block in the early afternoon and another after five where the missed rate jumps.
Those are not neutral losses. In this trade the caller has a list, and the next number gets dialled within a minute.
Fix the capacity before touching the website. An overflow answering service that can actually take a job address, a second person cross-trained for the afternoon block, or a documented rule for who picks up when the dispatcher is on another line.
Then measure it monthly as a headline number: calls offered, calls answered, calls booked. Owners who see that table once tend not to argue about the fix.
The second contractor to call back is quoting against a signed proposal
Speed to first human contact predicts who books the job more reliably than anything you can change on the page.
Form leads should trigger a phone attempt in minutes, not an autoresponder. The email that says "we have received your request" satisfies nobody and buys the competitor time.
Set an internal target, staff it, and check the actual times weekly rather than assuming. Response times drift the moment the season turns busy.
Text is underused here. A short text from a named person, immediately, holds the lead while your office finishes the call in front of them.
After-hours leads need a rule too. Whether that is a service, an on-call phone or a first-thing queue, decide it deliberately instead of by accident.
Cut the intake form to what a dispatcher can actually act on
Fields are a cost, so keep the ones that change what the truck brings or whether the job is real, and delete the rest.
Keep: service address, home or commercial, the symptom, system age roughly, owner or renter, best number.
Delete: budget ranges, preferred contact method, marketing source as a required field, anything your CRM can infer.
Order matters as much as length. Ask the easy identifying questions first and the qualifying question last, so an abandoned form still leaves you a phone number to call.
Two forms usually beat one. A short emergency form and a longer replacement form, because someone requesting a quote for a full changeout will tolerate more questions than someone whose house is at eighty-six.
Ask which utility bills the house before anyone on your team says the word rebate
Rebate eligibility in this county depends on the meter, not the mailing address, and getting that wrong on the phone costs you trust at the exact moment you were building it.
FPL's Residential HVAC Program Standards require an active FPL residential account for its $200 per unit rebate. Lake Worth Beach Electric Utility, a citizen-owned municipal utility, serves 27,809 customers across the City of Lake Worth Beach, parts of unincorporated Palm Beach County and the Village of Palm Springs.
So the intake script gets one extra question: who sends the electric bill. It takes four seconds and it prevents a quote built on a rebate the customer cannot claim.
Put the same question on the replacement form as a simple choice, and train the office to say "we will confirm eligibility before it goes in your proposal" rather than quoting the amount cold.
Worth re-checking the program terms periodically, since these standards carry effective dates and change.
Reconcile the dispatch board against lead source or you are optimising blind
Until booked jobs are attributed back to the channel that produced them, every conversion decision is being made on anecdote.
One tracked number per channel, consistent listings, and outcome tagging in the CRM the dispatcher already lives in. Booked, not booked, out of area, wrong service, no answer.
Then reconcile monthly against the actual board. Marketing reports and completed job counts usually disagree the first time, and the gap is the finding.
Track by job type, not just count. A month heavy on diagnostics and light on changeouts is a different business result even at identical lead volume.
This is the least exciting work on the page and the highest leverage. Everything else is guessing without it.
A single county will not give you clean split tests, so test bigger things
Contractor traffic volumes rarely support detecting small differences, which means the honest method is large changes, sequential comparison and listening to calls.
Do not run a button colour test on a few hundred sessions a week and call the result a win. It will not be one.
Test things large enough to move behaviour: form length halved, a phone-first layout against a form-first layout, adding a licence and credential block above the fold, adding online booking to the maintenance path.
Run them sequentially over meaningful periods, note the season, and accept directional evidence honestly labelled as directional.
Then supplement with qualitative work that needs no sample size: listen to twenty recorded calls, read the abandoned form data, watch five session recordings, and ask your dispatcher what people ask before they book. Most of the real findings come from there.
Questions we actually get
- We answer most of our calls. Is that good enough?
- Look at the log before answering that. Most contractors who say this discover the misses are concentrated in the busiest hours, which are also the hours the highest-intent callers dial. A ninety percent answer rate spread evenly is a different business from a ninety percent rate that collapses to sixty between two and five.
- How fast do we really need to call back a web lead?
- Fast enough that you are the first conversation. In practice that means minutes, and it means someone owns the alert rather than everyone assuming someone else saw it. We would set a target, measure the real times weekly, and treat drift as an operational problem rather than a marketing one.
- If we shorten the form, will lead quality drop?
- Sometimes, which is why the question is which fields you keep, not how few. Keep the ones that qualify (owner or renter, home or commercial, service address) and drop the ones that only satisfy curiosity. Volume rises, and the qualifying happens in the first thirty seconds of a call instead.
- Can you A/B test our website properly at our traffic level?
- Not for small effects, and we will not pretend otherwise. A single metro contractor rarely produces enough sessions to detect a modest difference with any confidence. What works is testing large changes sequentially over meaningful periods, combined with call listening and form abandonment data, and labelling the evidence honestly.
- What should the office actually say when the phone rings?
- Confirm the address and whether they own the property, get the symptom, state your licence and that you handle the permit where one is required, then offer a specific arrival window rather than asking what works for them. Ending with a booked time slot rather than a callback promise is the single biggest scripting change most contractors can make.
