ILLUSTRATIVE EXAMPLE — composite scenario; replace with verified customer data before publication.
You can't cut the marketing that isn't working if you can't see which marketing isn't working. Two agencies, two invoices, one owner staring at a lead count that never tied back to a booked job. This is a story about call tracking for garage door marketing — how seeing the calls plainly let a shop cancel the right invoice and come out ahead.
The shop here is a composite: seven trucks across two locations in Ann Arbor and Brighton, two marketing agencies running in parallel, and no honest way to tell which one earned its keep. Names and figures are representative examples, not a specific verified customer.
Both agencies sent monthly reports. Both reports looked fine. Impressions up, "leads" up, dashboards green. But the owner ran a garage door shop, not a marketing dashboard, and none of it told him the one thing he needed: which spend turned into trucks rolling and money in the bank.
The gap was the phone. Calls came in across two locations, got answered inconsistently by whoever was free, and got logged — when they got logged — in someone's memory or a sticky note. So a call that started with an Ann Arbor search ad and a call that started with a Brighton referral looked identical by the time it reached the schedule. The agencies could claim credit for anything. The owner couldn't prove or disprove any of it.
Which meant he was allocating budget on gut feel. He suspected one agency was carrying the other, but "suspected" doesn't cancel a contract. He needed the calls themselves to tell him true cost per acquisition, channel by channel, before he pulled the trigger on a $2,000-a-month decision.
He forwarded both location numbers to Ava, the 24/7 AI receptionist, and let her answer, capture, and log every call across Ann Arbor and Brighton in one place. Both existing numbers stayed put — no new lines, no changes to the ads either agency was running. Setup was done-for-you and live in under 24 hours.
Now every call got captured the same way: name, phone, address, issue, and the source when the caller offered it, all in an instant SMS and email summary and a single running log. The two locations stopped being two black boxes. For the first time, the owner could see every call that came in, whether it booked, and roughly where it came from — in one view instead of two agency spin decks.
He didn't need a data team. He needed the phone to stop losing the evidence.
These figures are illustrative and internally consistent — captured leads exceed booked jobs, and booked jobs times a plausible ticket approximate the revenue. Repair tickets run $150–$450, spring jobs $250–$500, opener installs $350–$650. Treat them as a model, not a promise.
Day 0–30 — unified tracking. Both locations' calls landed in one log. Across the two cities, Ava captured every inbound call and tagged source where the caller named it. The immediate finding wasn't revenue — it was visibility. The owner could finally count booked jobs against the channel that produced them.
Day 31–60 — true CAC appeared. With two months of clean call data, the cost per booked job per channel stopped being a guess. One agency's channel was producing booked jobs at a defensible cost. The other's "leads" mostly didn't book — the calls were thin, off-market, or never converted. The invoice that looked equal on paper looked very different once the phone did the accounting.
Day 61–90 — cut the right one. The owner cancelled one agency. Total marketing spend dropped about 18%. And because he cut the channel that wasn't booking — not the one that was — bookings didn't fall. They rose about 15%, partly from reinvesting attention (and a little budget) into the channel the call log proved was working.
The owner's line: "The call data told us which invoice to cancel."
Agencies grade their own homework — your phones grade it honestly. Impressions and "leads" are inputs. Booked jobs are the output that pays your trucks. Until every call is captured and logged in one place, you're trusting the vendor's version. Full, consistent call capture is the missing measurement layer under any garage door marketing program.
True CAC lives in the call log, not the ad platform. Cost per click and cost per "lead" tell you nothing about cost per booked job. Tie the call to the outcome and the real number appears — often very different across channels that looked equal.
Cut the channel that doesn't book, not the one that does. The whole point of tracking is to fire the right invoice. Cut blindly and you might kill your best source. The dollars-and-cents version of this — what a captured-versus-missed call is worth — is spelled out in what missed calls actually cost a shop.
A flat answering cost pays for itself in decisions, not just calls. The tracking that let this shop cut 18% of spend rode along with the answering service at a flat monthly rate. That combined value is worked through in the pricing and ROI of a flat-fee AI receptionist, and the broader habit of measuring where your work comes from is covered in where garage door leads come from.
The agencies didn't get worse. The measurement got better — and the invoice that couldn't survive it got cancelled.
Call the live demo and have Ava call you now — hear exactly what your customers will hear when they call your shop.