Collision leads that don’t book die at one of six points: they were the wrong leads, nobody replied in time, the first call ended without an estimate time, the estimate visit lost them, nobody followed up, or a price or insurance question went unanswered. Score your last 30 inquiries against those six stages. The biggest drop is your first fix.
Score your last 30 inquiries
Thirty is enough to show a pattern and few enough to finish in an evening. Pull them from everywhere they land: the phone log, call tracking reports, web form emails, texts and Facebook messages.
Count people, not actions. In an ad report, a conversion is an action such as a call or a form, and one person can take more than one. A driver who fills in your form and calls 10 minutes later can show up twice.
Cross out existing customers checking on their car, vendors and spam. Give each remaining person one row and mark how far they got:
| Stage | Count it if | Example | Yours |
|---|---|---|---|
| Inquiry | A new person asked about a repair | 30 | 30 |
| Fit | Collision work you want, from your area | 24 | |
| Reached | A person at the shop talked with them | 20 | |
| Estimate set | The talk ended with a time or a tow-in | 12 | |
| Estimate done | They showed up, or sent photos for one | 10 | |
| Booked | The car came in for repair | 6 |
Give everyone who didn’t book one lost reason: not a fit, never reached, no estimate time, no-show, total loss, went to the insurer’s shop, price or deductible, went silent, or still open.
Then find the biggest drop between two rows. In the example, 6 of 30 weren’t a fit, but the biggest drop is the phone: 20 people talked to the shop and only 12 left with an estimate time.
That’s eight drivers who wanted a repair and hung up without a time. Better ads won’t fix that.
Put a number on it. In the example, 6 of the 12 drivers with an estimate time booked. If those eight had gotten a time and booked at the same rate, that’s four more jobs, or $15,200 at a $3,800 example RO, from inquiries you already paid for.
Wrong leads
Before you blame the front desk, check what’s coming in. The usual misfits:
- Glass only. Windshields and side glass belong at a glass shop.
- Mechanical. Brakes, check-engine lights, AC. These come from ads and pages built on broad terms like “auto repair” or “car repair near me.”
- Too far away. Drivers 40 minutes out, because the ad radius or your Google listing’s service area is wider than people will drive for a bumper.
- Price-only shoppers. They want a number over the phone and won’t come in. Keep them apart from customer-pay drivers who ask about price first, because those are real jobs.
- Work you don’t want, like custom paint or restorations, plus job seekers and sales calls.
Each has a fix. Add negative keywords to your ads, which are words that stop an ad from showing, like windshield, glass, brakes, oil change, tires and jobs. Set the ad radius from your own records by pulling the ZIP codes off your last 50 ROs.
Then say plainly on your website and Google listing what you fix and what you don’t. A line like “Collision repair, paint and calibrations. No mechanical work.” stops a lot of brake calls before they start.
AE Collision Center in Tampa had this problem before we started: the ad account attracted searches that didn’t fit the repairs the shop wanted. We rebuilt the campaign around local collision repair searches, cut the irrelevant traffic and corrected call and form tracking.
Joniel’s team answered the inquiries, wrote the estimates and closed the jobs. For August, month one, the shop reports 9 closed jobs and about $40,000 in repair revenue, at about $132 in ad spend per reported job (ad spend only, before our fee). The AE Collision Center case study has the full numbers, and what one booked job really cost breaks down the math.
If 10 or more of your 30 aren’t a fit, fix targeting first. Every later stage is working with the wrong cars.
Slow or missed replies
Mark how many minutes each fit inquiry waited for a person. Missed calls live in the phone log, not your CRM or spreadsheet, so check both.
Then split them into fast replies, under 15 minutes, and slow ones, an hour or more. Thirty is a small sample, but if the fast group books and the slow group doesn’t, you’ve found the leak.
Look hard at the ones never reached at all: a voicemail nobody returned, a form sitting in spam, photos texted to a landline. Those are pure loss, and the cheapest to fix.
The fix is coverage and ownership: one named person per block of the week, and a backup phone that rings when they can’t answer. We lay out the setup, and what a missed call costs in dollars, in the real cost of a slow lead response.
The first phone call
This leak hides well, because the calls sound fine from the office. The CSR was polite and the driver said thanks. Nobody set a time.
Listen for the lines that end calls without one:
- “Call us back once you’ve talked to your insurance.”
- “Just bring it by whenever.”
- “We’re pretty backed up right now.”
- “It’s probably around $1,500.” A phone price either scares the driver off or sets a number your real estimate will blow past.
- “We’re not in network with them.” The driver hears “we can’t fix your car.”
Score 10 recent calls. If you record calls, play them with your CSR. If you don’t, role-play 10 with you as the driver and ask four questions of each:
- Did we get the name and number in the first minute?
- Did we find out whether the car is drivable?
- Did we ask where the claim stands: filed or not, whose insurance, any photo estimate yet?
- Did the call end with a specific time, picked from two we offered?
Every call that fails question 4 is a lead you paid for and handed back.
The full call flow, word for word, is in our front desk phone script.
The estimate visit
A driver who shows up for an estimate still hasn’t picked you. Some cars won’t be repaired anywhere: CCC’s Crash Course 2026 report found total losses hit a record 23.1% of all claims in 2025, so code those as total loss, not as lost to a competitor. For everyone else, check five things.
Wait time
Block estimate slots on the estimator’s calendar. A driver booked for 10:00 who waits 35 minutes while the estimator finishes a supplement call is already thinking about the other shop. If you’re running late, text before they arrive.
The process
Walk the car with them and point at the damage. Then explain the order: your estimate goes to the insurer, parts get ordered on approval, they drop off on a set date, and hidden damage found at teardown goes back as a supplement. Put that on one page with your number on it.
Insurance questions
Answer what you can: when the deductible gets paid, what the claim number is for, and why the insurer’s photo estimate is a starting point, not the final number. Don’t promise what the insurer will cover. Say, “We’ll send everything to your adjuster and call you when we hear back.”
Calibrations come up more often now. The same CCC report found 28.3% of repairable estimates included calibrations in 2025, up from 21.8%. When one’s on the estimate, explain it in a sentence: the cameras and sensors behind that panel have to be re-aimed so the safety systems work right.
Rental
Ask about rental coverage at the estimate, not at drop-off. If the other driver’s insurance is paying, that’s who to ask about a rental. Line it up for the morning they drop off, so they’re never without a car.
Photos
For photo estimates, tell drivers exactly what to send: all four corners from a few steps back, the damage straight on and at an angle, the VIN and the odometer. A vague request gets three blurry close-ups and a second round of texts.
Follow-up after the estimate
Plenty of estimates end with “Give us a call when you’re ready.” Meanwhile the driver is juggling the adjuster, the rental counter and work. Silence reads as “they don’t want the job.”
Run a short, fixed rhythm, owned by one named person, usually the estimator who wrote it:
- Same day: text a summary and two drop-off options.
- Day 2: call. Ask what they’re waiting on: the adjuster, a second estimate, the rental.
- Day 5: text one useful update, like parts availability or an open slot.
- Day 10: last check. If they’ve gone elsewhere, close it out with a lost reason.
Every touch needs a reason to reply, and “just checking in” isn’t one. A good one reads like this:
- “Hi [name], it’s [your name] at [shop]. Your [car] estimate is done and sent to [insurer]. We can take it Tuesday or Thursday morning. Which works?”
When someone goes elsewhere, ask where. “The insurer’s shop,” “a cheaper shop” and “the dealer” each point to a different fix.
On the scorecard, look at the estimates that didn’t book. If most got no follow-up within one business day, this is your leak.
Price and insurance objections
A price question is often a driver deciding whether to file a claim at all. A LendingTree survey of 2,000 US consumers in June 2024 found 39% of insured drivers who had an accident or incident paid for repairs without filing a claim. In the same survey, 44% said their deductible was higher than the cost, and 42% didn’t want their insurance to go up.
Treat those callers as real jobs. Offer a written estimate, in person or from photos, so they can decide with a real number instead of a phone guess.
“My insurance told me to go to their shop” is the other big one. In CCC’s “Moments of Truth” study with Magid (2,400 policyholders, reported in October 2024), nearly one-third of consumers said an insurer’s recommendation was the deciding factor in where they took their car. About two-thirds decided on something else, such as past experience, friends and family, price, location or reviews.
Your CSR’s answer:
- “That’s common. It’s your car and your choice where it’s fixed. If you’d like us to do it, we’ll work with your insurance company on the claim.”
Use that last line only if it’s true for your shop. If you’re not on DRPs, there’s more on growing without DRP referrals.
When a driver says your price is higher, find out what they’re comparing. If it’s the insurer’s photo estimate, walk them through the lines the photo missed: the calibration, the blend, the hidden damage.
Count your lost reasons here. If “went to the insurer’s shop” leads, work on the CSR’s answer and your reviews. If “price” leads on customer-pay jobs, look at how the estimate is presented before you cut the price.
What to do this week
- Pull your last 30 inquiries into one sheet, one row per person, from the phone log, call tracking, forms and texts.
- Fill in the six stages and give every row that didn’t book one lost reason.
- Find the biggest drop between two stages, and fix only that stage for the next 30 days.
- Score 10 calls with your CSR against the four call questions.
- Put a date on the calendar 30 days out to score the next 30 the same way.
Questions owners ask
What share of inquiries should turn into booked jobs?
We don’t know of a trustworthy collision-wide number. The rate swings with your mix of insurance and customer-pay work, how far drivers come, and how many total losses come through. The benchmark that matters is your last 30 against your next 30.
Should we pause the ads until this is fixed?
If the leak is the phone, the estimate or follow-up, more leads just leak faster, so fix that stage first or in the same week. If the leak is fit, change the targeting, not the budget.
Our CSR says the leads are junk. How do we check?
Look at the fit row. If most of the 30 were collision work from your area, the leads aren’t the problem and the drop comes later. If they weren’t, list the junk by type, like glass, mechanical or out-of-area, and fix the targeting for each.

