The 7 Numbers That Decide Whether Your Auto Shop Makes Money in 2026

Why these seven, and why the dashboard metrics aren't on the list

Walk into any shop or dealership on a Thursday at nine and you can feel whether it makes money before you ever see a P&L. The bays are full or they're not. The service writer is closing work or apologizing. The phone is getting answered or it's ringing into a voicemail box nobody checks. The seven numbers below are the ones that quietly decide the month, and most owners never look at five of them.

Car count, total revenue, and Google review stars are the metrics everyone watches. They're outputs. They tell you what already happened and give you nothing to grab onto. The seven here are inputs you can move this week with a phone call and a habit, and each one drops straight to the bottom line. Typical net margin in this business runs 15 to 20 percent, so a few points of leak on any of these is the difference between a good year and a flat one.

This holds whether you run an independent repair shop, a used-car lot with a service lane, or a franchise dealership pushing a thousand ROs a month. The names on the door change. These numbers don't. For each one you get the benchmark, why it decides profit, the one cheap lever that moves it, and which Auto Advisor agent does the moving. If you want your own version of these running live, the no-login demo shows the whole dashboard on sample data.

1. Average repair order: the number that multiplies everything else

Average repair order is total sales divided by the number of cars you touched. In 2025 the typical figure sat around $428, and healthy independents run a band of $350 to $500. ARO is the multiplier on your entire operation, which is why it sits at the top of this list. Every other number on this page either feeds ARO or gets multiplied by it.

Here's why it quietly decides profit. A shop doing 30 cars a day at $428 books about $12,840. The same 30 cars at $478, fifty dollars higher per ticket, books about $14,340. That's roughly $1,500 a day, close to $30,000 a month, on the same car count, the same bays, the same techs. You didn't work harder. You stopped leaving fifty dollars on each ticket.

The cheap lever is declined-work follow-through. Your techs already find the worn brakes, the weeping water pump, the cabin filter that looks like a forest floor. Most of that gets mentioned out loud, half-recorded, and never chased. Put every recommendation in writing on the estimate with a photo, then bring the declined items back up at the next visit. You're not selling harder. You're finishing the inspection you already did.

The Analyst agent moves this one. It watches ARO by writer, by day, and by job type, and tells you each morning where the number slipped and which declined jobs are still open. The Front Desk agent makes the follow-up call so the recommendation doesn't die in the system.

2. Estimate close rate: where quoted work goes to die

Close rate is the share of estimates you write that turn into authorized, paid work. Strong independents land around 70 percent. If yours sits at 50, you're not short on demand. You're letting one of every five jobs you already quoted walk out the door.

This one decides profit because the work is already found, priced, and sitting in front of you. There's no marketing cost to recover it, no new car to attract. A shop writing $40,000 in estimates a week at 50 percent closes $20,000. The same estimates at 70 percent close $28,000. That extra $8,000 a week cost you nothing but the conversation, and it falls almost entirely to margin because the parts and labor were already going to be there.

The cheap lever is speed and clarity on the approval. Close rate dies in the gap between sending the estimate and getting a yes. The customer is at work, the text sits unread, the car waits, the bay sits. Three things move the needle fast:

• Send the estimate with photos and a one-line, plain-English why, not a wall of line items a customer can't read.
• Follow up within the hour, by text, while the car's still on their mind and still on your lift.
• Offer the financing or the split (do the brakes now, the struts next visit) before they ask, so the answer is yes to something instead of no to everything.

The Front Desk agent owns this. It sends the estimate the moment the writer builds it, follows up within the hour by text, answers the customer's questions about the work, and routes a yes straight back to the bay. The Analyst shows you close rate by writer, so you can see who needs coaching and who's quietly carrying the floor.

3. Missed-call rate: the leak you can't see from inside the building

About one in four calls to an independent shop goes unanswered during open business hours, and it's worse after you close. This is the most invisible number on the list because you're inside the building when it happens. The writer is elbow-deep with a customer at the counter, the phone rings, nobody can grab it, and you never see the line that didn't connect.

It decides profit because a missed call is a missed customer, not a missed message. About 85 percent of voicemail callers never call back, and roughly two-thirds simply dial the next shop on their list. The car still gets fixed. Just not by you. A busy shop can bleed around $11,250 a month to missed calls, and even a smaller one loses about $2,400. That figure isn't a guess. It's miss rate times close rate times average repair order, the three numbers we just walked through, stacked on top of each other.

The cheap lever: nobody should ever reach your voicemail during a buying decision. You don't need to hire a front-desk person or pay a $3,000-a-month answering service that takes a message and hangs up. You need every call answered and booked, including the 7 p.m. call from the guy whose check-engine light just came on in a parking lot. That's the single highest-return fix in this entire post.

The Front Desk agent answers every call you'd otherwise miss, books the bay, and does it after hours and on weekends. It's the agent that pays for the whole platform on its own. If you want the missed-call math run on your real numbers, that's exactly what the Service-Drive Audit does first. The full set of these benchmarks lives on our auto repair shop statistics page.

4. Technician efficiency: are you paying for hours you can't bill

Technician efficiency is billed hours produced against hours actually worked. A flat-rate-competent tech should clear 100 percent, and the strong ones run well above it, turning eight clock hours into ten or eleven billed. The trap is paying a tech for a full day and only billing six hours of it. Those two unbilled hours are pure loss, every day, per tech.

This decides profit because labor is your highest-margin product and your second-biggest cost at the same time. A tech you pay for 40 hours who only produces 30 billed is running at 75 percent, and that 25 percent gap comes straight out of your margin whether the floor is slammed or dead. Multiply one slow tech across a five-bay shop and you're funding a phantom employee.

The cheap lever isn't pushing techs to rush, which buys you comebacks (see number six). It's killing the dead time between jobs: the walk to the parts counter, the wait on a part that wasn't pre-pulled, the hunt for a diagnostic procedure across four databases, the standing around for the next dispatch. Most lost efficiency isn't slow wrenching. It's a tech standing still, waiting on the shop.

The Analyst surfaces efficiency by tech and flags the dead time, so you're coaching the bottleneck instead of guessing. The Diagnostician cuts the lookup time by putting the whole professional diagnostic canon in the tech's hand on the bay floor, instead of sending them to a desk to dig.

5. Bay utilization: empty bays don't send you a bill, but they cost like one

Bay utilization is the share of your available bay-hours that are actually producing billable work. Your rent, equipment lease, insurance, lighting, and most of your overhead are fixed. They cost the same whether a bay's turning a brake job or sitting empty. A bay parked for two hours on a Tuesday doesn't send you an invoice, which is exactly why it's easy to ignore and expensive to keep ignoring.

It decides profit because fixed costs only get cheaper per car when you push more cars through the same bays. The shop next door with identical rent and one more turn per bay per day is quietly out-earning you on the same footprint. Empty bays at 8 a.m. and 4 p.m. while you're slammed at noon is a scheduling problem wearing a demand-problem costume.

The cheap lever is smarter scheduling, not more cars. Most shops cluster appointments midday and starve the open and the close. Book the after-hours and weekend callers into your slow first hour. Stage parts the night before so the bay isn't waiting on the counter. Spread the load and you find capacity you already owned.

The Front Desk agent captures the after-hours demand and books it into your soft slots, smoothing the day. The Parts Desk agent stages the parts ahead, so a booked bay actually moves at its appointment time instead of idling while someone runs to the supplier.

6. Comeback rate: the job you do twice and bill once

A comeback is a car that returns for the same complaint you just fixed. There's no clean industry-wide benchmark, but the rule on the floor is simple: lower is the only acceptable direction, and every comeback is a job you performed twice and got paid for once. It's the most expensive number that never shows up as a line item, because the cost hides inside your labor.

It decides profit on three fronts at once. You eat the redo labor, you eat the part if it wasn't the part's fault, and you spend a customer's trust you'll never fully earn back. A misdiagnosed P0420 that gets a fresh catalytic converter when the real culprit was an upstream sensor is a comeback, an angry customer, and a refund conversation rolled into one. It also drags down your tech efficiency from number four, because the redo eats hours you can never bill.

The cheap lever is diagnostic accuracy on the first pass, and it's cheaper than you'd think because the answer already exists. The professional canon holds millions of technician-confirmed fixes. The Identifix database alone carries more than 3 million, but the knowledge is split across ALLDATA, Mitchell 1, Identifix, and MOTOR. When a junior tech guesses instead of checking the confirmed fix for that exact symptom on that exact platform, you get a comeback. When they check first, you don't.

The Diagnostician closes this gap. It unifies the whole canon into one search and puts the confirmed fix for the symptom in the tech's hand before they touch a tool, so the junior tech diagnoses more like the veteran. There's a deeper walkthrough in our guide to cutting diagnostic time.

7. Parts gross margin: the profit hiding on the counter

Parts gross margin is what you keep on the parts you sell after what you paid for them. It's half of most repair orders, and it leaks in two quiet ways: parts sold at or near cost because nobody set a real matrix, and supplier pricing nobody has renegotiated since you opened. On a brake job where the customer sees one bottom-line number, a few points of parts margin is invisible to them and decisive to you.

This decides profit because parts margin is leverage you control without selling a single extra job. Tighten your parts matrix by five points across every ticket and the gain compounds across every car, every day, with no added labor and no added car count. It's the most passive margin improvement on this list once it's set, and most shops have never sat down and set it.

While you're auditing what you keep, audit what you spend. A point-tool software stack, one app for scheduling, another for inventory, another for reviews, often runs $1,100 to $1,500 a month, and the real cost lands roughly 40 to 60 percent over the sticker once the add-ons, per-seat fees, and integrations are counted. That overspend is a margin leak too, just on the cost side of the ledger.

The Parts Desk agent watches your parts margin by job and by supplier, flags the lines bleeding profit, and sees the shortage before it forces an emergency order at a premium price. The Analyst rolls parts margin into the morning briefing, so it stops being the number you check once a year at tax time.

The seven, on one page

Each of these is self-contained on purpose, but they compound. A missed call you never answer can't become an estimate, so it can't close, so it never lifts your ARO, so the bay it would have filled sits empty. Fix the front of that chain and the whole thing moves together.

1. 01Average repair order ~$428 — the multiplier on everything; move it with declined-work follow-through. (Analyst)
2. 02Estimate close rate ~70% — recover work you already quoted; move it with same-hour photo estimates. (Front Desk)
3. 03Missed-call rate ~25% in hours — the invisible leak; answer every call, after hours included. (Front Desk)
4. 04Technician efficiency 100%+ — stop paying for hours you can't bill; kill the dead time between jobs. (Analyst)
5. 05Bay utilization ~85% — fixed costs don't care if a bay is empty; smooth the schedule. (Front Desk)
6. 06Comeback rate (lower only) — the job you do twice; move it with first-pass diagnostic accuracy. (Diagnostician)
7. 07Parts gross margin — profit on the counter; move it with a real parts matrix. (Parts Desk)

None of this requires ripping out the software you run today. Auto Advisor sits on top of Tekmetric, Shopmonkey, Shop-Ware, AutoLeap, and Mitchell 1 on the shop side, and DealerCenter, Frazer, DealerSocket IDMS, Wayne Reaves, and AutoManager on the dealer side. Nothing gets pulled out. The agents read what's already there and act on it.

A crew, not a tool, with a human always in the loop

Auto Advisor is a 15-agent AI crew spread across the whole operation, from the sales floor to the service bays. Every agent runs in one of three modes you control: Off, Approve (it drafts, you confirm), or Auto (it acts inside the rails you set). The point isn't to replace your service writer or your parts manager. It's to make the team you already have run about 40 percent more efficient, because the boring, lossy work gets handled and your people do the work only people can do.

That human-in-the-loop principle isn't a slogan here. Founder Cory Salisbury spent his engineering career at Tesla, SpaceX, and Rivian, where autonomous systems have to be safe, have to cite their work, and have to keep a person in command. Same rules apply on your service drive. An agent that books a bay or orders a part shows you why, and the modes mean it never does anything you didn't authorize.

If you run a franchise store or a used-car lot instead of a repair shop, the same seven numbers and the same crew apply across your sales floor and service lane, on top of your existing DMS. Our guide to AI for car dealerships walks the dealer-specific version, and if you're weighing whether to bolt this on or rip and replace, start with replace or add AI to your shop software.