A repair shop advertises next to the largest DIY library on the internet. For every driver looking for somebody to do the brakes, several more are pricing the pads, watching a stranger fit them on video, or hunting a used caliper two states away. Every one of them can click your ad, and the account records all of it as traffic.
Google Ads works for an auto repair shop when the account is split by the service you sell, restricted to the vehicles and jobs you actually take, and judged on repair orders rather than clicks. The rest is upkeep: finding the searches that were never going to end up on a lift, and doing it often enough that they do not pile up.
This guide is that upkeep, run from a chat window. If your assistant is connected to the account, the reading takes one message and your time goes on the decisions instead.
Why does a repair shop lose more to bad clicks than other trades?
Three things are true here that are not true of most local services.
- The parts trade shares your vocabulary. Pads, rotors, alternator, timing belt, catalytic converter. Retailers and marketplaces bid on the same words, and the person typing them wants a box delivered, not a bay booked.
- Repair is the most-taught trade on video. How to, replace, reset, torque spec, wiring diagram. That traffic is enormous, it is cheap enough for broad match to keep buying it, and it never calls.
- The vehicle decides whether you are even eligible. Hybrid batteries, diesel, European makes, motorcycles, heavy trucks, transmission rebuilds you sublet. A shop that does not touch the job still pays for the search.
Put those together and a repair account can spend a third of its month on people who were never customers, while the reports still look busy.
How should an auto repair shop structure its campaigns?
The rule is the same one that governs every account. If you would be unhappy to see the budget move from one service to another on its own, they are separate campaigns.
- One campaign per service line whose money you want to control. Brakes, diagnostics and check-engine, maintenance and oil, and whatever your third earner is — suspension, AC, transmission. Smart Bidding will otherwise pour the budget into oil changes, because the cheap job converts most often.
- Not one campaign per make. Make and model belong in ad groups and in the negative list, not in campaign names. A Honda brake job and a BMW brake job are the same bay at a different labour rate.
- Give the towed-in emergency its own ad group. Will not start, overheating, broke down, towed. Different urgency, different ad, and the shop that answers first gets it.
- Schedule the ads around the service counter. A call at 7pm that nobody picks up is a click you bought and a car somebody else is fixing tomorrow morning.
- Keep the jobs you do not sell out of the keyword list entirely. Body work, paint, detailing, tires if you do not mount them, key programming if you send it to the dealer. To a searcher those are separate businesses.
Which searches waste a repair shop's budget?
Five kinds: parts and prices, DIY and how-to searches, vehicles and jobs the shop does not take, dealer and warranty questions, and people looking for a job or a salvage yard. Together they are the bulk of what a repair account spends on people who will never book.
Here is the shape of it on a small independent account over 30 days. The figures are illustrative, not a customer's results.
| Search term | Spent | Conversions |
|---|---|---|
| brake pads price | $41 | 0 |
| how to change an alternator | $23 | 0 |
| used auto parts near me | $16 | 0 |
| mechanic jobs near me | $14 | 0 |
Five categories cover most of it:
- Parts and prices. Pads, rotors, filter, battery, OEM, aftermarket, and any part name followed by price or cost. The single largest leak in this trade, because your service words and the parts catalogue are the same words.
- DIY and how-to. How to, replace, fix, diagram, torque spec, reset, forum, video. These people are reading, not booking.
- Vehicles and jobs you do not take. EV, hybrid battery, diesel, semi, motorcycle, classic restoration, engine swap, rebuild — whichever of those is not something you would accept on Monday.
- Dealer, warranty and recall. Recall, warranty, TSB, software update, dealer service, key programming. The answer is the dealership and the searcher usually knows it.
- Jobs, training and salvage. Mechanic jobs, apprenticeship, ASE certification, salvage, junkyard, part out, scrap.
Add the business-side ones while you are there: shop for sale, auto repair software, franchise, insurance. The free Google Ads waste audit runs this filter for you and returns the queries, the cost behind each one, and the negatives it proposes.
The general version of that filter is the spend-over-threshold check, and picking the match type and the level for each negative is its own decision. The same drill reads differently in another trade: in a cleaning business the biggest leak is job seekers, here it is a parts catalogue.
Find the wasted spend from the chat
With account access the whole review is one message. Fill in the brackets with your own services, your own area and the vehicles you turn away.
Review the last 30 days of Google Ads search terms for my auto repair shop. We offer [brakes, diagnostics, suspension and scheduled maintenance] in [service area]. We do not offer [body work, paint, tires or EV service], and we do not work on [heavy trucks or motorcycles]. Separate the searches into clearly irrelevant and uncertain, and for each one show the spend, the clicks and the recorded conversions. Flag anything that looks like a parts purchase, a DIY repair or a dealer-only job, and say which of the three it is. Recommend negative keywords with match types, say whether each belongs at campaign or ad group level, and explain what each one risks blocking. Tell me which service campaign each wasted search came from. Do not change anything.
Four clauses are doing the work. Naming the vehicles you turn away is what makes the report yours rather than generic, because nothing in the account says a diesel is somebody else's problem. Asking for parts, DIY and dealer as labels forces the reasoning into the open instead of a bare verdict. Splitting irrelevant from uncertain stops a confident model from blocking queries that were merely unproven. And do not change anything keeps the message a read: nothing reaches the account until you approve it. That is one prompt for one job; the 37-prompt ChatGPT library covers the campaign, copy and budget work around it.
Approve the exclusions one at a time
What comes back should be a decision list, not a data dump.
| Search term | Spent | Conversions | Verdict |
|---|---|---|---|
| brake pads price | $41 | 0 | Irrelevant — parts purchase, phrase negative |
| how to change an alternator | $23 | 0 | Irrelevant — DIY intent, phrase negative |
| transmission rebuild cost | $38 | 0 | Uncertain — you sublet rebuilds, your call |
| brake fluid change cost | $29 | 1 | Keep — priced the job and called |
The uncertain row is the one worth slowing down for. A broad negative on transmission clears the rebuild you do not want and the fluid service you do, and you will not notice for a month because nothing breaks — the impressions just stop. That is why the match type and the level are a separate decision from the exclusion itself, and why the assistant proposes rather than applies.
Is the tow radius the same as the drive radius?
No, and an account that treats them as one number overspends on the edges.
- A drivable car travels; a dead one does not. Somebody who needs a tow picks whoever is nearest or whoever covers the truck, so the outer ring of your radius rarely converts for emergency work and sometimes converts well for booked maintenance.
- Read the spend by location, not the radius setting. Ask the assistant for cost, clicks and conversions by location for the same 30 days, and the suburbs that bought nothing show up immediately.
- Set the campaign to *Presence: people in or regularly in your targeted locations*. Otherwise you pay for somebody researching your town from another state.
Count repair orders, not clicks
None of this means anything if the account cannot see an outcome, and a repair shop books by phone more than by form.
- Track calls from the ad, calls from the website and form submissions as three separate conversions, so you can tell which one the money is buying.
- Count one meaningful action per enquiry. Counting every page view teaches Smart Bidding to buy browsing.
- Get the booked repair order back from your shop management system where you can. Until then, cost per call is the ceiling of what anyone — you or the assistant — can honestly judge.
The half the assistant cannot decide
An assistant reads what the account records. It does not know which of those calls became a repair order, what the average ticket was, or that you have one technician out this week and no free lift until Thursday. So the uncertain list is yours, and the useful thing you can do is hand over the context it is missing:
- Which work you want more of right now, and which bay is already full.
- The jobs you turn away, and the ones you only take when the schedule is light.
- What a booked repair order is worth on average, so a cost per call can be judged against something.
Say those out loud in the same conversation and the next review is better, because the model stops guessing at the economics of a shop and starts using yours.
Make it a weekly check
Loose match types keep serving new queries, so the list refills. In ChatGPT you can save the review as a scheduled Task and it runs on its own; Claude has no scheduler, so treat the cadence as a habit and connect it the same way — the prompt above works unchanged in either one.
The weekly review folds this check in with five others and still fits in ten minutes. The same report has a profitable half worth reading: the searches that converted without ever being keywords, which the keyword research walkthrough promotes into the account. And when you want the full sweep rather than the search terms alone, the 30-minute audit covers the other six places money leaks. And when the budget question in your trade is seasonal rather than a filter problem, the landscaping walkthrough plans the quarter around the crew instead.
