Quick lube camera analytics runs AI on your bay cameras to log when each car pulls in, when work starts and stops, and when it leaves. Those events become car counts, service times, bay utilization and a daily check against POS tickets. It usually runs on cameras you already have, if they see the bay door and the hood. Accuracy depends on angles, lighting and clock sync, so prove it against a hand count at one store first.
Quick answer
- What it measures: car in, hood up, hood down, car out, plus queue and drive-offs if a camera covers the lot.
- What it needs: one fixed camera per bay that sees the door and the hood, stream access from your recorder, and clocks that agree.
- How to start: one store, about four weeks, with pass marks you set before you begin.
What vendor pages tell you, and what they leave out
Leverege's press release about PitCrew's NOLN feature runs about 160 words. It says PitCrew times services, tracks vehicle and technician activity, flags cars left unattended, and uses "edge-based processing" to keep video local (Leverege). Leverege's oil change page adds that it counts arrivals and drive-offs, works with existing cameras and connects to your POS (Leverege).
An operator's view helps too. Scott Hempy of Oilstop wrote that bay timers in his service centers cost under $50 a month per center, and that Oilstop had started testing cameras with computer vision in place of analog sensors (NOLN).
None of them explain camera placement, accuracy, the ticket match, what to tell staff, or how to run a fair test. That's the rest of this post.
What camera analytics measures in an oil change bay
| Event | What the camera must see | What you use it for |
|---|---|---|
| Car in / car out | The bay threshold and the whole car | Car count, total bay time |
| Hood up / hood down | The hood, from above or the side | Work start and finish |
| Tech on the upper floor | The space around the engine bay | Unattended cars, labour per car |
| Tech in the pit | A separate camera in the pit | Lower-bay work, safety |
| Car waiting in the lane | The approach lane or lot | Queue length, wait time |
| Car leaves the lane unserved | Lane camera plus the exit | Drive-offs |
| Plate read | The plate, straight on, at low speed | Matching visits to tickets, repeat visits |
| Activity outside posted hours | Any bay camera | After-hours visits |
From those events you get the service-time numbers most owners care about:
- Wait time: car joins the lane to car in.
- Time to first hood: car in to hood up. Long gaps often point to staffing.
- Hood time: hood up to hood down. This is closest to actual work time.
- Total bay time: car in to car out.
- Bay idle time: car out to the next car in, while cars are waiting.
Ticket timestamps can stand in for some of these, but only if advisors open tickets when the car pulls in.
Camera requirements and placement
You rarely need special cameras. You need the right view. Walk each bay with this list before you sign anything.
| Check | Why it matters |
|---|---|
| One fixed camera per bay sees the door and the full car | Car in and car out are the base of every metric |
| The hood is visible and not blocked by the tech standing at it | Hood up and hood down mark the work |
| The camera is mounted high and out of reach | A bumped camera quietly breaks counts |
| Low sun through open doors and headlights don't wash out the view | Glare causes missed and double events |
| Night mode (infrared) switch doesn't happen mid-service | Image changes at dusk confuse detection |
| Your recorder (NVR) gives stream access, usually RTSP or ONVIF | The software needs a feed to read |
| The store's internet can carry metrics or video to the cloud, if used | Some systems process on a box in the store, some in the cloud |
| The recorder and the POS use the same time source | A few minutes of drift breaks the ticket match |
A few placement habits help. Look down the length of the bay from the front or rear, not across it. Give the pit its own camera if you want lower-bay data, because an upper camera can't see under the car. Put a separate camera on the approach lane for wait times and drive-offs.
Accuracy: where it holds and where it breaks
None of the vendor pages above publish accuracy figures, and a number from another chain wouldn't carry to your stores anyway. Accuracy depends on your angles, your light and your process. These problems show up most:
| Problem | What it looks like in the data | Fix |
|---|---|---|
| Car repositioned in the bay | Two short visits instead of one | Merge visits with a gap of a minute or two |
| Tech blocks the hood | Hood time cut short or split | Raise or move the camera |
| Glare from open doors | Missed car in or car out at certain hours | Adjust angle, add a shade, or change exposure |
| Camera bumped or moved | A bay's counts drop to near zero | Daily camera health check and a snapshot alert |
| Recorder offline | Gaps with no events at all | Alert on missing feeds, not just on events |
| Clocks out of sync | Tickets match the wrong car | Sync recorder and POS to the same time server |
| Dirty or angled plates | Plate reads missing or wrong | Treat time-window matches as the default and plates as a bonus |
The honest test is a hand count. Pick two busy periods, tally cars in and out on a clipboard, and time a sample of services with a stopwatch. Ask any vendor to show you that comparison at your own store.
How it ties to POS tickets
The camera shows what happened in the bay. The POS shows what was sold. The match works on bay and time window first, then plate or VIN where you have it. For that, your POS export needs open time, close time, bay if it records one, plate or VIN, services, total, and every $0 ticket and void.
Once the match runs daily you get:
- Visits with no ticket: work done that wasn't rung up, or no-charge work with no $0 ticket.
- Tickets with no visit: a camera gap, a ticket on the wrong bay, or work that never happened.
- Service time by ticket type: how long a full synthetic with a filter really takes compared with a basic change.
- Revenue per bay-minute: ticket revenue divided by minutes with a car in the bay.
We walk through the match, the false positives and who should review flags in our guide to off-book oil changes.
Privacy and your staff
Keep it overt. The Office of the Privacy Commissioner of Canada calls covert video surveillance highly privacy-invasive and says it "must be considered only in the most limited cases" (OPC). Visible cameras, with staff told what's measured, are far easier to defend.
Put it in writing. In Ontario, employers with 25 or more employees on January 1 must have a written electronic monitoring policy in place before March 1 that says whether, how and why employees are monitored. Existing employees must get a copy within 30 calendar days (Ontario ESA guide). Update it before the pilot starts. Other provinces and states have their own rules, so check with your employment lawyer.
Report on bays first, people second. Store and shift numbers are for managers. Individual times are for coaching, with the clip in hand and a chance for the tech to explain.
Set retention and access. Decide how long clips tied to flags are kept, who can view them, and who can search plate reads. Treat plates as customer data. Leave camera audio off unless your lawyer has signed off.
What drives the cost
Vendors rarely publish prices, so compare quotes by what's driving them:
| Cost driver | Why it moves the price |
|---|---|
| Existing cameras usable as-is | New or moved cameras add hardware and install time |
| Processing on a box in the store or in the cloud | Edge boxes add hardware; cloud adds bandwidth |
| Number of bays and events tracked | More views, more tuning |
| POS connection | A daily export is simpler than a live connection |
| Custom reports and alerts | Your own scorecard takes build time |
| Review time | Someone on your side has to look at flags each week |
| Clip storage and retention | Longer retention costs more storage |
For each quote, ask for setup, monthly cost per store, hardware, contract length, and what happens to your data if you leave. If you're weighing a packaged product such as Leverege's PitCrew against a custom build, our PitCrew comparison lays out where each one fits, and the bay camera payback calculator puts your own numbers against a quote.
How to pilot it at one store
Four weeks at one store is enough to see whether the numbers hold. Write your pass marks down first, so nobody argues the result after.
- Week 0, setup. Run the camera check. Sync clocks. Pick three metrics, such as car count, hood time and unmatched visits. Update your monitoring policy and tell the crew.
- Week 1, run quietly. Collect events without sharing them yet. Do two hand counts at busy times.
- Week 2, tune. Fix angles, merge rules and the time window. Start the daily POS match.
- Weeks 3 and 4, use it. Send the store manager a daily summary. Review the unmatched list weekly with someone outside the store. At the end, check results against your pass marks.
Keep pass marks simple: camera count within your chosen tolerance of the hand count, an unmatched list you can review in under an hour a week, and a manager who changed at least one thing because of the report.
Worked example (made-up numbers, not real store data)
One store, three bays, Saturday 9 a.m. to noon.
| Line | Example value |
|---|---|
| Cars counted by hand | 41 |
| Cars counted by camera, first run | 44 |
| Extra visits that were repositions | 3 |
| Camera count after a merge rule | 41 |
| Hood time, stopwatch average on 10 cars | 13.5 min |
| Hood time, camera average on the same 10 cars | 14.0 min |
The count matches after one rule change, and hood time is close enough for store comparisons. If one bay alone had missed cars, you'd fix that bay's angle first.
Once the numbers hold, feed bay time and utilization into your weekly scorecard. Our quick lube KPIs guide has the formulas and a template.
Where Trilix fits
We build camera analytics for quick lube operators on the cameras most stores already have. It logs visits per bay, times services, matches visits to POS tickets and sends daily reports. We check your angles first and tell you what each store can measure before we quote. To talk through your setup, get in touch.
Want to see this on your stores?
We’ll look at your calls, bays and tickets with you and show what we’d build first.