Car wash analytics: measuring bay occupancy and wash duration with cameras

The questions every car wash operator asks

A car wash manager knows the day’s takings, but rarely the operational reality behind them. How many vehicles turned up between 5 and 7 pm? How many had to wait, and how many drove off when they saw the queue? Is one bay sitting idle while the others are overloaded? Does a wash actually take as long as it should, or does it drag on at peak times?

These questions drive real decisions: adding staff on Saturday mornings, changing how vehicles are directed to the bays, tailoring an off-peak offer or making the case for an extra gantry. Without measurement, they are settled by gut feeling.

Many sites already have cameras, usually installed to protect the payment area and vacuum stations. They can answer part of these questions, provided they are well placed and analysed with the right approach.

Vehicle counts, bay occupancy and cycle times

Three families of video-derived indicators are usually useful at a car wash:

• Vehicle counting: a virtual line at the site entrance, or at each bay entrance, counts vehicles per hour, per day and per bay. It is the foundation for understanding real traffic and when it peaks.
• Bay occupancy: a zone drawn over each bay shows whether it is occupied or free, and for how long. The result is an occupancy rate over time that highlights under-used bays and saturated time slots.
• Cycle duration: measuring the time between a vehicle entering a bay and leaving it gives an estimate of wash duration. Compared with the expected duration of each programme, it can reveal bottlenecks, without being used to assess individual staff.
• Upstream waiting: a zone on the access lane estimates how many vehicles are queuing and roughly how long they wait before reaching a bay.

Constraints specific to the site

A car wash is one of the toughest environments for video analytics. The difficulties are well known and should be dealt with during the audit, not discovered during deployment:

• water, steam and foam, which temporarily hide vehicles and can disrupt detection inside the bays;
• spray on the lens, which calls for a suitable housing, a sheltered position and a regular cleaning routine;
• night-time lighting, often uneven, with harsh contrast between lit and dark areas, plus vehicle headlights;
• glare during the day, especially for open bays facing the setting sun;
• humidity, heat and chemicals, which shorten equipment life and shape the choice of cameras and housings.

Getting camera placement right

In many cases it is more reliable to measure at the bay entrances and exits than in the middle of the wash. A camera mounted high, out of the spray zone, that sees the access lane and the bay entrances often delivers more usable data than one lost in a cloud of steam.

That is exactly what the audit is for: deciding whether the existing cameras will do, whether reframing is enough or whether a dedicated camera is needed. It also checks stream access (RTSP, ONVIF, recorder), night-time image quality and whether an edge AI unit can be installed in a dry, ventilated technical room, so footage is processed on site and only indicators reach the dashboard.

On the data protection side, the goal is to count vehicles and measure durations, not to identify people. Privacy masks exclude the public road or staff break areas. As data controller under Law 09-08, the operator must inform customers and staff and complete any applicable formalities with the CNDP.

Cross-checking with the payment terminal or dosing pumps

Video measures what happens physically: vehicles arriving, waiting, occupying a bay and leaving. It cannot tell which programme was bought or how much product was used. That information lives in other systems.

Comparing the number of vehicles detected with the number of payment terminal transactions, for example, can surface gaps worth investigating: vehicles leaving without paying, drive-offs from the queue, transactions with no matching visit. Likewise, matching observed cycles with dosing pump data can help track chemical consumption per wash.

These comparisons require an integration with the payment terminal, wash controller or pumps, which depends on their export or API capabilities and calls for a dedicated study. On the dashboard, video-derived indicators and data from those systems must stay clearly separated: a gap is not proof of fraud, it is a question to look into.

Running a pilot on one site

Whether you operate one car wash or a network, the safest approach is to start with a pilot on a representative site. It should cover weekdays and weekends, sunny days and, ideally, evenings, so the difficult conditions are properly tested. A typical sequence:

• set two or three precise questions, for example: at what times does waiting exceed an acceptable level, and which bay is used least?
• audit the cameras, adjust framing or add a protected camera where needed;
• validate counts and durations with manual spot checks, by day and by night;
• track a handful of indicators: vehicles per hour, occupancy per bay, average cycle time, estimated wait;
• decide, based on observed results, whether to extend, adjust or stop.

The bottom line

Measuring bay occupancy and wash duration gives operators an objective view of their business, as long as they account for a harsh environment, put cameras where they can see clearly and keep what video measures separate from what the payment systems report.

Curious about what your cameras could measure on your site? Have a look at our car wash page, then request an audit: it is the starting point for a well-scoped pilot.