By Tom Gresham
When Hoffman Car Wash, which has approximately 30 locations in the state of New York, began to see more pileups inside its tunnels, company leaders knew they needed to identify a solution to reduce the chance of collisions. Fixes such as slowing production or placing staff members in the tunnel to monitor the process were imperfect, harming efficiency and morale.
So, in late 2016, Hoffman Car Wash installed software to help prevent pileups at its Queensbury location, one of the company’s busiest sites. The solution uses machine learning to prevent accidents in car wash tunnels. The software combines site data and live video of vehicles in the tunnel to detect potential accidents. If any potential collision-causing problems are processed, such as a vehicle unexpectedly stopping, the software halts the conveyor transporting vehicles.
The solution proved instantly successful, and Hoffman chose to expand its use of the technology, called NoPileups, at additional locations. AJ Davison, Director of Information Technology for Hoffman Car Wash, said it was simple to adopt for the company and helped operations run more efficiently and safely. Machine learning proved to be a valuable solution when a human-based solution couldn’t offer the same advantages.
“The key is it gives us visibility into potential issues where we often don’t have people,” Davison said. “Since we don’t have people in the tunnel, and we can’t control what customers do down the conveyor, we had a hole in our anti-collision system. [This software] fills that void to prevent collisions in the tunnel.”
Davison said Hoffman Car Wash wasn’t wary of adopting machine learning – in fact, company leaders embraced the opportunity to be on the vanguard in their industry.
“Our CEO, Tom Hoffman Jr., always had a vision that a camera-based anti-collision system would be huge in the car wash industry,” Davison said. “He’s long suggested that this technology should be developed.”
Machine learning, a form of artificial intelligence, provides the capability to extract useful patterns from data and make decisions (or suggest decisions) based on those patterns with minimal human involvement. Typically, machine learning is used to focus on optimizing a particular process or dimension of the business, such as understanding and predicting customer behavior. Machine learning technology is starting to gain an initial foothold in the car wash industry, though Alan Nawoj, Founder and CEO of Beacon Mobile, said machine learning technology and artificial intelligence are still “very new” to the industry and not yet widely used.
“However, some bigger brands and software companies that serve the car wash industry and collect large amounts of data are starting to think about how to apply these technologies to improve their operations,” Nawoj said.
Although machine learning in car washes is in its infancy, Nawoj said he foresees an increased application of the tool in every aspect of car wash businesses – from marketing and sales to equipment and chemicals. Beacon Mobile is developing a machine learning software platform called Carwash.AI that is designed to improve the customer experience at car washes, “laying the foundation for a new way customers will interact with their local car wash,” Nawoj said.
“There are always ways to improve and optimize a business and many clues can be found by identifying patterns in data,” Nawoj said. “This is where machine learning and AI comes into play. These technologies are almost like keys that can unlock greater efficiencies in a business, and by making a car wash more efficient, everyone can benefit… from the operator down to the customer.”
Anoop Kanthan, Co-founder and COO of omniX Labs, which has a machine learning platform used in a variety of industries, said machine learning carries the potential to give car wash operators new insights and fuel more sophisticated decision-making.
“Imagine having the best manager running each of your locations every day to the highest of standards and alerting you to any issues in real time and taking steps to proactively manage the situation,” Kanthan said. “Imagine being able to predict – within say a few percentages – how your business will run tomorrow: How much volume, how many pay-as-you-go versus unlimited and even how many you will convert? Conversely, also what’s likely then to be your costs for that day? Imagine what you could do with that information.”
Nawoj said machine learning and AI have the potential to help car wash operators boost revenue, streamline operations and improve the customer experience. Machine learning tools can track information and activity at a car wash such as the identity of vehicles (by reading license plates or using RFID tags), the length of queues and the length of stays, among other details, and then evaluate the data that has been collected. The more data an operation gathers, the more insight gathered for the machine learning analysis. With that data, machine learning solutions can help operators better understand the intricacies of their operation, leading to recommendations and new clarity about opportunities and inefficiencies.
“You can start to get a really rich picture of what’s going on,” Kanthan said.
Machine learning offers the capability to allow operators to track the behavior of specific customers, including repeat customers, and to receive guidance into how to sell to them. Kanthan said machine learning provides the detailed insight into customers that operators crave.
“[Operators] really want to know who their customers are, especially the pay-as-you-go ones as opposed to the ones that are subscription members, and they really want to convert them in targeted way,” Kanthan said. “Right now, they don’t generally have the capability to do that … Obviously, there’s a huge revenue opportunity for them once them once they understand that customer and they can try to sell them a better package.”
By tracking customer purchase history and car wash usage, Nawoj said machine learning can deliver personalized offers to each customer using data from previous marketing offers.
“When a customer feels like you are talking directly to them and giving them an offer that is tailor-made just for them, they are more likely to act on it,” Nawoj said. “There is a huge opportunity for machine learning to deliver a more personalized and optimized user experience for all car wash customers, which will in turn help drive more revenue to the operator.”
Machine learning analysis of the available data also can lead to anticipating future conditions, helping car wash managers prepare to operate their facilities more efficiently, such as with scheduling staffing, Kanthan said. Nawoj said machine learning and AI algorithms can help operators be proactive with equipment management.
“By tracking data on equipment parameters (e.g. vibration frequencies, average time until equipment failure, etc.), it is possible to apply machine learning and AI algorithms to help wash operators be more proactive with their equipment maintenance and upgrades rather than reactive,” Nawoj said. “For example, by looking at patterns in a large data set, machine learning could alert an operator when they may want to order a replacement part before that part even fails based on past history. This application of predictive analytics would help to reduce or even eliminate wash downtime when the part does actually fail, therefore sparing both operator and customer frustration.”
Nawoj said machine learning requires access to a lot of data to work properly, meaning a car wash operation needs to be able to “collect meaningful data in an organized way” while also identifying the problem the data will be used to solve.
“For example, if you’re trying to use time-boxed coupons to smooth out traffic volumes at your car wash, you will need to start by collecting detailed traffic data for your site as well as coupon redemption statistics so you can see what works and what doesn’t work,” Nawoj said. “You would then design a machine learning algorithm that would take this data as its input.”
Still, the complications of machine learning likely won’t fall to operators, Nawoj said.
“Machine learning is not necessarily something that a car wash operator should have to worry about directly,” Nawoj said. “Rather, car wash operators should identify vendors who specialize in this area and can provide solutions that analyze their data and provide insights that can help them to boost their revenue, streamline their operations, and improve their customer experience.”
Kanthan recommends that car wash operators try “to walk before they run” when considering machine learning’s possibilities.
“I think start with the basics and evolve as you are comfortable,” Kanthan said. “Get the basic view on your operations. You will start to see things about your business you didn’t probably realize or had a hunch but couldn’t prove it. Get some of your trusted managers to start to use it for day-to-day needs like looking up the last time a particular vehicle was at your facility to help with a claim or seeing how the days’ counts compared to your [point-of-sale] transactions. Then start to adopt some of the advanced capabilities like the ability to target repeat customers that are not yet unlimited members or use the predicted volumes for tomorrow to better manage your staffing or figure out how to make it better.”
In addition to NoPileups, Davison said Hoffman Car Wash has employed other machine learning tools to continue to pursue sophisticated solutions that can strengthen their operation.
“We utilize sonar sensors to detect a pick-up truck in our wash tunnels to adjust product functions, saving on chemicals, and turn off blowers pointing directly into the bed of a pickup,” Davison said. “We also monitor amperage of our spinning brushes to stop the conveyor if we detect the cloth is caught on something to prevent damages.”
Davison doesn’t expect machine learning’s role will stop there for Hoffman Car Wash and other car wash operations like it.
“I think, like any industry, machine learning will be a huge part of the future,” he said.