The cars on the road are older today than they were 20 years ago. According to the U.S. Department of Transportation, the average age of all cars on the road is a little over 11 years, up from 8.4 years in 1995, which means drivers have a greater incentive than ever to properly maintain and repair their cars.
But Pitstop CEO Shiva Bhardwaj told PYMNTS that most drivers and businesses aren’t maintaining their cars properly, though it’s not for lack of desire. It’s for simple lack of knowledge.
The problem, he said, is that the automotive industry has a bit of a broken feedback loop when it comes to gathering and analyzing vehicle data. It’s a problem he first began noticing early in life — because his father runs service shops and dealerships.
“And I realized that vehicles are being embedded with computers at a faster rate than ever before. But once vehicles get on the road, not one is actually collecting and analyzing all that data coming from vehicles to figure out how to make the product work better.”
This, he noted, was incredibly different from the tech industry where he went on to work as an adult.
Every six months, he noted, a new chip would be released and within two week there would be firmware updates going out to then perfect those chips. It was a feedback loop that made the products more powerful, and when he began to incept the idea for the Pitstop platform, the idea was basically to fix that broken data loop so that the incumbent auto industry could respond as proactively to data as the tech industry does.
“What we do is predict vehicle failures before they happen. It is a prognostics platform that we’ve built.”
Building To The Right Use Case
The standard maintenance model most people rely on is whatever comes recommended for the vehicle. The driver looks in the owner’s manual, looks up the maintenance schedule for their year, make and model and follows it to the best of their ability.
The problem, Bhardwaj said, is that those rules don’t work well in every context — because when it comes cars context is everything.
“We go in believing what we have been told, that this is the right way to do things. No one is analyzing data from the vehicle.”
And that is leading to losses. Bhardwaj said there is $70 billion a year in car maintenance that doesn’t happen though it should, $20 billion in recalls ever year and things like emissions scandals that make drivers and fleet owners feel like they can’t trust dealerships or mechanics.
The method Pitstop is pursuing, he explained, is to train its machine-learning algorithms with information to integrate connected vehicles, dealerships, fleet managers, suppliers, original equipment manufacturers, insurance companies and the aftermarket supply chain. The firm also, he noted, groups information differently. Instead of looking at the car alone, it is really looking at the whole use context for that machine.
“When you start to compare different driving styles you start to realize the trends between someone who is an Uber or Lyft driver operating a passenger vehicle like a fleet vehicle against my mom, who uses a car to go the grocery store,” he said. “You do see a major difference when you start putting use in separate buckets. There is very different maintenance required in those two scenarios.”
The same car, same year and same model — but the different use cases make all the difference. A simple example, he noted, is that ridesharing drivers need to do things like replace their spark plugs with about twice the frequency of typical car owners operating the same car because of how often their automobile is in use.
“It’s because of how often and aggressively they are driving those vehicles. When the car was designed and the recommendations were made, it was thought of as only a passenger vehicle for a family,” he said. If the Uber or Lyft driver doesn’t know that, they can be setting up for bigger, more expensive problems like head gasket failures or oil leaks in the future.
Expanding The Platform
The data picture Pitstop is looking to paint requires a lot of streams, beyond what comes from the car itself. Dealerships and mechanics have insight into what issues they are seeing recur, while suppliers and manufacturers have a full picture of part condition and what defect rates are.
In addition, “Then you have to go and try to understand things like the weather in any individual area and how that plays a role.”
And those are just some of the data streams Pitstop looks to tap into as it teaches its learning engine to properly prognosticate — and every data source is different, as is access somewhat different. They don’t all have APIs that are easily plugged into, Bhardwaj noted, and beyond the technical challenge itself is also the business challenge of convincing some entities that they want to be part of sharing their data.
“We have to really show them the value of bringing all this data together, and that the net data they are going to get from the reports and analysis is greater than them holding on to the data themselves,” he said.
Convincingly selling that, Bhardwaj said, can be challenging because there is a mindset around data that it is as valuable as “gold or oil.” And while that idea is on the right track, he said, it is incomplete — data can be that valuable, but generally isn’t on its own. It’s when it is combined with the right tool set to extract its value that it really lives up to its fullest potential.
“Moving people to that idea takes time, but it is a value that becomes clearer when they see that there’s only potential gain here — and they aren’t giving up anything IP-related to get it. And I think a lot of firms are seeing that they have been dominant leaders in the auto industry, [but now they are] looking at the next 10 years and wondering if the fuel pumps they are making will keep them in that place as a leader — or if that pump is going to become the commodity and their data is the real value.”
What’s Next
The ridesharing space, Bhardwaj said, is changing. There are firms and entrepreneurs purchasing 30 cars at once and renting them out to Uber and Lyft drivers for ridesharing use. And it is in that evolution, he said, that firm sees its largest future opportunity going forward.
“What we are seeing more of is networks of vehicles being leveraged as though they are a fleet, even though they weren’t really designed for that use,” he said. “And we are seeing a question in those firms about how to maintain those passenger fleets, as we are really working to cater to that market.”
Today, he said, that means having insight into vehicles and how they will function that the firm believes their manufacturers don’t even know about yet. And tomorrow, he noted, the firm hopes to be able to push customized software updates for vehicles that will be able to reduce their maintenance impacts going forward.
“What we think will set us apart moving into the future will be updating the vehicle with software updates customized to the driving habit of that vehicle,” he said. “I think that type of future value is really the direction we want to go in.”