Taxi hailing app company Uber is aiming to provide more accurate ‘estimated time of arrival’ to customers in India and other markets by better understanding their traffic conditions, a top executive said.
“We need to make a much better job with ETAs in India and other markets. There is an expectation that we set with ETAs and we want to make sure we live up to that,” said Manik Gupta, head of product, maps at Uber.
For improving accuracy, Gupta said they are using a mixture of purchased mapping data like real-time traffic, imagery and topology, among others, from third party companies and blending them with historical trip data-—like drop-offs and pickup locations— and typical traffic speed, among others.
He said one of the biggest contributors to estimated time of arrival inaccuracy is traffic. “If you don’t have a good understanding of how bad or good traffic is on a certain route, you cannot estimate it better. Hence, there is a lot of work that we are doing to improve how we understand traffic better,” he said.
Although Uber has tie-ups with several partners like Google and Tomtom for mapping content, Gupta, a former Google Maps executive, said they are investing a significant amount of research and resources in their own mapping efforts since their problems are different from that of other mapping providers.
“For example, parking lots matter a lot to Uber, since a lot of pickups and dropoffs happen there. However, there are not that many companies that invest in parking lots since it’s probably not costeffective for them.” Gupta said.
The company is also using this data to solve the rider pickup and dropoff problem. “For instance, we know in a hotel, pickups typically happen in the lobby, so we can request riders to go to the lobby for future rides” Gupta said.