The guidebooks highlight San in Francisco's Hayes Valley neighborhood for its lively bars and restaurants promoted by the removal of an earthquake-damaged expressway and swelling tech industry salaries. Near Uber's headquarters, data scientists working on the company's food delivery service, Uber Eats, look at the scene through a more numerical lens.
Their logs indicate that restaurants in the region need an average of 12 minutes and 36 seconds to prepare. In the evenings Pad Thai orders – that's 3 minutes and 2 seconds quicker than in the mission district to the south. That number may seem dark, but it's at the heart of Uber's bid to build a second giant company to stand alongside its ride-hail service.
Uber fights with other well-funded startups and listed GrubHub in the fast-growing food applications market. In order to gain market share and make the business profitable, some of the future forecasts depend on the preparation time of each pasta dish. If you do something wrong, it means cold food, unfortunate drivers or disloyal customers in a ruthlessly competitive market.
Uber Eats mobile apps and competitors like DoorDash list menu items from local restaurants. When a user places an order, the delivery service passes it on to the restaurant. The service tries to send a driver to arrive as soon as the food is ready, and resorts to a pool of independent contractors, such as the rideshop. In the meantime, the customer is given a prediction until the next minute when his food arrives.
"The more detailed we can model the physical world, the more accurate we can be," says Eric Gu technical manager with the Uber Eats data team. The company employs weather forecasters to predict the impact of rain or snow on orders and delivery times. To refine the predictions is also tracked when drivers sit or stand, drive or walk and join the growing employers who monitor their every move.
Improved accuracy can be converted directly into dollars, for example, with the help of Uber combine the orders so that the driver carries several meals without cold. The drivers get a small bonus when you combine several orders in one trip. "We can save delivery costs and return some savings to the eater," says Gu.
Four blocks away, Uber-rival DoorDash has its own team of data-mates working on an AI-powered crystal ball for food delivery. One of their findings is that the sunset is important. People tend to order dinner at dusk, which means they eat later in the summer and change their habits as the clocks change in the spring and fall. Like Uber, the company keeps track of schedules and weather forecasts and keeps track of the preparation times for the dishes offered in various restaurants. The company's data indicates that Pad Thai is two minutes longer from Friday to Sunday than the rest of the week, as the kitchens are busier.
Rajat Shroff, Vice President of Product, says the DoorDash data also highlights the relationship between accurate delivery predictions and customer loyalty. "That's a big part of our growth," he says. The company was valued at $ 7 billion this month for investors who raised $ 400 million in new funds.
DoorDash has also been working to better understand what's going on in restaurants, for example, by connecting its systems to Chipotle's internal software, which can be sent more smoothly, and DoorDash can track how they are doing. The company has built a grocery delivery simulator that plays past data to test various planning and forecasting algorithms. Both DoorDash and Uber use their data to help motorists spend more money on areas where demand is expected to be strong.
The analytics company Second Measure says credit card data show that DoorDash overtook Uber Eats second place in November behind Uber Eats behind GrubHub. According to January, GrubHub had made 43 percent of its revenue from food, compared to 31
Nevertheless, DoorDash claims that it receives orders from customers within 35 minutes. This is a bit slower than the 31 minutes that Janelle Sallenave, head of Uber Eats for the US and Canada, states that her service to the US is average.
Uber's data researchers may have a big advantage over their competitors: the rich live and historic traffic data from the company's corporate network. The company also delves deeper into its data on restaurants and Uber Eats drivers.
A project involves the analysis of the language of the restaurant menus. The goal is for algorithms to predict the preparation times for dishes that do not yet have good data by dragging data from menu items with similar ingredients and cooking processes.
Chris Muller, Professor at Boston University, says The central view of eating Uber Eats and its competitors is contributing to a big upheaval in the restaurant business. "This is the biggest transformation since we've seen fast-growing chains like Chipotle promise fast-food of higher quality than fast food."
Joe Hargrave, who built a farmer's stall in five Bay Area taco stores, lives through The Food App Transformation He designed his Tacolicious stores for people who share his love of good food that you can eat with your fingers as you watch baseball – now more of his customers eat their tacos, and delivery at home
Orders from apps like DoorDash and Caviar account for about 12 percent of Hargrave's business, and helped drive sales up 8 percent last year, even though the store is about a quarter up He appreciates what the apps do, but it was not easy to account for the supply boom.
"I've spent my entire career figuring out how to present the best product to people, "says Hargrave. "Now I've been thrown with this curveball where I have to put it in a box." In the meantime, every restaurant often has one person working exclusively with packaging and delivery orders.
Muller and Hargrave say the food's app-and-algorithm approach can suppress traditional restaurants and could even bring some things to a standstill. The default Uber cut for each job is 30 percent. This is a significant bite in a traditionally low-margin industry. Even restaurants like Tacolicious, which offer delivery services, must also serve the people who walk through the door.
This is one reason why Uber promotes the development of "virtual restaurants" that emanate from the kitchen of an existing restaurant but sell only through the app. Uber said last year that it worked with more than 800 virtual restaurants in the US. Many work during business hours when a restaurant's home business is loose or closed, allowing for more efficient operation and more efficient use of the property.
Uber and DoorDash also work with so-called dark kitchens, operations that are only served via deployment apps, and much more efficient and predictable than traditional restaurants. DoorDash operates a 2,000 square meter kitchen area in the Bay Area, which it leases to such operators.
Muller compares the arrival of Uber Eats and others with how online travel websites confused the hotel industry and forced hoteliers to tailor their business models for a market where consumers are more involved, resulting in more visits, but at lower prices.
How lucrative this new form of restaurant business will be is unclear. Uber has previously stated that its service is viable in some cities, but the financial data released for the final quarter of 2018 did not contain any details about Uber Eats. Overall, the company lost $ 940 million, 40 percent more than in the previous quarter. In the third quarter of 2018, the company stated Uber Eats accounted for 17 percent or $ 2.1 billion of its gross global bookings.
GrubHub has been consistently profitable since its IPO in 2014, selling $ 1.4 billion worth of food in the final quarter of 2018, an increase of 21 percent over the previous year. But there was also a small loss after a big increase in marketing spend. GrubHub management told investors that competition would not affect growth, but analysts interpreted the company's findings that the rise of DoorDash and Uber Eats would put pressure on the deployment apps.
Uber and DoorDash refused to provide more details, but their businesses are rapidly expanding their reach. DoorDash says it covers 80 percent of the US population, and Uber Eats is said to have reached more than 70 percent, in addition to more than 100 cities in Africa, Asia and Europe. Sallenave, the head of the Uber Eats for the US and Canada, predicts that food via apps will become the norm everywhere, not just in urban areas. "We believe in principle that we can do this business economically, not only in big cities, but also in small towns and suburbs," she says.
Other great WIRED stories