Big Data road for car giant Ford

"The data will set you free," says Alan Mully, chief executive and a Mulally with 2007 as CEO of Ford Motor Company. At the time, Ford Motor Company was like a sinking ship, losing money and failing to find a direction. 6 years later, Mr Mulally's remark has become a mantra for the carmaker, and no one should come to see Alan without conclusive data, whether good or bad.

"Ford was rarely exposed to bad news in the past." "The change that Allen brings to the company is that we're going to use the data to talk and not shoot the head," said John Ginder, a 22-Year-old Ford. I've seen this change over the last five or six years, and our CFO and COO are the biggest supporters. Today, all decisions are data-driven. ”

Indeed, data and analysis have permeated every activity in Ford, from predicting the price of goods to understanding what consumers really need, from what models the company should produce for its customers, to what parts the car should be purchased, and whether to add cars and trucks. Behind these activities are Ford's nearly 200 large data and analysis specialists, from different disciplines working in Ford's Department of Excellence, which will be involved in various parts of the Ford company, including marketing, research and development, and credit services.

Since 2007, these analytical experts have played an important role in the company's various strategic adjustments and tactical decisions. For example, decide which brands and models of cars to shut down, where to buy parts and raw materials, and how to get distributors to adjust their stock to increase sales.

Ford's subsequent quarterly financial data, stock prices and the company's latest annual report attest to the correctness of these strategic adjustments and decisions. In 2009, Ford renewed its first profit in 4 years after a record loss of 4.6 billion dollars a year. Just this year, Ford launched 25 new car production lines, the United States since 2007 to sell 2.3 million cars and trucks, and become the first domestic car brand. By October 2013, Ford had achieved a 17th consecutive quarterly profit.

There is no doubt that data analysis is an important competitive tool for all carmakers, whether it is analyzing customer needs, automotive performance and production processes, or improving product quality and forecasting markets. In General Motors, for example, it has been collecting vehicle diagnostic data and other information through its OnStar system over the years. Ford's data analysis focuses on customer preference analysis.

Thilo Koslowski, Gartner's auto analyst, sees Ford as a leader in the field. "Today, all car manufacturers use data analysis to understand the user's evaluation of their products to varying degrees." "he said. "I've seen Ford do it earlier than anyone else, especially in terms of understanding customers ' preferences," he said. ”

At Ford, "There's a lot of urgency and it's going to push people out of their minds." "Nothing like a crisis has made us realize the importance of analysis," Ginder said. The 2007 and 2008 companies in crisis were a major turning point, both for companies and for the understanding that analysis can play a role. After the crisis, new management and new systems allow us to look at things in a completely different way. ”

At present, Ford's most significant return on data analysis is focused on three areas: determine what the customer wants, manage the complexity of the vehicle, and deliver the appropriate dealership to the appropriate distributor to ensure that users in a particular geographic area can buy the desired car.

Deliver the right car to the right dealer

For decades, like all carmakers, Ford has relied on a wide range of market research, surveys and interest groups to explore people's deep feelings about driving cars.

"However, these methods do not always give us a complete picture, because to compare, we must standardize the data." "One way the company solves this problem is by monitoring social media to get more specific information and customer feedback," said Mike Cavaretta, head of the Ford Company's predictive analytics program.

"The advantage of social media is that it expresses what people really think. And the content is also rich, not just cool. "Cavaretta said.

For example, Ford has monitored social media to see how people view change lanes as turning lights flashing 3 times this problem. In this process, I know a lot of things that I didn't know before. "We found that the vehicles on some lines were not set high enough or in the wrong place at all," he said. When people talk about flashing lights, they really care about other things. Cavaretta explained that Ford was able to incorporate such feedback into decisions about new products and functions.

Providing the driver with the desired function is one thing, and making sure that the dealer has the same key to the car that the customer wants to buy is absolutely the key to ensuring the company's profits. "Many customers walk into a dealership and hope to be able to drive a new car home that day, and not a lot of dealership cars," he said. Bryan Goodman, a research scientist at Ford, explains that Bryan Goodman is responsible for analyzing systems to support sales and marketing as well as related material planning and logistics. "We have to make sure that we have the right vehicles in every dealership, that they use the right engine and that they have the right configuration." "he said.

Data consolidation

To deliver the right car to the right dealer, Ford must consolidate and analyze multiple data streams, including data on the cars it has already produced and sold, data on cars that have been sold but not taken out of stock, and visits to the company's Web site to search for car-related configuration. These data are then combined with economic data (including employment rates) to predict the sales of the vehicle. Ford says the system is an intelligent inventory management system or Sims.

"Over the years we have been experimenting with a variety of experiments and applications to achieve the desired results, but until recent years there has been substantial progress, thanks to a significant increase in computational capacity." "There are 20,000 computing cores in our data center that have 1.5TB of memory," Goodman said. It is precisely because we have powerful computing resources that we are able to handle various types of data. ”

In fact, Sims is recognized as a key factor in Ford's turnaround and a key to the overall success of Mulally's global "one Ford" strategy.

"As our products go global, our sourcing comes from around the world, which means we face new management complexities." Goodman said that in the past, if you want to launch a function in a market, our production system must be able to match, make it feasible. Considering the roof height of different, different interiors, and even different wheels, so to provide an astronomical combination, is basically 3000 times the result of 3000, and customers because too many choices overwhelmed.

"And now we want to match our customers ' needs with our global supply chain." It should be said that this is a common challenge for the industry as a whole, and once realized its value is very big. "The key, he says, is to understand the customer's core preferences, and then sort out and simplify the company's various models, rather than building a very broad mix of models and features." The ultimate goal is to make sure that the cars that most customers most expect to have are produced.

This is where Sims really works. In this way, the timetable and forecast accuracy of the Auto assembly plant can be improved significantly because of better algorithms. Goodman said, "2007, we can do this, but also need two, and assembly workshop need 20 minutes to get results." And today, we can finish it in two minutes. ”

Ford dealers also benefited from Ford's data analysis. Some 3,500 Ford dealers receive reports weekly, and those who subscribe to Sims's recommendations say they can sell cars at higher prices and sell faster.

Big data analysis on the road

But Ford analysts say they are just beginning to touch a bit of big data. In the automotive industry, the next frontier for big data is the data flow of the vehicle itself.

"The amount of data produced by vehicles is huge. Ginder said Ford's Fusion Energi is 108 miles per gallon for plug-in hybrid cars, and it can generate 25GB of data an hour. Ford currently offers three kinds of hybrid cars. All cars are equipped with embedded modems, customers can choose to open, so that the vehicle-related data back to the manufacturer.

"We collect each customer's charging information. So we know where they charge, how far they drive, how much of it is powered by gasoline, and whether they travel often. This data will help define the next product. "But the relatively small number of vehicles in our data so far is certainly going to be getting bigger," said Mike Tinskey, director of Ford's vehicle electrification and infrastructure. ”

Even so, the collection of data is valuable, especially when combined with other data from social media sites.

"Because it gives us a better understanding of our customers." "If we know how people use their cars and how they talk about it, and see how it integrates with our internal business processes, we can do a lot of things," Cavaretta said. For example, once integration is achieved, we can clearly know where to improve and where to build. ”

Ford will also provide a report to drivers. "We are collecting large data, processing them and then providing a monthly report to all interested parties," he said. "Customers receive a link every month, a customized monthly report detailing their driving conditions and the performance of the vehicle itself," Tinskey said.

Connect to the Internet of Things

Experts from other Ford analytics teams, such as Tinskey and Cavaretta, predict that more and more vehicle data will be combined with other types of internal and external data via the Internet of things. This could enable customers to adjust their car engines to their own needs in the future and choose the right time to recharge their cars for lower price discounts.

"The coolest thing is that we know where these customers charge and pay, and we know the power generation in each ZIP code area because we have a power supply database in all parts of the United States and we can integrate it with the power billing data." Tinskey explained.

"Our initial idea is that when a regional power supply reaches a tipping point, the vehicle can respond to a call from the power supply department without electricity," he said. One scenario is that customers postpone charging their cars when demand is high and power is tight. Thereafter, the power supply will be compensated for at a lower price. He said the program is similar to the domestic heating and cooling policies that are now in place. Tinskey explained, for example, that his home air-conditioning was measured separately, and that if it was to enjoy a lower price, it would be necessary to agree to switch off or lower the level at certain times when the power supply department was in demand. ”

Tinskey that the future popularity of Internet of things allows them to integrate internal and external data and vehicle data, resulting in some new forms of service.

"The likely future scenario is that the vehicle becomes a mobile sensor platform that is connected to sensors or streetlights on the road to monitor traffic conditions, weather conditions and energy use." This information can then be returned to the vehicle to conduct traffic to the driver's behavior and guide the transportation. The real key is to get information from sensors embedded in the surrounding environment, and then better understand how to adjust the vehicle, or optimize the business process. "he said.

These are complex and difficult to achieve, but for Ford, at least it has started, and Ford has tied its future tightly to big data analysis.