In a recent interview from Gigaom Structure Show podcast, Ford Data Science Leader, Michael Cavaretta discussed lessons learned in harnessing the power of big data, (large complex data sets requiring advanced processing), and how to best put it to use. “Big data” at Ford can mean anything from data coming off the car and truck production lines to customer feedback at dealer locations and customer service centers. In this interview, he talked about how data can impact automotive design; how to determine the analytic processes and tools to use, weighed against overall costs to produce useful output; and where to find the appropriate resources to hire or utilize for data analysis.
Data can influence Design
One of the ways that “big data” has been put to use at Ford is through the adoption of a lightweight aluminum frame instead of steel, in an effort to reduce fuel consumption and therefore reduce co2 emissions being put into the environment. Data analysis helped generate a move toward the co2 glide path program, which came about when technology options were weighed against cost and benefits.
Adds Cavaretta,“We take a look at the costs, we take a look at the benefits…and then we kind of put them all into this gigantic hopper and we run an optimization on them. We kind of sort everything around and find [what] the best combination that we can look at for the next few years. So, light weighting and even aluminum was one of the technologies that went into the hopper there and obviously then got picked.”
Another way that design was influenced by data at Ford was through Social media. It was used to determine why Ford Fiesta US owners didn’t like the 3-blink turn signal as compared to European owners. After marketing asked Ford’s data scientists to probe further into what Americans didn’t like about the 3 short blinks, it turns out that they actually had no issue with the turn signal itself, but it’s placement on the steering wheel.
Weigh Value-added against Cost
According to Cavaretta, “[W]hen Alan Mulally came in, one of his tenets was ‘the data will set you free,’ and [he] has been always very, very focused on data.” Consequently, Cavaretta’s team has responded to the doubled data analysis being requested by executives at Ford. They have been tasked with presenting the data in a clear usable fashion, with understandable analytics performed accurately to support the decisions being made.
Ford is asking themselves and their data scientists tough questions: Are they keeping all data? What is the most valuable data being stored? Are historical data sets too large? Should they be implementing tools like Hadoop, which allows for easily searchable and better organized structure for quick data analysis? What is the real value for Ford? For large enterprise like Ford, it may be too costly to rip out the existing infrastructure and replace it. Ford has determined that they need to realize an immediate value analysis for cost-justification.
Data Scientists don’t have to be “Unicorns”
Some might hear the term “Data Scientist’ and liken it with “demi-god”. Although this position does require good computer science, customer service and analytical skills, Ford has found that the blending of these skills doesn’t necessarily require a “unicorn”; a one of a kind, high-salary individual. They have been able to utilize many internal resources in these roles with great success. In putting this infrastructure together, Ford has been able to pull data sets from customers into one area and run analytics across all data sets to improve both processes and customer satisfaction. In doing so, data scientists at Ford are utilizing the power of “Big data” through analytical tools, cost-justified decision-making, and the implementation of the right resources.