BIG DATA PLAN

Ford data crunchers help dealers fine-tune inventory

Related Topics

Ford Motor Co. is trying to take the guesswork out of dealership inventory ordering, using advances in big data technology to offer its stores sophisticated recommendations for their orders.

Some dealerships, however, have been hesitant to embrace the data.

"I order my own vehicles," said Ford Hall, fleet and commercial sales manager for LaFontaine Ford in Lansing, Mich. "I've been doing this for 16 years, and I know what I want in my parking lot. They don't."

Other dealers use the system, which Ford supplies free, but the varied reception is not surprising to experts. Using big-data tactics to fine-tune inventory orders is in its infancy, they say. One analyst says effective use of big data could save $100 or more per car sold.

Ford rolled out its Smart Inventory Management System in 2009, providing relevant regional data and inventory recommendations to dealerships in the United States and Mexico. The company's predictive analytics group sees the tool as a way to cut the days that cars sit on dealer lots by better predicting consumer demand. While other automakers use some analytics for dealership inventory, Ford has made it a priority.

"Our goal is twofold," said Bryan Goodman, a Ford research scientist. "To help dealers have better, faster-turning inventory and to better predict what dealers are going to

order to better manage the supply chain." 

Related: Outside companies make inroads with inventory software

Five years after the system's introduction, however, independent companies have begun to develop their own inventory and pricing software.

And one executive for a competing system notes that dealers tend to be suspicious of factory recommendations. (See related story, at right.)

Karen Newman, vice president of the automotive and aerospace industries for IBM's Global Business Services in the Americas, said the auto industry produces massive amounts of information that lends itself to data crunching.

"There is a real value around the use of big data and using these kinds of tools," Newman said. "It's how we can do things smarter, better, faster and cheaper."

The predictive analytics group at Ford, led by data scientist and manager Michael Cavaretta, decided to take big data to Ford dealership inventories in 2007. According to Newman, there has not been much activity from other automakers in the use of data analytics to shape dealership inventories.

"I think this is a great place to apply big data and analytics to help OEMs and dealers solve problems," she said.

Anil Valsan, global automotive lead analyst for EY, said that although most automakers use analytics in inventory management to some extent, Ford's Smart Inventory Management System, or SIMS, is a more sophisticated program. As a whole, he said, the industry has not matured in the application of big data to inventory.

"From our engagement and discussions, a number [of automakers] are thinking of having similar programs," Valsan said. "Ford is a reflection of the benefits that such analytics provides."

He added that though inventory management analytics is in its early days of development, it has the potential to save two- to three-digit figures per car on the dealership lot.

 

What is big data?
Big data is information generated in amounts so large that traditional techniques cannot be used to analyze it.

As the technology to process data has improved, the auto industry can apply new information throughout a vehicle's lifetime, saving time and money.

For instance, a manufacturing plant can analyze pressure readings and other sensors from machines to anticipate problems and avoid production delays. As in-car connectivity improves, drivers can be alerted in advance to maintenance issues.

Shawn Nelson, a senior solution architect for Computer Sciences Corp., said applications can range from advertising to targeting vehicle recalls. Possibilities could go as far as receiving coupons in your car for a specific store as you pull into its parking lot.

“It's like the moon launch,” Nelson said. “The concept is way out there, but there's value, and it's possible.”

"I think this is a great place to apply big data and analytics to help OEMs and dealers solve problems," she said. 

Anil Valsan, global automotive lead analyst for EY, said that although most automakers use analytics in inventory management to some extent, Ford's Smart Inventory Management System, or SIMS, is a more sophisticated program. As a whole, he said, the industry has not matured in the application of big data to inventory. 

"From our engagement and discussions, a number [of automakers] are thinking of having similar programs," Valsan said. "Ford is a reflection of the benefits that such analytics provides." 

He added that though inventory management analytics is in its early days of development, it has the potential to save two- to three-digit figures per car on the dealership lot.

Cavaretta: Went to dealers in ’07

Sifting through data


SIMS examines a dealership's past sales, as well as the sales patterns of other Ford dealerships in the area. The data, stored in Ford's supercomputers at its product development center in Dearborn, Mich., are analyzed to generate recommendations for inventory orders, including model, trim and feature combinations.

Relevant data are then sent to dealers on a weekly basis, and it is up to them whether to follow the recommendations. Ultimately, however, Goodman said dealers are in the best position to fine-tune orders because they can better evaluate additional information on the ground than the supercomputers in Michigan can.

Goodman said there is about a 98 percent match between what SIMS recommends and what ultimately is ordered by the dealerships, and many times dealers "swap without realizing": Even if a dealer does not order a recommended vehicle, there is a good chance the vehicle he or she ends up ordering has been recommended to another dealer in the area.

"There is a tremendous amount of information," Goodman said. "There are 5,000 Ford and Lincoln dealerships combined in the U.S. and Mexico. That's a lot. SIMS looks at every single vehicle and every single feature and every single option."

Cavaretta added that there is an "explosion" of combinations for vehicles that can be ordered.

While feature combinations are seemingly limitless, the commodities to build them aren't. Though a certain feature may be optimal for a specific region during a specific time, production limits orders, and compromises may be made.

"Let's say there's high customer demand for the EcoBoost engine, but Ford can't make enough to meet demand," Goodman said. "We have to figure out the best place to put those and the best alternative.

"It's not the best situation for anyone -- dealers aren't happy, customers aren't happy, we're not happy."

However, SIMS has been able to adjust alternative recommendations if there is a shortage. For instance, there is a high demand for the F-150 Lariat with a moonroof, but there are no Lariats available for a specific dealership. Instead, SIMS may recommend the F-150 XLT, which may not be as popular of a model when equipped with a moonroof. But SIMS can recommend more attractive features specific to the XLT.

For newly introduced models, evaluating sales history to make inventory recommendations isn't a possibility. Experimenting with the introduction of the Fiesta in 2009, SIMS turned to Ford's build and price consumer Web site to find the most attractive combination of features.

"We developed ways of filtering data to represent a good view of what people are really shopping for," Goodman said. "We can use that data to guide recommendations."

“There is a real value around the use of big data and using these kinds of tools. It's how we can do things smarter, better, faster and cheaper.”
Karen Newman
IBM Global Business Services in the Americas

'Enough flexibility'


SIMS was introduced as a pilot program to 30 dealerships and then was extended to all 5,000. Since the rollout in 2009, Ford has been receiving feedback from dealerships to continually improve the tool.

"We received a lot of very positive feedback, also some feedback that was less than positive," Goodman said. "A few of the people ordering felt threatened. We never intended to replace them, just give them tools."

Like Hall at the Lansing dealership, some inventory managers are reluctant to take suggestions from an outside source after developing their instincts over the years.

"I don't even look at them," Kerry Petratos, commercial and fleet sales manager for Chapman Ford Sales Inc. in Philadelphia, said about the SIMS inventory recommendations.

For other seasoned managers, such as Alan Beery, sales manager at White's Maibach Ford in Orrville, Ohio, the data tell them what they already know.

The recommendations "are pretty much in line, but it's not always the way we stock them," Beery said. "They're in line with what's selling in the area, but I don't like to leave [orders] up to them."

"It is what it is," said Tom Lammel, fleet sales manager for Fallon Auto Mall in Fallon, Nev. "It doesn't save time or cost time."

Though some dealers are skeptical or indifferent toward Ford's data analysis, others are embracing it as another tool for generating orders.

Terry Kidd, owner of Kidd Ford in Morrison Tenn., said intuition gained from experience is still valuable, and when used in conjunction with SIMS, can produce optimal results.

"When we're doing something that we're accustomed to doing, it's hard to realize, 'Wow, this is being generated in the right way,'" Kidd said. "Your intuition can still work, but it's even more interesting to look at orders being generated and try to figure out why."

Kidd said SIMS has been particularly helpful in ordering complex feature combinations for trucks and choosing vehicle colors.

"Is it absolutely perfect? No," he said. "But there's enough flexibility."

“There is a tremendous amount of information. There are 5,000 Ford and Lincoln dealerships combined in the U.S. and Mexico. That's a lot. SIMS looks at every single vehicle and every single feature and every single option.”
Bryan Goodman
Ford research scientist

Still developing


One advantage of data technology is its ability to be improved constantly. Though SIMS has been around for five years, it still is a developing data system, and as related technology becomes more sophisticated, so will Ford's data tool.

"We're still working to improve the system," Goodman said. "Our computational ability has gone up."

He added that the team is looking at ways to extend order predictions out to eight to 12 weeks and gather more information on upcoming commodity shortages.

Kidd, the Tennessee dealer, said the switch to Web-based ordering for his inventory has made using SIMS recommendations much more effortless. Suggestions adapt to his past order patterns.

"I've got to be able to change with the times, too," Kidd said. "It's a good program that's being constantly improved. It's not perfect yet."

Contact Automotive News


advertising
Have an opinion about this story? Click here to submit a Letter to the Editor, and we may publish it in print.

Or submit an online comment below. (Terms and Conditions)


Newsletters & Alerts
Latest Headlines