Features

Case Study: 3M

3M Innovates

The manufacturer retains accounts and expands sales through the power of predictive analytics.

The manufacturer retains accounts and expands sales through the power of predictive analytics.

“Innovative Technology for a Changing World” is more than a slogan; it’s the driving force behind 3M and everything it does. The manufacturing giant serves its customers through an organizational model consisting of six autonomous business segments that share technological, manufacturing and marketing resources to increase speed and efficiency.

In 1996, 3M began its global enterprise data warehouse (EDW) initiative, using a platform from Teradata. The EDW ultimately grew to encompass data about customers, products, sales, orders, financials, the supply chain and more. Having built a strong foundation as a data-driven organization, the St. Paul, Minn.-based company is taking the next step in business intelligence (BI) to move beyond basic reporting into predictive analytics.

A Solid Platform

Because 3M had developed a multi-sub­ject, enterprise-wide data environment, the choice to reuse information already available for predictive analytics was clear, according to Global BI Manager Jeff Robinson.

At a Glance

3M

 

 

The Company: Minnesota-based manufacturing giant 3M produces thousands of innovative products and is a leader in scores of markets including healthcare, highway safety, office products, abrasives and adhesives.

The Challenge: As a data-rich company, it sought to leverage its multi-subject data environment in new ways through predictive analytics.

The Solution: Building on its global enterprise data warehouse foundation, the manufacturer gained valuable insights into products, markets and customers.

The Results: 3M can now predict business at-risk with unbelievable precision and accuracy. Through in-database analytics, in one case, it reduced processing time from 60 hours to 2.5 minutes.

“Many companies look at analytics to isolate a specific problem,” notes Robinson. “However, with our EDW from Teradata, 3M can process millions of various cus­tomer/product combinations. Most complex companies are performing predictive analytics for one business unit, product line or geography with multiple data warehouses. However, our integrated environment lets 3M do predictive analysis globally across multiple product lines and lines of business.”

Accelerated Performance

By shifting to in-database analytics, 3M has compressed job streams that used to take 28 hours to only 30 minutes. Previously, the business intelligence (BI) team would extract data from the enterprise data warehouse and put it into a separate database system.

To accommodate a single user, the previous system frequently had to backlog all other queries. With users around the world trying to access specific information, it could take days to get the requested data.

“The elapsed time on one of our predictive applications was 60 hours when using a previous architecture,” notes Global BI Manager Jeff Robinson. “After moving it to the Teradata environment, the elapsed time went from 60 to four hours—a big improvement. Then we took a look at the code to optimize it, and we were able to run the same process in 2.5 minutes!”

This strategy earned the company a significant competitive advantage. “Without an integrated data warehouse, I don’t know how we would attempt to ask for the right answers, because we wouldn’t even know how to identify the right questions,” he says.

Work of ART

As the manufacturer moved beyond stan­dard reporting into the realm of analytics, it soon uncovered numerous business uses for these capabilities. One of the first was account retention. The ability to uncover and analyze buying patterns was key to this effort.

“Without an integrated data warehouse, I don’t know how we would attempt to ask for the right answers, because we wouldn’t even know how to identify the right questions.”

—Jeff Robinson,
global BI manager, 3M

“We started predictive analytics a few years ago in the Industrial and Transportation business,” explains Robinson. “We developed an Account Retention Tool [ART] that looks at true buying patterns over a period of two years. This helped us focus on seeing where 3M could be at risk of losing a sales opportunity.” The development of ART was a tremendous step toward building an enterprise-level approach to predictive analytics. Executive spon­sorship helped facilitate the change-management process.

“We are now able to predict business at-risk with unbelievable precision and accuracy,” says Global Enterprise Data Warehouse Manager Connie Garritsen. “We frequently deploy ART at a country level, which requires analyzing huge vol­umes of data. For a small country alone, we typically process at least 15 million rows of data in order to identify custom­ers and products at risk.

“Because the company has six busi­ness sectors and operates in more than 65 countries, the sheer number of customer and product combinations is staggering,” she adds. “Managing account retention is only possible because of our Teradata platform and the level of analytical capabil­ity that our predictive team provides to our business clients.”

Driven by Data

Successfully working in tandem with busi­ness users has been a key lesson learned for the team. At times, helping business clients articulate their needs has been a challenge. However, having a reliable data platform and a culture driven by data makes all the difference.

Predictive Analytics Impact

In addition to gaining insight into its products, markets and key customers, 3M is applying its predictive analytics environment in multiple areas:

  • Corporate sales projections. The platform is credited with helping 3M forecast more pre­cise and reliable sales performance, operating incomes and earnings per share.
  • Succession planning. The company’s research scientists, analysts, product development specialists and others have spent years refining products and market positioning. To retain this vast corporate knowledge base, 3M is using its enterprise data warehouse to preserve and share knowledge as the company identifies and trains future leaders.

“The correlations are frequently very different from what the businesses thought,” Robinson explains. “They some­times make incorrect assumptions based on indicators. So while a business might think X, we can take the data and discover the reality is Y.

“By quickly showing them their assumptions are wrong, we can help them better anticipate potential business problems,” he adds. “The 3M IT team is really the advocate voice of predictive analytics, evangeliz­ing its benefits to the broader organization.”

“We are now able to predict business at-risk with unbelievable precision and accuracy.”

—Connie Garritsen ,
global enterprise data warehouse manager, 3M

Only the Start

The future of predictive analytics is bright for 3M. The company plans to expand into social media to capture more consumer data. “Small segments within the businesses are already doing it,” says Garritsen. “We want to use the data to look for trends and event-driven triggers. We are also considering the application to predict defects on an assembly line in a manufacturing site,” she adds. “And at the distribution center, we will be able to better plan for the ordering of raw materials.”

By moving from an environment where business units worked autonomously to a cohesive organization where deci­sions and insights can be made based on enterprise data, 3M is getting the most out of its EDW and predictive analytics. “That ongoing evolution continues to have a sig­nificant business impact on our company,” notes Robinson.

“We are investing in predictive analytics at 3M,” says Janel Haider, IT director, Global Applications Center of Excellence. “We have the EDW platform in place and the opportunity to support timely busi­ness decisions with good data. We have businesses lining up to take advantage of our capabilities and we want to support them as best we can.”


Your Comment:
  
Your Rating:

Comments
NetApp Q1-2014