Lukas Gojda/123RF

4 Companies Already Leveraging Big Data to Shape the Future of Manufacturing

  • 24 October 2017
  • Dave Rauschenfels

“Do many calculations lead to victory, and few calculations to defeat” 

-Sun Tzu, The Art of War

In some circles, modern business has been compared to warfare. The cleverest corporations usually defeat and destroy the less intellectual businesses. In order to lead an intellectual company, the leaders must be experts in their own business and that of their competitors. This can, of course, be very difficult with the complexities of running a modern manufacturing business. Despite the challenge, four companies are already making strides to develop full transparency into their production with the use of big data. In enterprise, big data is just another term for intelligence. In order to lead your army to victory, you must have full data and analytics on the internal operation of your own company.

Tesla Motors

The boutique electric automotive maker has ambitious plans for growth in the next three years, with plans to build one million cars by the start of the next decade. Unfortunately, this has been a formidable task even for Elon Musk with Tesla only building 76,000 vehicles last year. This is far below projections and sources place the company in the red. This problem is only compounded by the automaker’s plans to produce the Model X SUV and the Model 3 in coming years.  

Aleksei Potov/123RF

To address this, the company is feeding its production data into the Manufacturing Execution System or MES. Within the factory, MES operates like an all-knowing plant manager. It knows when a part arrives, and where it belongs in the assembly line. The software also monitors all production orders and quality issues and records the measurements to an Oracle/SQL server database. On top of the SQL database, customized results are stored in a separate MySQL test database.

Tesla also uses a web-enabled QuickBase for continued monitoring of quality issues and Microsoft Excel for the occasional report. It is reasonable to expect that all this division of figures will cause confusion, and it has. To solve this, the company is merging Tableau into its network. The software lets users with no technical expertise create data visuals and share them with interested staff.

General Electric

It turns out that conglomerates can be very resourceful when money is on the line. As you must know every time you fuel your car, gas can be expensive to purchase. Producers also pay a steep price when the machinery goes down. A single unproductive day at a liquefied natural gas platform can cost twenty-five million dollars. On average, a platform can be down five days a year. Oil and gas is part of GE’s enterprise, and they’ve developed a solution to combat the unplanned downtime.

Predix software is a cloud-based software template for writing industrial Internet applications. While GE has long built sensors into its machines, only recent advancements in computation have made it feasible to monitor inputs from an entire fleet of industrial machines.

The scale of real-time data that needs analysis is astounding, with over 50 million data points from 10 million sensors across the industry. The company is also mining the data for signals on the configurations of machines operating at their best or worst. This has already paid bonuses when the company noticed that jets operating in the harsh Middle East needed more downtime. After their scientists consolidated the jet engine performance according to the region, it soon became clear that dust in the region was clogging the engines leading to overheating. All that was needed was more frequent cleaning to save the airlines seven million dollars per jet in lost fuel.

Sergey Kohl/123RF

This discovery served as a benchmark for the company in its efforts to automate the pumps and machinery of the oil and gas industry, which is by itself worths twenty billion dollars in sales. Even for a company with GE’s reputation, many customers have been too private to share their data. The problem is only compounded by the diverse suppliers to the gas industry, which makes systemic monitoring difficult. Instead of selling complete monitoring. The company has chosen instead to market lightweight pilot plant solutions to specific producer concerns.  


For many people, anticipation is the best part of shopping. You just can’t wait for the new shiny toy to arrive. As it turns out, it can be equally pleasurable for Amazon to predict your next purchase. They have even filed a patent for an algorithm to forecast where a specific good will ultimately be sold. In summary, it directs the merchandise to a convenient warehouse in the state or region where those goods are expected to be a hit. Then when the orders arrive, the remaining distance to your door is short. This algorithm isn’t just academic. One or two-day shipping can be real money when the product needs to be shipped cross-country.

Within the same city or region, it is trivial. The algorithm uses browsing history and past purchases to anticipate future orders. Then when the holidays come along, they already know what the next big toy will be.


Not to be left behind, Siemens has developed its own Manufacturing Execution System for monitoring and analyzing production performance. The software provides an interface or OEE cockpit for visualizing the real-time performance of production machinery. Besides visualizing data, the MES system pushes data into quality management systems and performs an initial analysis of the results. Industrial functions can be analyzed in terms of cost, frequency of occurrence, or the duration of the process. Customers have already received bonuses from the technology, with Caterpillar reducing production delays by half. Machine uptime has also been increased by thirty percent, which has risen earnings by over a million dollars.

Максим Кузубов/123RF

Siemens will soon be rolling out software for the complete digitalization of manufacturing from initial design to the execution of production. They plan to get this done with Digital Twin, an application for virtually simulating every step of the assembly process.

Adoption of full digital manufacturing is expected to be slow with some businesses still using equipment over twenty years old. Effective data modeling also needs ninety-nine percent cloud reliability, a steep price for some manufactures. Siemens is pursuing cheaper hybrid options for manufactures to install sensors on their equipment and connect to the Microsoft Cloud.

About Dave Rauschenfels