Challenge

A historic global supply chain crisis brought on by the COVID-19 pandemic shed light on the inefficiencies of traditional procurement practices. Change agents at TimkenSteel saw it as an opportunity to rip off the proverbial Band-Aid and “lean up” through what Supply Chain and Commodity Modeling Manager Nicholas Galbincea calls an “optimization mentality.”

Solution

TimkenSteel has since implemented “Business 4.0” – an initiative led by Galbincea that applies the principles of Industry 4.0 to procurement strategies for steelmaking commodity inventory. JMP® statistical discovery software became a critical partner in the company’s transformation as it provided a lean solution to automate and consolidate entire workflows into one end-to-end platform.

Results

Within the last three years (2020 – 2022) the modeling and inventory tools Galbincea built have saved TimkenSteel approximately $7,300,000 – a figure he anticipates will only grow as he continues to build additional models and tools in other areas of the business. These savings are attributed to the inventory models that Galbincea has built for optimization of electrodes, ingot molds, and MRO inventory. In addition to financial value for the MRO inventory modeling, the process determining item reorder point (ROP) and economical order quantities (EOQ) has been dramatically streamlined, enabling TimkenSteel to move from a manual process of reviewing each item individually to an algorithmic based analysis allowing all MRO items to be reviewed in minutes. “I built the models, reviewed the logic with our VP of Engineering, Manufacturing Excellence, & Reliability (Andrew Bissot), Reliability team and other Procurement team members (Val Johnson, John Weekly & Steve Murphy), and after several iterations and lots of debugging, we saw them work and profit the company,” Galbincea reflects. “To me, that's awesome. That’s making a difference.”

“It’s very easy in any industry with as much history [as steel manufacturing] to get caught up in doing things the same way year after year. But that just can't happen anymore,” says Nicholas Galbincea, TimkenSteel’s Supply Chain and Commodity Modeling Manager. “One of the most unhealthy things you can say in any business is: ‘This is how we've always been doing it.’”

TimkenSteel – a leader in high-performance alloy steel bar and tubing – has in recent years embraced ambitious sustainability targets and is leveraging lean practices and continuous improvement to reduce environmental impact and improve the efficiency of its annual melt of roughly 1.2 million tons.

Thanks to Galbincea’s unique vision and the innovation-minded leadership backing him, TimkenSteel is borrowing a strategy from the world of engineering – Industry 4.0, a smart manufacturing philosophy – to transform other aspects of the business. This optimization mindset is now helping the company gain traction as an early adopter of big data in procurement and supply chain strategies.

“Analytics is the future, and that future is now,” says Galbincea. “If you don't get on board, you're going to be left in the dust. And I know that sounds cliché, but it's truer now than ever.”

Scaling lean ideals beyond engineering 

Galbincea, who has degrees in both mechanical engineering and materials science, brought a computational perspective to TimkenSteel where he started in an R&D role. In early 2020, he moved into the company’s purchasing department and now manages direct, consumable materials procurement. The timing of this move, however, isn’t lost on Galbincea.

“Lo and behold, it would be a crash course in one of the hardest global supply chain issues that had probably ever faced the world,” he says. The global COVID-19 pandemic, which quickly spread globally in 2020, caused the most significant supply chain crisis in the history of the post-industrial world. Manufacturers in every industry on every continent were hit by staggering disruptions beyond their control, and TimkenSteel was no exception.

Faced with a historic challenge only weeks into his new role in commodity management, Galbincea saw an opportunity for TimkenSteel to make a clean break with the way supply chains and commodity inventories were managed in the past and instead embrace analytics transformation. “It made total sense. It's not rocket science to correlate what we should be buying – our inventory – based on the demand and the needs of the company."

Galbincea’s plan was to use statistical methods common within Industry 4.0 to optimize commodity inventories. “It's taking an analytical approach to what we should be consuming based on a master plan forecast and correlating that directly with what we should be buying,” he explains. “I started applying these analytical models to ask how we could reduce our inventory. 2020 was the time to reduce inventory and lean up because business was slow.”

Having studied computational approaches to materials science during his graduate training, Galbincea was no stranger to the idea of Industry 4.0 – at least in technical applications. Applying algorithmic approaches in purchasing, however, was relatively uncharted territory, not just at TimkenSteel but across the entire industry.

“We began optimizing our inventory for the current business situation, and here we are three years from that [inflection point] and the company is doing very well. 2021 was actually a record cash flow year for TimkenSteel since its inception,” he says.

JMP® is an all-in-one solution for automating data workflows

At any given time, TimkenSteel houses maintenance, reliability, and operations (MRO) inventory worth millions of dollars in a controlled warehouse setting. Everything from pumps to nuts and bolts must be on hand in the event of a breakage or failure in the manufacturing system so as to avoid costly downtime. Managing thousands of inventory items can be daunting, even when the global supply chain is running without disruption.

Traditionally, inventory management is the purview of a team working by hand to index supply and initiate procurement. However, TimkenSteel’s lean approach meant Galbincea would need to create new efficiencies through statistics.

“[At first] I started building inventory models for specific direct commodities – electrodes, ingot molds, etcetera – because I didn’t know how much inventory we should be carrying for the current state of the business,” Galbincea recalls. That’s when he had an idea for the thousands of items in MRO inventory: automate the entire system. “I said let's build a model. Let's build an algorithm to handle the whole thing!” To do this, he would need a tool, and that tool was JMP®.

Having used the statistical discovery software in his prior R&D role, Galbincea says JMP was the natural choice. “JMP in itself is a very user-friendly tool. Just from an interface standpoint, it’s beautiful. It's relatively easy to navigate and start graphing your data, doing linear correlations. Even machine learning – the neural net module – is straightforward.”

Moreover, he adds, the learning curve associated with any new tool is eased considerably in JMP by intuitive instructional resources. “The help tab at the top has short tutorials and the search tool is especially helpful for figuring out which formulas are already built in,” he explains. “JMP has a huge database of [readymade] formulas that you can use instead of having to code or figure it out yourself.”  

“The visualization is great too. I can't give enough kudos over how easy Graph Builder is to use. Showing data with a graphical representation is so much more powerful than just presenting numbers, and with inventory, I can show our management visually – and in a single graph – where the model is saying our inventory levels should be. It’s very powerful because it makes sense.”

$7,300,000 of cost avoidance in 3 years

Starting with major steelmaking commodities like electrodes and ingot molds – and seeing the possibility in other categories – Galbincea began developing an MRO inventory optimization model. “From the Query Builder in JMP, I was able to query from the controlled warehouse's database, collecting the historical consumption of each item,” he says. “I then used JMP scripting (JSL) to massage these data tables and do a bunch of linkages to connect each item’s data in order to perform analytics for calculating reorder metrics," he explains. This workflow was easily automated with a few scripts.

Now, with the click of a button, the system accesses, and queries data from TimkenSteel’s database, runs an optimization algorithm for each item based on historical consumption and current lead time calculations, and produces a set of recommendations: when items should be reordered and in what quantity. When the balance of any one inventory item drops below a certain threshold, the system will trigger a buy. Thresholds are set so as to account for global supply chain events, including longer than average lead times.

“Working in JMP, I was able to, over several months to a year, build [an MRO] model, [automate] data query, work out bugs and then go live with it,” he says. “Now it's as simple as clicking, running two scripts, and [the dashboard] spits out new numbers [taking into account] updated lead times and each item's most recent but still historical consumption.”

The MRO inventory optimization model went into effect in March 2022 and within its first 6 months alone, Galbincea says, generated more than $700,000 cost avoidance. “It’s a real golden goose because over time, it will keep saving us money,” he adds. In addition, the other commodity inventory models have generated approximately $6,600,000 in cost avoidance since their creation in 2020.  

A living, breathing model delivers value to the company year over year

Prior to the introduction of Galbincea’s model, a key failure point in supply chain management was the absence of optimizing inventory and procurement strategies dynamically based on business needs. Now with much of the process automated – and a dynamic model created – order metrics for thousands of MRO items can be updated in minutes compared to weeks. Better data tools support more accurate and optimized purchasing practices. The system now processes a buy with a lead time calculation that, over time, adds to supply chain knowledge by continuously refining the algorithm.

“The model is living and breathing – it keeps getting updated along with the data,” Galbincea says. “Lead times changed this year with demand so high and supply still being choked because of labor, inflation, fuel costs, transportation, and freight… And [with the model], boom – all those numbers are updated instantaneously.”

In its second phase, Galbincea is developing a more advanced reliability model that will enable forecasting and optimize a preventative maintenance schedule and related purchasing. “That's where things start getting really cool,” he says.

There will likely be more AI and machine learning involved in this next phase, as failures are not always linear from a data perspective. “When you can correlate preventative maintenance with your buying schedule, then you can really optimize risk and what inventory is needed to support manufacturing.”

Automation enables engineers to invest their time in more value-adding tasks

More even than a reduction in procurement budget, Galbincea’s innovation means TimkenSteel is also saving on engineering labor hours that were previously devoted to repetitive manual tasks. Whereas queries were previously one-offs, standard scripts in JMP now enable engineers to query the data as frequently as they need without investing additional time.

“Before the MRO inventory model, an engineer would query and massage data manually to set an opinionated reorder metrics. Now it's all automated, so you can see how [this project] has been proof of concept,” he says, adding that achieving TimkenSteel’s current level of optimization manually would have been so time-consuming even for a team of engineers that it would not have added value.

“It's not just in the dollar amount [that we see value from the model],” Galbincea explains. “It's also freeing people up to solve other problems and create value through new projects. When you free up that employee's time, they can work on more value-added tasks.”

“If you can automate it, there's no excuse to do it manually”

In transforming supply chain management, Galbincea and his advocates at TimkenSteel have perhaps hit upon the next industrial revolution – one that involves the entire business enterprise. “Ultimately, if you can automate it, there's no excuse to do it manually,” Galbincea concludes. “That’s Business 4.0: digitize the entire enterprise and you’ll be able to analyze almost any facet of your business and see potential optimization opportunities. You’ll see where you should focus [time and investment], and you'll get the most bang for your buck.”

The results illustrated in this article are specific to the particular situations, business models, data input and computing environments described herein. Each SAS customer’s experience is unique, based on business and technical variables, and all statements must be considered nontypical. Actual savings, results and performance characteristics will vary depending on individual customer configurations and conditions. SAS does not guarantee or represent that every customer will achieve similar results. The only warranties for SAS products and services are those that are set forth in the express warranty statements in the written agreement for such products and services. Nothing herein should be construed as constituting an additional warranty. Customers have shared their successes with SAS as part of an agreed-upon contractual exchange or project success summarization following a successful implementation of SAS software.