Supply chain disruptions are nothing new. They can be caused by a wide variety of events, ranging from extreme weather to labor conflicts. However, 2020 and 2021 have seen disarray in global supply chains at an unprecedented level. Brexit, the COVID-19 pandemic, an ongoing semiconductor shortage, and then the blockage of the Suez Canal by a massive container ship have had catastrophic results that still echo throughout the global supply chain. Events like these will continue to happen, confounding sourcing professionals. However, there is a way to minimize the damage caused.

Predictive sourcing is the key to minimizing the impact of supply chain disruptions. Made possible by advances in data analysis and machine learning, predictive sourcing leverages the reams of data that previously hid valuable information. With predictive sourcing, procurement gains end-to-end visibility and sustainability. This technology provides resilience that allows organizations to successfully navigate supply chain disruptions. In many cases, predictive sourcing can anticipate classic disruption scenarios (such as popularity-based shortages), helping sourcing teams avoid the crush altogether. In short, predictive sourcing is poised to revolutionize procurement.

Eric Buras is the Head of Data Science and Machine Learning at Bid Ops, a leader in AI-powered procurement software. With a background in mathematics, Buras started his career in cybersecurity before moving to the logistics side of supply chain management. He worked on machine learning models used to track the location and progress of container ships and delivery trucks. The vast amount of data involved in procurement offers both opportunity and challenge. That is the kind of situation that attracts someone interested in complex problems in data science. Buras notes:

“I think procurement is one of the areas where there’s still a lot of impacts to be made. Bid Ops’ purpose here is to really shoot for the moon in terms of changing the way big companies do their procurement.”

According to Buras, the events that were so disruptive to sourcing in 2020 and again this year may have people thinking they were once-in-a-lifetime occurrences. However, that is wishful thinking.

“These big events, they seem like random, rare events to us, but I think they’re going to be more and more common, just because of the way the world works. Hopefully, not another pandemic, but there are always going to be events like Brexit, oil shortages, or ships getting stuck in canals. Those are just human issues.”

Because supply chain disruptions will continue to happen—and perhaps on a larger scale than ever before—the old way of dealing with them isn’t going to work. Companies have traditionally dealt with these disruptions reactively. The eventual realization that there is a problem leads to scrambling to figure out exactly what is going on. That’s followed by a race to source an alternative supply. This approach is riddled with problems, not the least of which is the challenge of trying to lock down alternative suppliers when competitors across the globe have come to the same conclusion and are trying to do the same thing. The key is to be ahead of the disruption and use a predictive sourcing strategy instead of being reactive.

Buras has thoughts on what this new approach entails:

“Instead of reacting to these events, companies are going to need to be proactive, having supply on hand, or maybe just a plan B for sourcing locally. So, just getting a hold of all their historical purchasing data and diversifying their purchasing is really important, because you don’t want to run into these global event shortage issues again.”

For many companies, this is easier said than done. Buras says he frequently sees procurement teams that are still trying to track their sourcing data locally; on spreadsheets and in email. Those files are typically going to be spread across multiple computers, multiple locations, and in multiple formats. It’s all but impossible for human analysts to make sense of this much data, let alone do so under a time crunch. Even advanced machine learning technology can’t spit out accurate analysis in these conditions. So step one is moving all that data to a centralized location in the cloud. That way, everyone on the sourcing team can access it. As part of the process of moving the data to a single repository, the data must be cleaned up and normalized.

In these conditions, machine learning models are extremely effective.

Bid Ops provides that centralized platform and data repository. Buras points out that companies adopting Bid Ops then benefit from low-hanging fruit, including data visualization. From there, basic analytics identify important metrics like supplier diversity. As this data is incorporated, Bid Ops then shows the power of machine learning and artificial intelligence, supporting the goal of predictive sourcing. Buras explains the transition starts with the computer finding patterns but then progresses to the point where Bid Ops can do the heavy lifting:

“As the price from one supplier increases, we predict that the price from another supplier is also going to increase, but historically doesn’t rise as much as the competitor. So, go talk to the alternative supplier, and we predict you’ll get a better price there.”

While pressure is always on sourcing teams to reduce costs, there are other factors at play. Strategic issues, including sustainability, environmental impact, and carbon footprint, are increasingly important. Predictive sourcing can be applied here as well. Whether a procurement team is being measured by cost savings, the impact of their decisions on a company’s carbon footprint, or a combination of factors, being reactive just won’t cut it. Bid Ops has been designed to fully support predictive sourcing:

“Letting a computer predict the price and make a recommendation and just reduce manual effort is the gold standard we’re shooting for. Enabling their teams to focus on the most strategic sourcing initiatives and letting a computer handle lower value POs or more tedious transactions is what we’re trying to accomplish.”

At the end of the day, while Eric Buras works for Bid Ops and would obviously love sourcing teams to adopt the product he’s put so much into, he has a wider goal:

“More people working in this space is better, even if they’re competing with Bid Ops. More visibility into what can be done in this space will help all of the companies competing here… ‘A rising tide lifts all ships.’ Just having more and more work in this area will enable all of our customers to see the value in it and lead to better things for all of us. So, I’m just trying to increase visibility in the analytics and data science space in procurement. Because I want other smart people to work and help improve procurement.”