How Math is Becoming Cool in Supply Chain — Again!
When I started my career in the late 1990s, the supply chain was a cool place for math geeks! As a discipline, supply chain planning was not really that big in the universities. The traditional pathway into supply chain was through a major in operations research, industrial engineering, or business — often with each offering courses in their silos.
Of course, some managed to sneak in because of their experience in mathematical modeling, as algorithms were all the rage! I fell into the latter category. Or… let us just say someone believed I did.
We tackled some of the most complex and hairiest supply chain problems using mathematical techniques. Statistics, optimization, and heuristics-based approaches were gaining traction then.
User experience, either as a term or as a concept, was unheard of. As we were building algorithms, we shipped the software on a CD (or a bunch of CDs). The software did come with a user interface. However, for the most part, the users skipped the UI altogether, having their IT organizations write custom user interfaces while embracing the algorithms that, we, the vendors, built and shipped.
The drive towards modernizing IT infrastructure in the wake of Y2K helped us tremendously. Life was good!
Then the tide started turning. The economy hit some headwinds which worsened with 9/11 and the Enron scandal. A wave of consolidation followed in the supply chain software industry which resulted in vendors acquiring a mixed bag of technologies that did not speak to one another.
Most of the vendors’ time was spent on rationalizing and integrating these technologies, the success of which was mixed at best. Lots of time and precious R&D dollars were spent on providing a unified frontend through a common UI while in the backend, the data and algorithms sat on islands. Algorithmic innovation, for the most part, took a backseat.
Scalability was also a big challenge, causing us to artificially partition a supply chain model into horizontal and vertical slices. That resulted in anomalies. The physical supply chain and the digital supply chain representation were never in sync. The digital model at best was an ugly patchwork of a bunch of systems cobbled together. The user community’s skepticism grew as they had to spend more time reconciling algorithmically generated results with their views of the world.
Besides, as the world started becoming more dynamic, the user community could no longer operate their supply chains based on a single plan recommended by a supply chain analytics algorithm. They had to prepare for a range of scenarios, and most supply chain systems struggled to enable this scenario planning.
This resulted in user disenfranchisement with algorithmic approaches, resulting in Excel as the system of choice for most planners.
However, all this is starting to change. I am seeing a renaissance in algorithms in supply chain. A confluence of factors is contributing to this. Let us examine these.
1. Rising Complexity and Volatility Require Algorithmic Intelligence
The pen-and-paper models or Excel bypasses miss out on more opportunities as supply chains become complex and volatile and as supply chain disruption has become a part of everyday life.
As supply chains are faced with accelerating disruption, rising nationalism, explosive growth in the SKUs, channel complexity, rising consumerism, and calls for sustainability among other factors, they will need to make supply chain decisions in an increasingly holistic manner as opposed to operating in silos. Multiple tradeoffs need to be considered in arriving at decisions that do not have unintended consequences.
Making decisions in Excel sheets and manual methods now results in blind spots and missed opportunities.
2. The Rise of Cloud Computing
While the concepts of neural nets, unsupervised learning, and such have been around for many years, cloud computing is making it feasible to unleash these algorithms on massive amounts of data, augmenting human intelligence.
There is an increasing acknowledgment within the organizations of the need to harness the power of the data that their systems and increasingly digitized products and services generate. As more devices and products turn digital, companies are building massive troves of data assets that can be mined and monetized. If data is the new oil, then algorithms are the heavy machinery needed to extract information.
3. Money Speaks
Algorithms now surround us 24/7, hidden in plain sight in the form of smart devices. Uber, Netflix, Google, and Airbnb are all examples of companies that operate in an asset-lite mode while harvesting the power of data through algorithms. These companies achieved massive valuations in a record amount of time, created entirely new markets and delivery mechanisms in their respective domains.
This is attracting plenty of venture money and inspiring entrepreneurs to extend the power of algorithms to the enterprise software domain, helping companies monetize their data assets. Accordingly, careers in data science and machine learning are turning out to be quite lucrative. If algorithms are the heavy machinery, the data scientists and engineers are the makers of the machinery!
4. A Better Appreciation of Algorithms Amongst the Young Professionals
I have an opportunity to interact with and — from time to time — mentor students in supply chain programs in universities. It is not uncommon for these youngsters to take coursework spanning industrial engineering, computer science, applied math, and business, often opting for a dual major. When coupled with intern/co-op experiences, these digital natives are entering the workforce with a much better appreciation of analytics and algorithms than those of previous generations.
5. The End-to-End Digital Representation of Supply Chains is Now a Reality
Thanks to Moore’s law, storage and processing power has grown geometrically over the last two decades. Cloud computing continues to bring unparalleled scale along with metered pricing, helping tackle large-scale computational problems much faster than before and at a very affordable cost.
This has emboldened disruptive companies to bring to life large end-to-end reference models of supply chains. They connect all the nodes including sourcing, production, distribution, and last-mile delivery, essentially creating a digital twin that brings the key supply chain information to the executives’ fingertips in a highly interactive and visual manner. Organizations no longer need to cut corners.
6. The Democratization of Algorithmic Intelligence Through Hyper-Personalized User Experience
The algorithms powering consumer-oriented platforms manifest as easy-to-use consumer-facing apps. A lot of the mathematical complexity and horsepower is hidden behind a simplistic user experience. We don’t need a user manual to understand how Netflix works.
Such simplicity is now entering the enterprise apps with a hyper-personalized user experience.
Emerging technologies are placing the power of algorithms in the hands of, what Noha Tohamy of Gartner refers to as “citizen data scientists.” These citizen data scientists can tap into emerging platforms that give them the ability to blend data from disparate sources in a visual environment, access to an algorithm library, and a zero-code visual environment in which they can rapidly develop and deploy apps at an enterprise scale, thus democratizing the power of algorithms.
Looking Forward to the Future of Supply Chain
I am very encouraged by what I am seeing. As math becomes cool (again), it does open up some very exciting opportunities for those of us in the supply chain space. Life is good — again!