It’s amazing to think that, for almost 50 years now, the supply chain management profession has embraced sales and operations planning (S&OP) as a method for balancing supply and demand in complex networks. Along the way, there have been significant challenges associated with both developing and sustaining the process. But with so many exciting and innovative new tools emerging in our increasingly global and digital landscape, the future looks bright for the next generation of S&OP.
Unfortunately, according to Gartner’s four-stage, higher-maturity, S&OP model, nearly 70 percent of companies are stuck operating in the reacting and anticipating phases. These areas are most often centered around making a plan and maintaining a regular S&OP meeting in order to balance supply with demand for the good of the enterprise. Consequently, they are totally focused on inward processes and tend to require at least four years to solidify.
To address the uncertainty, complexity and risk that come with global, digital supply chains, it’s time to make the jump to the more mature S&OP stages: collaborate and orchestrate. Collaboration involves expanding the S&OP process to a company’s suppliers and customers; orchestration describes a process that is driven by demand sensing and shaping, as well as enterprise trade-offs such as risk-reward analyses. In other words, the focus shifts to profitability.
The present global, digital landscape is ominous. According to the “10 Key Marketing Trends For 2017” report from IBM Marketing Cloud, 90 percent of the data in the world today was created in the last two years alone, at 2.5 quintillion bytes of data a day! One quintillion bytes of data is equal to one billion gigabytes. That’s overwhelming enough. And yet, it will continue to grow into something even more astounding, called a yottabyte, in the next five years. A yottabyte is equal to one septillion, or a million raised to the seventh power, bytes. This is theoretically massive enough to fill one million datacenters and will cost $100 trillion — more than the combined gross domestic products of the entire world.
The best way to look at and better understand big data is to dissect it into four V’s: volume, variety, velocity and veracity. This breakdown helps clarify what big data is really all about:
- Volume: Big data is increasing exponentially, both for people on a personal level and within our enterprises and supply chains.
- Variety: Many new forms of data now exist. The information comes from computers, phones, photographs, texts, the internet, streaming video, sensors in everything and much more. Eight years ago, about 80 percent of all data was structured, similar to rows and columns in relational databases. That is the case no longer. Today, 80 percent of all data is unstructured, free-form chaos.
- Velocity: The speed with which data is now passing through all of our devices is increasing exponentially. Most existing systems simply cannot handle the pace.
- Veracity: In the new global, digital landscape, data is imprecise, amorphous and therefore must be vetted for reliability before it’s possible to use it to make informed and efficient decisions.
Map to minimize risk
Because today’s networks have an escalating number of nodes, there’s a much higher probability of a risk event occurring. Supply chain visibility thus has become a critical part of top-notch execution and solid risk management. Visibility is key, and this is best achieved by digitally mapping supply chains.
To produce an effective map, start small with a product line maintaining a couple of stock keeping units. Graph suppliers, then plot manufacturing plants or contractors. Next, add distribution centers, followed by any warehouses — both the company’s and its customers’. Finally, map the customers themselves.
When the map is complete, connect the dots to display the flow of material. Next, use this information to identify risks that have occurred or might occur. Put the risks on the actual supply chain map, and assess them in order to determine the most impactful risk-mitigation strategy. At this point, it’s also possible to run what-if scenarios that can help evaluate how a supply chain will react to certain stimuli. This is known as stress testing.
Make the jump
Big data, mapping and stress testing are a great place to start. However, to cope with mounting uncertainty, complexity and risk, additional resources are needed. Predictive analytics, probabilistic modeling, programming languages such as Python and R, the open-source software utility Apache Hadoop, and the open-source computing framework Apache Spark are some of the available tools supply chain management professionals should familiarize themselves with. (See sidebar.) These solutions enable users to
- digitize their supply chains
- better handle uncertainty
- more effectively read and make sense of unstructured data
- quantify how their supply chains will react to challenges
- develop statistically strong patterns associated with supplier and customer buying habits and sentiments.
Perhaps most importantly, these exciting tools can successfully support the jump to the more mature, outside-in S&OP stages. Indeed, as they enable users across all kinds of networks to accelerate learnings and enhance the precision of next-generation decision-making, the face of S&OP will change dramatically — and a lot sooner than many professionals may think. And with more and more supply chain management professionals leveraging these tools and embracing next-generation S&OP, companies will benefit from progressing through the predictive analytics maturity model:
- Phase 1: Descriptive (What happened?)
- Phase 2: Predictive (What might happen next?)
- Phase 3: Prescriptive (What should I do about it?)
- Phase 4: Cognitive (The system learns.)
Along the way, the objective must be for supply chain management professionals and the networks they guide to become better positioned to master complex decision-making processes. As S&OP continues its dramatic shift — from a structured, linear tool to a more ad-hoc, event-driven process — it will be the ideal support mechanism for these strategic imperatives. Supply chain management professionals would be wise to maximize innovations that help them achieve high-frequency, high-impact decision-making and accelerated learnings across the enterprise.
SIDEBAR: The S&OP Innovations to Watch
Predictive analytics involves mining information from data sets in order to identify useful patterns and forecast the likelihood of future events. As increased supply chain complexity puts pressure on even the most well-run S&OP processes, accurate demand planning via predictive analytics is a valuable advantage. Supply chain management professionals can benefit by using predictive analytics to analyze and consolidate data from all areas of the enterprise in order to make informed, effective, real-time decisions.
Probabilistic modeling involves statistical analysis of historical data in order to estimate the chance of an event occurring again. In S&OP, it makes it possible for supply chain management professionals to identify unknown risks and develop contingency plans.
Python and R are open-source programming languages. R is most often used to solve statistical problems, machine learning and data science. It also offers packages that enable users to perform time series analysis, panel data and data mining. Python offers a more wide-ranging, data science methodology. Both can meaningfully enhance the S&OP process.
Apache Hadoop and Spark can process large data sets with simple programming models. These solutions can create more accurate analytic decisions in response to the internet of things, artificial intelligence, the cloud and mobile. Hadoop also helps users optimize enterprise data warehouses, drive better decision-making and cut costs by moving “cold” or siloed data to a data lake.
Gregory L. Schlegel, CPIM, is founder of the Supply Chain Risk Management Consortium, executive-in-residence at Lehigh University, and an adjunct professor of enterprise risk management at Villanova University. He is also the former S&OP process owner at three Fortune 100 manufacturers. Schlegel may be contacted at firstname.lastname@example.org.