Back to overview

Taming Supply Chain Bullwhips with Artificial Intelligence

By Mimin Adhikary - Jul 1, 2022 10:30:36 AM

Bullwhip effect is a common term used in supply chain to describe how a small change in consumer purchasing pattern, or, for that matter, a change in any demand pattern can cause significantly large impacts to supply. It is an analogy that transcends from the pattern in which a cowboy slowly turns the whip until it gathers enough momentum to crackle and emit a sonic boom which ‘tames’ the cattle it is intended for.


A turbulent climate for global supply chains


In the recent days, starting at the onset of the pandemic, and accelerated with the war, the world has had many more examples of classic supply chain bullwhips compared to any other time in recent history. We had the infamous toilet paper outage that began with consumers detecting small shortages of paper rolls in store shelves triggered by paper shortage at the manufacturing companies. This, in turn, led to hoarding behaviors, exacerbating the shortage, signaling manufacturers to produce more. The paper shortage recovered fairly quickly within 2-3 months of the start of pandemic, but the nationwide hoarding behaviors lasted half a year into the pandemic triggering excess production and at the end, a huge surplus of toilet paper rolls in the United States.

Macro Economical Context contributing to the bullwhip effectThere is a similar problem brewing beneath the surface right now, with the over-abundance of global freight containers and slowing demand of freight trucks. Retailers across the world have surplus inventory of nearly every kind of consumer goods – clothing, electronics, housewares. So much so that some big retailers are allowing customers to keep their returns instead of adding to the already excess inventories. Combined with the high oil prices due to the war, and softening demand for freight trucks, the high demand for freight containers across global and local supply chains in waning. This is from the bullwhip effect of the pandemic, at the onset of which consumers limited purchases to essentials. With all the ensuing demand once vaccines became available, manufacturers overproduced inventory causing excess stock. All of this is happening at a time when consumer sentiment is not very positive from the speculation of impending recession. Each one of these events have a ripple effect, contributing to the bullwhip.


Early detection of anomalies in demand with AI

Stepping away from the granularity of these supply chain breakdowns, what can we do to prevent, or better control these bullwhip effects in the future? Adoption of artificial intelligence to build demand forecasts is a very powerful, less explored solution to control the bullwhip from ballooning to proportions hard to manage and costly to control. AI models have evolved significantly over the last decade, owing to the availability of relatively inexpensive cloud data storage which makes it possible for organizations to collect, process and store useful data at micro levels which Excel-based tools lack as capabilities. These models can detect subtle anomalies in consumer sales data a lot faster and with many-fold improved accuracy. Therefore, if the demand for toilet paper ever goes up unusually again, AI will make it possible to trigger demand forecasts that take into account the short-term effects of hoarding behaviors and normalize longer term demand for realistic consumption, preventing over-production and over-stocking.

Traditional demand forecasting uses historic sales data to model future sales. Historic data, based on events of the past 28 months, is full of exceptions, misleading signals and more volatility in every aspect of life than the world has seen in the last several decades. Hence, to factor in last two years of historic sales, without carefully interpreting short-term consumer behaviors or orders being placed by B2B businesses would be irresponsible and short-sighted. AI, when augmented with Actionable Demand Sensing, is a powerful platform to help companies build proactive supply-chains.


Key Takeaways: AI for planners and supply chain professionals


In summary, AI in demand planning can help significantly by:

  1. Catching macro & micro-trends in consumer behaviors quickly and accurately
  2. Removing the sole dependency on historical sales in the face of volatile business landscape
  3. Reducing reaction time to adapt to drivers of change in demand, therefore directly impacting the bottom line of organizations

Even if AI in supply chain management is still in its infancy, there are proven AI solutions in the market that consistently raise the bar for planning functions. If you have not explored them yet, no time is better than now.


Mimin Adhikary

Chief Product Marketing


Get in touch with Garvis.

follow us