The phenomenon of demand amplification in the supply chain has been recognized since the early 1900s – long before Dr. Jay Forrester coined the term in his book Industrial Dynamics in 1961. Over time, it has been called the “whiplash”, the “lead-time effect” and, since 1997 as the “bullwhip effect”.
The causes of the bullwhip have been well documented (these are Forrester’s terms with my paraphrased definitions):
1) Demand signal processing
Basically, this means that there are time delays (lead-times) with in-transit quantities and inventories between supply and demand for any node in the supply chain. Because of these delays, the supply chain requires the use of forecasts for planning. The inaccuracy of these forecasts (up or down) will cause imbalances in supply and demand which lead to large corrections downstream as the passage of time plays out.
2) Order batching
Three words: economies of scale. Even though only 10 units may be required, it’s much more efficient from a handling point of view to pick and pack a case of 12. A customer may need a half truckload today and another half truckload tomorrow, but it’s much more cost effective to consolidate both shipments onto a single load. As this batching behaviour works its way back through the supply chain (e.g. consumers buy one at a time, the store orders in cases of 12, the DC orders in pallets of 288 and the manufacturing plant ships in truckloads of 2880), product movement becomes more infrequent and “lumpy”.
3) Rationing and gaming
Whenever there’s a shortage (or a perceived shortage) or a supplier is chronically unreliable, people will order and inventory more than they need earlier than they need it to hedge the risk.
4) Price fluctuations
Drop the price and demand will spike during the time that the price drop is in effect, particularly if everyone knows that it’s a “limited time offer”.
Just looking at these root causes can be a bit disheartening. While there is no question that work can be done on them to somewhat lessen the effects of the bullwhip, it’s virtually impossible (that’s a word I try to use very sparingly) to eliminate it altogether. Let’s look at them individually:
1) Demand signal processing - Fat can be trimmed from the administrative portion of lead-time in most supply chains, but we’re still bound by the laws of time and distance. This point is especially poignant to retailers who source product from the Pacific Rim. There’s simply no cost effective way of getting product from China to the West in 24 hours.
2) Order batching - You can certainly negotiate smaller minimum order quantities and pack sizes from your suppliers, but this can often result in higher acquisition and handling costs that serve to negate the inventory savings. At the end of the day, the only way to fully defeat this bullwhip-causing constraint is to sell, distribute, transport and manufacture in “onesies”. Not likely to happen in our lifetime, and with good reason.
3) Rationing and gaming – Okay, you got me on this one. This “constraint” is all negative with no positives and, over time it can and should be completely obliterated.
4) Price fluctuations – To the extent that this activity does nothing for the consumer at the back end of the supply chain, it should be minimized if not stopped altogether. However, on the retail end, I have yet to find a company where the marketing department decided to abolish promotions because it makes the supply chain department’s life more difficult.
Okay, so 3 of the 4 main root causes of the bullwhip effect can, at best, be reduced but not completely eliminated. In other words, as supply chain professionals, we’re all doomed. Unless...
If you can’t “whip” it, model it!
Maybe it’s time to start thinking about the bullwhip effect a little differently. Given that it will likely always exist in the supply chain, despite our best efforts to mitigate it, perhaps we should instead turn our attention to attacking its ability to do serious harm. The existence of the bullwhip is a problem, but what makes it really damaging is that the timing and severity of it are unknown until it hits you.
If the bullwhip-causing constraints between the consumer and the factory in the supply chain are known (and they are), they can be explicitly modeled into the plans. The bullwhip still exists, but the timing and severity can be predicted with a high level of accuracy well into the future, because the actual constraints that cause it in the first place are being used to predict it.
On April 12, 1912, 11:40 PM – in the middle of the night – the RMS Titanic hit an iceberg in the north Atlantic. If she had approached that same iceberg at the same speed with the same captain at the helm a mere 8 hours earlier, she never would have sunk.
Sometimes the difference between a minor course correction and a disastrous tragedy is the ability to see it coming. And it’s a lot easier to steer around an iceberg than it is to completely melt it.
