Supply Chains are so damn interesting to me.
To design a machine, you need to be technically smart. To design a supply chain, on the other hand, you need to be wise.
Let me explain.
If you're designing a circuit, applying basic Kirchoff's Law and RLC calculations will tell you everything you need to know about the circuit's behaviour. Likewise, if you're designing a truss for a bridge, material properties and load calculations like stress, strain, yield, moments, etc. will tell you how the structure will behave in reality.
You can design these systems and as long as all your calculations are correct, they will work as predicted without too many hiccups.
These are static systems; their performance can be simulated really well.
This isn't true of supply chains.
Because supply chains are dynamic systems. Even when you design and optimize individual stages and put them together, the system as a whole can behave differently when it interacts with the complexity of the real world. Unknown variables will come into play and start affecting the system in meaningful ways. Lots of hiccups will have to be work around, stuff will have to be mended on the fly, and the system will soon assume a life of its own. It will change and evolve with time.
Recently, I wrote a two-part essay on how buildings learn and grow over time. The same applies to supply chains, and everyone — right from the retailers and distributors to the manufacturers and wholesalers — is continuously adapting their strategy based on changing market conditions.
A supply chain develops many emergent properties as a part of its functioning. And all the tacit and contextual information that makes any supply chain run efficiently is locked within the hundreds and thousands of people involved in getting a product from the factory to the end consumer. It is locked in the thousands of decisions every actor in a supply chain makes when it comes to sourcing, buying, and selling decisions, which in turn are fully influenced by the surrounding market and geopolitical environment.
So, when it comes to dynamic systems, Dynamic Thinking is needed, in contrast to Static Thinking.
Static thinking leads people to focus on particular events in isolation. Dynamic Thinking, however, sees problems as unfolding over time — as if setting in motion a chain of dominoes with far-reaching effects, many of which may not be easy to predict.
Today, we are going to discuss a prime example of one such dynamic problem supply chains face, which will serve as a good entry for you into the world of complex systems thinking.
It's called the Bullwhip Effect.
Let me prefix this discussion with a small business case study that will put things in context.
In the late 90s, when no one could build out network capacity fast enough, Cisco was the industry titan that made the underlying gear at the heart of networks being built by many service providers and installers. Things were going so fast that multiple telco companies were bidding for the same network projects.
So, if four companies were all bidding on a network install, they all told Cisco they needed gear. But in reality, they were all competing for the same network project so there was only really one network that needed Cisco's gear.
But to Cisco, these orders looked like 4x the volume of what was really needed by the market. So, what Cisco ended up doing is it ended up manufacturing tech according to this flawed sense of demand. And since things were moving so fast, it would take just 6 months for the tech to become obsolete. The result was that as of August 2001 when the telecom bubble popped and the stock market crashed, Cisco was left holding $2B+ worth of excess inventory that was now already obsolete.
This is just a small glimpse of what incomplete information can do to individual players in a supply chain.
In our example case, Cisco only had visibility on the orders being placed by network companies, but it didn't know that all of those orders were ultimately being placed by companies competing for a single network project.
Let's consider a more relatable example.
Say a major sporting event is coming up on Sunday. The bar owner senses a spike in demand and hence places a larger-than-normal order of beer on Thursday to stock up for Friday because they think people will start prepping for the "big game."
All other bar owners and places selling beer do the same, for the same reason. As a result, the distributor gets hit by several large orders at once. So, the distributor fulfills whatever orders it can and similarly places larger-than-normal orders on the brewery so that this same inventory shortage doesn't happen on Friday. But there's more than one distributor too, and they're all doing the same thing, so the brewery gets slammed with a yuuugeee pile of orders, which they can't fill either, so they fulfill as much as they can and start brewing even more beer so that this doesn't happen to them again on Saturday.
But when all the hype around the Sunday game has died down on Monday, the retailer is back to their normal stock but has pissed off some customers because they ran out of beer over the weekend. The distributor has a huge overstock because the Monday orders are back to normal. And the brewery is sitting on a giant pile of beer with no customers.
In essence, what happened here is one small action set off a series of related actions that had a pronounced effect up the supply chain.
A small increase in demand on the retailer's end led to a higher increase in demand on the distributor's end which led to even higher demand on the manufacturer's end.
This is called the Bullwhip Effect.
A small glut or squeeze in demand can lead to huge spikes and volatility up and down the supply chain. The result is that a lot of players in the chain face a massive shortage of inventory while other players have huge sums of cash now stuck in excess inventory that they can't sell. A liquidity crisis ensues.
In basic control systems theory, it is known that delays and missing feedback loops between different parts of the system make systems more unstable.
As every stage in the chain doesn't have full information about the status of other stages in the chain, it is working with incomplete information based on the limited data it is receiving from the market and adjacent stages one level up and down the chain.
As a result, what ends up happening is that the farther up the supply chain you are (the farther away from the end user), the wilder swings in orders you see. This happens because each intermediate step in the chain tends to aggregate orders and work on a longer lead time, so their buffers and fear of running out of stock get larger.
The basic flow goes as follows:
- Retailers sense a shortage of inventory and wish to plan ahead.
- To plan ahead, every retailer places orders for larger-than-normal inventory.
- The manufacturer up the chain gets flooded with orders.
- To meet demand, the manufacturer increases manufacturing capacity to fulfill these orders.
- By the time it starts fulfilling the orders, multiple things can happen: the market demand for the thing can wane, which can lead to many retailers canceling their orders.
- Or, retailers can get stuck with excess inventory, post which the manufacturer faces a sudden large dip in orders.
The latter happened to Motorola during the Christmas shopping seasons in 1992 and 1993. Motorola could not meet consumer demand for handsets and cellular phones, forcing many distributors to turn away business. Distributors, anticipating the possibility of shortages and acting defensively, drastically overordered toward the end of 1994. Because of such overzealous ordering by retail distributors, Motorola reported record fourth-quarter earnings in January 1995. However, once Wall Street realized that the dealers were swamped with inventory and new orders for phones were not as healthy as before, Motorola’s stock tumbled almost 10%!
Let me pause for today. But we aren't done yet.
In upcoming essays, we will discuss the factors that lead to the Bullwhip Effect and what strategies companies have employed to mitigate this problem.
(Keiretsu Jutsu, anyone?)
