The Bullwhip Effect Revisited

The academic journal Management Sciences recently developed a list of its 50 best articles of the last ten years, and among the winners was "Information Distortion in a Supply Chain: The Bullwhip Effect", published in 1997, by Hau Lee, V. Padmanabhan and Seungjin Whang, all from Stanford University.

I thought the timing of this recognition, combined with high expectations around RFID and its potential to reduce supply chain information latency, and general emphasis on more "demand-driven" supply chains (e.g. AMR's demand-driven supply networks concept), meant it was worth revisiting Lee's seminal research.

The basic concept, if not the details, should be familiar to most supply chain professionals. Relatively minor variations in end consumer demand result in increasingly wide order swings back up through the supply chain, which when graphically represented, resemble the rolling shape of a bullwhip in action. The result: excess inventory buffers up and down the chain.

Lee identified four sources contributing to the Bullwhip Effect:

	Demand signal processing: Use of recent demand to generate a forecast, which tends to exaggerate both high and low swings, especially as that forecast is propagated back up the supply chain.
	Order batching: The end retailer/distributor combined potential smaller orders to gain efficiencies in administrative costs, volume-pricing opportunities, and/or transportation savings.
	"Gaming" of tight supply products: Buyers at each level "over order" to ensure supply or mitigate against allocations.
	Price variations: Deals offered at any level promote orders that exceed true demand, followed by long periods of no orders while true demand continues.

Individually and together, these factors cause spikes and valleys in orders in the face of relatively consistent and predictable end consumer demand. This basically confuses various members of the supply chain as to the true state of affairs, causing them to produce and stock goods according to the spikes, resulting not only in excess inventory, but often more manufacturing and distribution assets than should be required.

The solution? The key is passing true end user demand back up through the chain, and then using this information to drive continuous replenishment strategies. Also, significantly reduce the use of pricing promotions, and be more collaborative around actual requirements and production capacities for short supply products. Voila, most of the problem is solved.

OK, so much for today's lesson. Quiz tomorrow.

So, how far have we progressed? Continuous replenishment and VMI have clearly become mainstream, but are from universal practices. "Point of Sale" data is reasonably available and modestly used in retail-consumer goods, but rarely exploited in other industrial supply chains. Use of trade promotions and "forward buying" have abated a bit, but are still prevalent. "Collaboration" is still in the very early stages in most trading relationships.

The real question is why we haven't made more progress. It's why a lot of us are asking whether RFID is really the answer to solve many of these problems, as it is often positioned, or if it's almost a distraction in moving towards these goals.

With many observers concentrating on the manufacturer-distributor interface, I think two of the most overlooked factors are the actual changes in both supply chain processes and manufacturing systems required to actually take advantage of improvements in information flow and collaboration. We've seen this theme explored for example, in pieces we've recently run, citing thoughts from Scotts Company's Sumantra Sengupta on the huge organizational change required to move to a pull-based model, and P&G's long-term initiative to radically improve manufacturing flexibility.

Seven years later, to get rid of the Bullwhip Effect, it's "back to the future". We've known how to get there for some time, before anyone was even thinking about RFID. In the end, it's about senior company leadership deciding they are going to transform the way their supply chains work.

Do you think we've made real progress in the past seven years on reducing the Bullwhip Effect? What are the real barriers to further progress?
Let us know your thoughts.

Changes at the Top for i2 and Manugistics
View i2 Press Release>>

View Manugistics Press Release>>

Coincidentally, two of the supply chain industry's leading solution providers, i2 and Manugistics, announced changes at the top.

i2 Chairman and CEO Sanjiv Sidhu announced his company had contracted with a well-known executive search firm to find a new CEO. Sidhu will move back to just being Chairman, a position he maintained during the company's years with Greg Brady in the CEO role.

Manugistics announced it had hired supply chain industry veteran Joe Cowan as CEO, with current CEO, Greg Owens, also moving to the Chairman's spot.

We'll let the financial and software industry analysts sort out what it really means. Obviously though, these two companies, which have been so associated with the evolution of supply chain management, are entering a new phase. Both have also encountered some financial challenges, some of their own making, some the result of significant changes in the technology buying climate.

i2 appears to be stabilizing. Manu had been on a bit of an upswing, but the stock took a hit after an unexpectedly tough most recent quarter, which probably led to the move.

Supply Chain management is really still in its infancy. The future can be bright for both companies, though each will have to find a way to prosper in a new type of market. In the end, that's probably what these moves are all about.

How can software companies achieve financial success in this different type of customer buying climate? Do you think the future can be bright for most supply chain software companies? Let us know your thoughts.

Feedback of the week - "On Software Prices in the Toilet - Is this a good thing?"

Your commentary on the erosion of Supply Chain Execution Systems (SCES) pricing is right on the money! As an independent observer of countless software sales pitches, a plethora of live demos, and an abundance of closing tactics, I must say that many of those witnessed over the past 24 months have been almost laughable (if they were not so pathetic). One SCES provider recently competed with itself by dropping its price three times in succession when my client failed to return a phone call for a two week period. When my client returned from his vacation, he literally giggled sadistically while listening to the three desperate voicemail messages from the sales rep dropping the price by more than $150,000.00 in total.

The competition between best of breed supply chain execution solution vs. "fully-integrated" ERP offerings does continue to fuel the price war fire and yes, the market is a tighter one. But it is my experience that "ugly" deals (not mutually beneficial) spawn "ugly" implementations and "ugly" business relationships and often "ugly" lawsuits.

My advice is simple (if anyone wants it):

Functionality rules! First and foremost, the systems under consideration must be a good functional fit. Assuming this is true...

Software vendors should put their best offer on the table the first time; one that is fair and reasonable; and one that provides value to both parties. Dropping prices without altering the nature of the products and services quoted erodes credibility. Of more dire consequence, dropping prices below the threshold of profitability is a recipe for disaster. In this case, there is a high probability that when the sales team leaves, the implementation team will be looking to shave dollars from the project cost or seek additional revenues (both come at the expense of the client).

Buyers of supply chain execution software systems should realize that they are not buying a disk with a program on it. Buyers should be prepared to enter into a mutually beneficial long-term business relationship with a trusted vendor capable of providing ongoing support as their business evolves. This cannot be viewed as a commodity purchase, and attempts to treat it as such are rarely successful. Risk mitigation and cost-avoidance should be factored into the price equation.

Bottom line: The relationship between buyer and vendor has to be a "win-win" situation since there is no such thing as a "win-lose" scenario. Neither player has that kind of leverage when it comes to software and implementation services. This is a marriage between two organizations, a symbiotic relationship that by its very nature will produce a "win-win" or a "lose-lose". There is nothing in between.

Lawrence Dean Shemesh
OPSdesign Consulting