In its 10-K filing (page 33) Amazon.com (NASDAQGS: AMZN) reported spending $1,174 million on outbound shipping in 2007. That amounted to 7.9% of revenues. The good news is Mr. Bezos also earned $740 million on those outbound shipments, a lot of it from the Amazon Prime Club annual membership dues.
A LITTLE BACKGROUND
In the ninth of this series of my ten posts on airline mergers, Passengers, Packages: The Paradox of Air Transport, I reported that FedEx (FDX) and UPS (UPS) had Michigan ACSI customer satisfaction scores around 80% in each of the last ten years. In the same period, American Airlines (AMR) and United Airlines (UAUA) barely broke the 60% customer satisfaction barrier. Claes Fornell and his colleagues have shown that higher ACSI consumer satisfaction scores are associated with high returns and low risk. Prompted by these results I asked:
Wouldn’t it improve consumer satisfaction if passenger airlines took baggage out of their equation? And specialized in transporting passengers? In economics this is called the “division of labor.”
In his Enquiry into the Nature and Causes of the Wealth of Nations Adam Smith wrote in the first paragraph of Book 1, Chapter 1:
The greatest improvement in the productive powers of labour, and the greater part of the skill, dexterity, and judgment with which it is any where directed, or applied, seem to have been the effects of the division of labour.
You may think taking luggage handling out of the passenger carriers’ operational equation is a great idea, or you may think it’s ridiculous. Either way you will agree it can’t happen unless express shipping can affordably be married to airline reservation systems. The purpose of this post is to chart a demand curve for express luggage services without any historical data. In doing so, I hope to suggest the impact specialization might have on the price of express shipping services.
THE VOID IN PRICING DECISIONS
Price elasticity may be the most widely taught, yet little used, concept in marketing science. As evidence, price elasticity is not even mentioned in the text of two classic pricing cases involving radical innovations: digital cameras and deregulated airlines. The reason is not a failure of the underlying theory (see Chapter 10 of Lilien and Rangaswamy’s Marketing Engineering). Rather, it’s due to the inability of management to connect price theory to the market realities of new product introductions.
The demand for brand new consumer products depends heavily on their price. Digital camera demand is a contemporary example of this dependence. Here's what Gary DiCamillo, CEO of Polaroid, had to say about digital imaging in 1997:
... the consumer market will be slower to evolve. I don't think we'll see something major next year or the year after or maybe even by the year 2000…Will it become a big deal? I don't know…I think there's revenue there, but I'm not sure about profit (Polaroid Corp., Rosenbloom and Pruyne 1997).
The uncertainty expressed by Polaroid's CEO was based partly on the difficulty of articulating the relationship between price and consumer demand. Implicit in the CEO's statement "… I'm not sure we can make a profit… " is the trade off between increases in sales volume associated with decreases in price.
Perhaps more directly related to charting the demand curve for express luggage shipping is this comment by Rollin King, EVP of a startup called Southwest Airlines (LUV):
Pick a price at which you can break even with a reasonable load factor… a load factor that you have a reasonable expectation of being able to get … without leading yourself down the primrose path and running out of money (Southwest Airlines, Lovelock 1975).
I’ll bet both Mr. DiCamillo and Mr. King had economics 101 in college. So why wasn’t price elasticity used in their assessment of demand for their innovations? Because there’s a huge knowledge void between drawing demand curves on a blackboard and estimating price elasticity in practice. Drawing demand curves is an art. Estimating price elasticity is a science.
The folks at the airline carriers who are responsible for seat-specific pricing use a vast amount of historical data to price those seats in a way that optimizes revenues. Yet, given less complete (and less applicable) data, one can still make predictions about possible express luggage pricing.
LIMITED INFORMATION DEMAND CURVE
There are no data on what the large scale demand curve for express shipping of luggage might be if it were adopted by passenger carriers. But it’s possible to craft one without such data from the limited bits of information currently available. How? Begin by bracketing the limits of what demand might be. I identified the lower limit in my last post:
I ordered a new HP Laser Jet P2015 printer from Amazon.com. The retail price was $449. I paid just $289.99. The next-day Amazon Prime delivery charge for this 28.9 pound package on UPS was $3.99. Based on this rate significant scale economies must exist in AMZN’s agreement with UPS as a result of huge shipping volumes.
One reader complained that I “conveniently neglected” to mention that Amazon (AMZN) Prime membership costs $79 a year. Actually, I didn’t neglect to mention this, I simply forgot how much it cost and didn't bother to look it up. In case you’re not a member here’s what Amazon Prime membership gets you:
Overnight Shipping for only $3.99 per item
-- order as late as 6:30 PM ET
Ship to any eligible address in the contiguous United States
I think you will agree that $3.99 for shipping a 30 pound package from Dallas to New Orleans next-day air is likely the lower limit that Amazon would pay for this service.
It’s just as easy to get the upper limit. Using the UPS shipping screen I priced a 30 lb box from New Orleans to Dallas standard overnight express at $142.50. I joined UPS online shipping to get this number, so it is the rack rate for a single shipment.
I also need a few points in between the upper and lower limits. The first two points I got from the logistics manager of a company that does $500,000 in UPS shipping services per year. He pays $89.64 to overnight a 30 lb package from New Orleans to Dallas. That’s a base rate of $73.20, plus a declared value charge of $1.80, plus a fuel surcharge of $14.64.
I also got an estimate of $42.75 for that same overnight package from another shipper that books a little over $1.5 million a year with UPS. Scaling that number back to $1 million per year gives a price of about $56. If these last two prices are representative, it is likely when a shipper doubles annual volume from $500,000 to $1,000,000 they earn a 37% discount. So I ran this discount assumption up the scale doubling volume each time till it got close to that $3.99 AMZN cost. That happened at $64 million per year in annual shipping volume.
CHARTING A DEMAND CURVE
Because of its restrictive properties the constant elasticity demand curve works better than the linear demand curve in this context. It is expressed as a power function in the following chart.
In this function annual shipping volume [q$] equals a constant [a] multiplied by the price to ship a 30 lb. package from New Orleans to Dallas [P] raised to the power beta, which is the constant elasticity of demand over the range of the function. Extrapolating this equation using the limited information described above creates the demand curve in this chart.
Based on the limited information and assumptions behind this chart, AMZN would reach a net per package rate of $3.53 at an annual shipping volume of $64 million. Note from the opening paragraph in this post that Amazon had a net outbound shipping volume of $434 million in 2007. It just may be that Mr. Bezos pays even less, on average, than $3.53 for that nationwide shipping rate! Only he knows for sure.
Remember there are no data on shipping luggage in the volumes that would occur if baggage were moved from the passenger transport into the package transport system. Such a move probably would create tonnages well beyond the current capacity of even the largest express shipper (DHL revenues are 50% greater than UPS). If this chart is indicative of that future, demand is highly price elastic: beta is -1.5.
THIS AIN’T TEA LEAVES
In pricing a new product or service it is necessary to imagine a future that cannot be projected from the past. Charting demand curves from limited information brings the power of theory to bear on the relationship between price and volume. As Peter Drucker famously said “The only way to predict the future is to create it.” But before you can create it you must imagine what it looks like. Isn't that what theory is designed to do?