Pricing and Amazon Prime

We first got an Amazon prime membership on a trial basis, and because I was buying something that the shipping would have been absurd on: a weight bench that wasn't stocked at any of my local sporting good stores. However, it quickly became so addictive to be able to get any thing sold by Amazon shipped second day without any fee for shipping, that we've kept it for the last six years or so.

I'm sure that we're some of the reason that Amazon loses over a billion dollars on Prime every year. For instance, the cheap earphones I'd passed down to the kids so that they could listen to audiobooks on my old iPod finally broke, so rather than go out to the store I got on Amazon, picked out a nice pair for $15 (which would have cost me much more in store) and they showed up two days later to the kids great delight. The fact that it's a pain for us to get out the door to go shopping, especially for something that we'd have to drive down into Columbus to buy, means that free second day shipping is a huge value for us. (And we even end up using the discounted overnight shipping a lot for last minute birthday presents and such.)

As such, I was rather chagrined to read that Amazon is thinking of raising the price on Prime by $20-$40 in the near future. The feedback has been overwhelmingly negative, and I myself have wondered if we should cancel if it happens. However, reading Rafi Mohammad's piece on pricing Prime in Harverd Business Review, I think I probably disagree with his alternative suggestion.

The solution to Amazon’s pricing dilemma is straightforward: unbundle the components of Prime. By unbundle, I mean instead of making a “take everything in the package or leave it” price hike, allow customers to purchase what works best for them.

Much like cell phone carriers serve customers with varying usage needs, Prime should provide different shipping options. The premium option can be the current unlimited service, albeit at a steeper price. A lower-priced option could require a purchase minimum of, say $25, to get free two-day shipping. This will curb shipping losses from irksome small dollar orders such as $5 scissors. Another variant is to offer a classic two-part fixed/variable pricing strategy. For instance, there could be a $50 annual Prime membership that gives members the right to receive two-day shipping at a variable price of $1 – $2 a package.

Given that streaming and the Kindle library have little to do with shipping, there’s no clear reason to include them in the Prime bundle except to elicit a “wait, there’s more” urge to buy. Why not simply offer streaming and the Kindle library a la carte or as an extra option to those who purchase a shipping option?

Unbundling Prime and providing choices better serves customers, which increases the popularity of this critical growth program. Amazon’s challenge provides two key pricing lessons to all managers. First, understand and respect that customers have different needs. One price does not fit all – provide pricing options to serve the diverse needs of customers. Just as importantly, when implementing a significant pricing increase, it’s critical to offer customers lower priced options. No one likes “take or leave it” ultimatums.

The reason, from Amazon's point of view, for keeping Prime bundled is that it allows them to get people used to extra value. Take the instant video streaming service. We already subscribe to Netflix, and the things available overlap to a great extent, so I don't think I would have ever tried Amazon streaming if it hadn't been included in Prime. As it is, though, they selection is slightly different. (For instance, the had the BBC production of Emma on streaming while Netflix did not.) And so we've gradually fallen into using it more. Now I'd be more inclined to drop the Netflix streaming and pay for Amazon if they split it off into a separate service from Prime, but the key is: I never would have got to that point if they hadn't offered it as a free service to Prime members for quite a while.

Offering free add-on services to Prime rewards Amazon's most loyal and profitable customers, and it encourages those customers to broaden the number of services they turn to Amazon for. Splitting Prime into a set of a la carte options would probably significantly slow adoption of new services which Amazon would like to launch.

The one split out that would potentially work would be if there are a small number of users who are having a massively disproportionate effect on the cost of Prime by shipping huge numbers of items. (For instance, people who are doing ordering for some kind of small business.) If 10% or less of Prime members are accounting for a huge percentage of the shipping costs, and if it's possible to come up with a way to charge for for "unlimited two day shipping" versus some sort of more modest plan like "50 free second day shipments a year, with additional shipments at $2.99 each" or "unlimited free 5-6 day shipping and 25 free 2nd day shipping credits per year" I think that could potentially work as a segmentation. However, while as a consumer I'd happily not have have to pay for the Kindle lending library (I don't own a Kindle), I think that in general Amazon would probably not be well served by segmenting their Prime offering.

As for the size of the increase: They're going to face major backlash when they increase Prime pricing regardless, so it's probably a good bet for them to make a price increase such that they won't have to do it again for another 5-7 years. In that sense, it's probably smart to do a single $20-40 increase and deal with whatever percentage of people actually cancel (probably much smaller than the number who say they'd cancel), than to set themselves up for having to do smaller price increases every 2-3 years and deal with the backlash each time.

How Pricing Fairness Can Hurt You

Megan McArdle has a great piece up dealing with one of the things that often bothers people about price discrimination: finding out that other people are consistently paying less than you. Since we live one of the wealthiest and largest countries in the world, and since many of the companies we buy products from do business internationally, this often means that US consumers find out that the products they pay a lot for here in the US are being sold for much less overseas.

Case in point, the US Supreme Court is considering the case of Kirtsaeng v John Wiley and Sons which deals with whether companies can get into the business of buying US-made textbooks overseas and shipping them back to the US. Why would this be advantageous? Textbooks are fairly expensive to produce (as in, to get all the content put together and ready to print) but the cost of producing additional textbooks is low (the cost of printing and binding.) The result is that publishers sell textbooks for far, far more than the cost of printing a book -- that is, far above the marginal cost (the cost of producing one more product.) This isn't because it allows the publishing executives and their investors to lie on beds of gold coin like dragons, publishing companies are not actually massively profitable. It's because the revenue produced by selling the textbooks needs to cover the cost of producing the book content as well as the cost of producing the book itself. So they sell textbooks for what the market will bear, which with something as expensive as college is a heck of a lot.

However, since publishing companies are also constantly on the lookout for more revenue (again, to cover their production costs, as well as to produce profits for their dragons to lie on) after pricing college textbooks in the US at what the US market will bear, they'll go into other countries and sell the same textbooks there for what that country's local market can bear, so that as that price is still above the marginal cost of producing more books. The result: US textbooks can often be bought for much less in other countries. Thus creating the opportunity for some enterprising company to buy the books abroad and ship them back, if only they can get around the legal restrictions thereon. But will that save everyone money? McArdle describes why not:

It is possible for a firm to make money with some of its customers paying less than the average cost (but more than the marginal cost of producing an additional unit). It is not, however, possible for a firm to stay in business with all of its customers paying less than the average cost.

But we want to believe that it is possible. Indeed, no matter how often it is explained that we cannot all be the marginal cost consumer, someone will insist quite loudly that it is possible; that only greedy companies, incompetent bureaucrats, or bad laws stand between us and the joys of marginal cost pricing.

She draws out an example of how this situation would equalize out to the higher price with lower overall unit sales, taking as her example a situation in which ordinary US consumers are the beneficiaries of price discrimination: air travel.

If you're skeptical that this is true, consider an area where you're probably the beneficiary of price discrimination: business travel. The first class and business passengers provide the bulk of the profit on airline tickets; they pay for the extra amenities, and the ability to make last-minute arrangements for short stays during the week. Restricted economy travelers often generate very little revenue over and above the cost of transporting them--but since the seat is a wasting asset, you might as well sell it even if you'll only make a few bucks on it.

Let's say that businesses who travel a lot got a law passed to eliminate this sort of price discrimination, forcing airlines to set a single price for a given route, no matter what date or time the flight was. Such a law might well pass, since people hate confusing airline prices.

The result would not be that everyone got the tourist class price, however. Imagine that you've got 10 business class passengers paying $100 apiece, and 100 tourist class passengers paying $5 apiece. Assume that the marginal cost of taking on an extra business passenger is $2 and the marginal cost of a tourist passenger is $1. Assume your fixed cost of getting the plane off the ground is $1300, so you've got a total cost of $1420 and revenue of $1500.

Now say someone comes along and mandates that everyone has to be charged the same ticket price. What happens? Well, you charge all 110 passengers $13.50 and keep your margins about the same . . .

But wait! The tourist class passengers are very price sensitive. They're on average only willing to pay $5-$8 for a ticket. When you raise the price to $13.50, half drop out. Now you've got 60 passengers, which gives you revenue of $810. But your cost has only dropped by $50. Now you're losing a big hunk of money. You'll have to raise prices. $23 should cover it.

Gee, $23 is more than 4 times what the tourists were paying. They don't want to see Grandma that much. Another half of the tourists drop out. Now you've got 35 passengers. But your costs have only dropped by $25. You'll need to raise prices again. Maybe $40 would cover it?

You can see where this is going; pretty soon almost all you have left is business passengers, paying what they were paying before. In fact, depending on how many tourist class passengers drop out, the business class passengers might end up paying more.

someone has to pay for all those planes, and airline profit margins are far from reliably spectacular (which is why they keep ending up in bankruptcy). No, the tickets for your next vacation would probably cost three or four times as much. Obviously, this is a made-up example, with numbers chosen more for ease of calculation than versimilitude--in the real world, most people would snap up a $38 plane ticket. But empirical research bears this out; ending price discrimination doesn't necessarily mean consumers get a better deal. It can easily mean they get a worse deal.

But there's something deep within us that resists that insight. We hate the feeling that someone else is paying less than us, even if it's not costing us anything. So while a decision for the textbook importer might not make consumers any better off in the pocket, it may make them feel better about the high prices they're paying.

The Ethics of Price Discrimination

To our modern ear the very term "discrimination" sounds bad. However, in the pricing discipline, "price discrimination" is not necessarily a bad thing. The term refers to creating a pricing and product structure which finds a way to charge customers with different willingness or ability to pay different amounts.

Examples include:

The price of clothes in a department store vs. in an "outlet" store.
The price of a newly released novel versus a backlist novel.
Name brands vs. store brands in grocery stores.

In most cases, price discrimination involves creating a situation in which customers can make a trade off between convenience and price or quality and price in order to get fairly similar products for very different prices. This ends up being good for both customers and companies. Companies are able to maximize their sales, allowing them to pay their employees, grow, and produce profits for their investors. They're able to do this by getting the most money from the customers most able to pay, and providing customers with the least ability to pay with the best deal.

However, some forms of price discrimination work more like an intelligence test. One of these I was reminded of recently when setting up travel arrangement for some upcoming trips. Most car rental companies now offer an option where the customer is offered the chance to pay for a tank of gas up front, usually at a rate $0.10 to $0.40 lower than the standard gas station price in the area. This seems like a great deal, until you remember that it only works out as a savings if you return the tank entirely empty. If you return the tank half full, you get no refund for that half tank, and you end up effectively giving the car rental company an extra $20 or so. By making money on the pre-sold tank of gas, the car rental companies are able to offer lower daily rates, allowing them to win more business, pushing other companies to lower their rates and make their money back in other less obvious ways like pre-paid gas, car rental insurance, etc.

This does offer customers who can think through all the offers they're given an opportunity to pay less, so it clearly benefits some customers. However, rather than doing so based on the willingness of customers to make small trade offs in convenience or quality, it does so based on their ability to problem solve. The pre-paid gas is particularly tricky, because at first pass the pre-pay at the lower price actually seems like a way to save money, but in most cases it will turn out to be an extra fee.

These kinds of price discrimination seem more ethically troubling, and yet it's very hard for companies to resist such approaches when they catch on, because "hiding" profits through this kind of pricing allows the companies that adopt such tactics to offer lower up-front prices.

What's it worth? What's the price?

Brandon has a post up over at Siris which I can't help linking to, given the subject matter. The origin in this bloggish chain of being is a post over at First Things, showing the world's most expensive photograph:

This picture sold for $4.3 million

Clearly, one might ask, is it worth it?  Several commenters make the claim that it must be worth $4.3 million, since that's clearly what someone paid for it.  Brandon responds:

I find the automatic response of some commenters interesting: the claim that things are worth simply what people are willing to pay for them has the direct implication that nothing can be overpriced or underpriced; that there are no bargains and no bad deals; that, in fact, it is impossible to price anything unjustly as long as someone will pay it. Nobody actually believes such nonsense; everyone attributes value and disvalue to things by comparisons that have nothing to do with the actual price paid, and we all can make perfect sense of saying that somebody paid too much for something, or that a price is too high even if someone will pay it. The conditions under which price paid can reasonably be said to track the worth of the thing in question are not universal; in part because other things beside the thing bought can be factored into deliberation about price. Status signaling, for instance, which can at times have virtually nothing to do with the thing bought. Nor is it plausible to identify actual worth with attributed worth in the absence of any consideration of practical or moral rationality. But it is interesting how easily people will swallow such an incoherent principle, merely because they have the notion that it’s ‘economics’.

This brings up an interesting conflict between price and worth. Something's price is determined by some balance between how much those who want an item are willing to pay for it, how much those who currently have the item are willing to accept in return for it, and the supply of items on hand. Something like this photograph becomes particularly tricky because the article being sold (a particular print, mounted a particular way, with, one assumes, some implicit agreement as to how many identical items the original photographer will or will not allow to be produced) is unique. Thus, it only takes one potential buyer with (to use my mother's phrase) "more money than good sense" to set the price very high indeed. Because it is determined by willingness to pay, price is necessarily volatile at times, especially when you're dealing with something like this.

But what is its worth?

This is an altogether different question. Certainly, by all our actions we seem to suggest that things do have a worth, and that to some extent, we know it. As Brandon points out, we could never suggest that something was a good or a bad bargain, that something was over or under valued, if we did not have some sense that a thing has some knowable degree of worth.

I'd like to think for a moment on the intersection of these points: the way that price is determined by many people making assessments of worth. Take any given item. I have an ipod sitting on the desk here, so I'll use that. How much is it worth? There's a huge range of different prices at which you could sell some number. But at each price, the numbe that you sold would be a function of the number of people who considered the price 'worth it' and the number who didn't. You might, theoretically, be very successful in selling a product that the vast majority of people did not consider worth the price, so long as enough people did consider it worth to the price to make your business structure work. (Things fall apart when a company expects far more or far fewer people to consider a price 'worth it' than turn out to be the case.)

In this sense, markets (the "economy" as we tend to think about it) are only necessary or useful because there is a great deal of disparity out there in what people consider things to be worth -- both in an absolute sense and in relation to their available means. Pricing and markets are a method of getting around differing preferences and amounts of knowledge when allocating scarce resources among a large and diverse group of people.

As for the world's most expensive photo? No, I don't think it's worth it. I'm out of the running.

All About Pricing: Price Discrimination

I thought it might be interesting to do an occasional series on how pricing works -- since it's what I do all day and I (clearly a biased source) find it interesting. I flatter myself this might also be educational in that many people don't think much about how pricing works in making a business successful.

One of the important-but-illusive concepts that I deal with a fair amount as a pricer is price discrimination. This is a term which different people use in different ways, and on its own it sounds a bit scary (we don't normally think of "discrimination" as something we want a business to be doing) so I'll define it roughly here at the beginning and then discuss some examples: Price discrimination consists of selling similar or identical products or services to different customers for different amounts of money.

Why would you want to do that?

To be successful, a business needs to meet both its fixed costs and the cost of the goods it sells. Take a restaurant. The fixed costs include rent, the kitchen and wait staff, utilities, advertising, etc. The costs of the goods it sells are the costs of the food ingredients used to make meals for its customers.

If the restaurant gets really, really busy, it might need more staff, but most of the time the cost of keeping the restaurant open for business are the same whether 100 people or 150 people come in for lunch.

So, you're running Egan's Irish Pub and right now your business is just breaking even. It's worth it to you to offer lower prices to new customers if that would increase your traffic, because your fixed costs are already met. But you don't necessarily want to discount all the business that you already have, because then you might find yourself no better off. What do you do?

Well, you could offer a coupon. If you get the coupon mostly to people who aren't already your customers, you increase your sales and the lower price those additional customers are paying is okay because you only have to meet your cost of goods on those sales: your fixed costs are already covered.

You could also offer time dependent discounts. The "happy hour" is a classic example of price discrimination. You pick of a time of day when not many people normally come to your business and offer people who come at that time a special discount. People who really care about getting a good deal will come at that time, while customers who care more about convenience will come at their usual times.

While not a classic example of price discrimination, certain types of product differentiation can be motivated by price discrimination-type thinking. The idea here is to offer one version of your product for value conscious customers, at a lower price, in order to win their business, while steering other customers to a very similar, higher priced product.

In addition to helping businesses make money, there's an odd sort of social justice angle to price discrimination as well. While the purpose of the strategy is to help companies make more money, the result can end up being that those who are truly short of money (and willing to make certain trade-offs as a result) get to pay less for virtually the same product, while those with more money pay more.

Buying A Car To Save Money

Cars that get over 40 miles per gallon in fuel efficiency are, reportedly, becoming all the rage, with more models from American and foreign car makers being introduced at the latest Detroit Auto Show.

So I got curious, having just started a 18-mile-each-way commute, what exactly are the savings one can achieve by buying a more fuel efficient car? I assumed a situation faily like mine: My car is paid for and costs me only minimal maintenance to keep up (a 14-year-old Toyota Camry) and a 20 mile each way commute.

Say you're considering buying a new car which gets 40mpg for $20,000. That seems moderately standard for these cars. Assume a 40 mile daily round trip commute, and an additional 40 miles of weekend or additional driving. Assuming a current care actual efficiency of 20mpg. Assume the price of gas goes up to $4/gal. How long would it take for you to make up the cost of that new car in fuel savings?

I calculated savings as follows:

(([Daily Commute]/[Old MPG])-([Daily Commute]/[New MPG])) x [Price of Gas]) = Daily Gas Savings

[Cost of New Car]/([Daily Savings] x 6 x 52) = Years to make up cost of car

How long would it take? 16 years.

Then I decided to go at it the other way. What changes would allow me to pay off the expense of a new car in gas savings over a normal financial window of five years?

- If the cost of a new car gets below $6,500
- If the cost of gas goes over $12/gal
- I tried coming up with an increased MPG efficiency answer, but it's impossible. Even if you could get 100,000 MPG, it would still take eight years (assuming gas prices of $4/gal) to pay off the expense of a new $20k car via gas savings. If you assume a 100MPG car and seek for the gas price which get you to break even in five years, you need a gas price of $8 per gallon.
-If you vary the length of commute, a 130 mile daily commute (assuming 780 miles per week of driving total) would allow a $20k car to pay for itself in gas savings in five years with gas prices of only $4/gal.