Skip to content

Get In Line! The Ups and Downs of Excess Demand

April 8, 2011

Three weeks ago, Apple’s magical and revolutionary iPad 2 went on sale. Although Apple has not yet released any sales figures, demand appears to be extremely robust. Supply, on the other hand, has failed to keep pace – just ask anyone who has tried to pick one up.

I am certainly a member of that group. I have made three trips to my local Apple Stores since the iPad launch. According to one of the salespeople at the Palo Alto store, folks have been lining up starting at 4 or 5 in the morning just to plop down their $500 (or more – the average selling price for the original iPad was $645 in Q4).

Analysts were quick to call the iPad 2 launch a success. Piper Jaffray’s Gene Munster estimated that Apple sold between 400,000 and 500,000 for the weekend, while Chris Whitmore from Deutsche Bank called it “one of the company’s most successful product launches to date.”

Here’s the billion-dollar question: Is Apple leaving profit on the table with its handling of the iPad 2 launch? Does the fact that customers are willing to spend 5+ hours waiting in the rain constitute a dead-weight loss for Apple and its shareholders? And, most importantly, could Apple (or any other company) monetize this exuberance for their products?

First, a quick microeconomics refresher. The two fundamental market forces, supply and demand, set a price. If the price is too low, there will be excess demand (and/or not enough supply), while if the price is too high there will be excess supply (and/or not enough demand). Intuitively, many people will want to buy iPads if they are cheap, however Apple will not be willing to manufacture iPads to sell below cost. On the other hand, Apple would move any mountain to manufacture any number of million-dollar iPads, however they would find a very limited market.

The Case For Dynamic Pricing

Imagine, for a second, the Apple Store in Economistland. iPad 2 demand is still robust, and iPad 2 supply is still limited. But, this being Economistland, everyone behaves perfectly rationally. Understanding their opportunity costs and utility curves perfectly, and therefore is able to submit to Apple exactly how much they would be willing to pay for their preferred iPad 2 model on each day of its availability. Apple could then allocate each day’s availability of iPads to maximize revenues and profits, while each citizen of Economistland would get the opportunity to purchase his iPad at a price and on a date that he agreed would be acceptable. (Equivalently, Apple could hold a uniform price auction for each day’s inventory of iPads.) The result, at least for the Economistland Apple Store, would be increased profits, while customers would be no worse off, having exchanged one commodity (in this case, waiting in line for several hours in unpredictable weather conditions) for another (in this case, cash), and thereby implicitly valuing their time. Apple shareholders are happy, and the citizens of Economistland are warm, dry, and still get their iPads – a Pareto improvement.

We’ve already seen several notable examples of dynamic pricing. The San Francisco Giants started using dynamic pricing for all seats last season. allows hotels and airlines to clear unsold inventory at below-market prices. Finally, Smart Meters will allow public utilities to dynamically price electrical power to (hopefully) reduce peak demand. All of these cases share a common thread – limited, expiring inventory (baseball tickets or airline seats have very low marginal costs, so selling them for even $1 is better than leaving them unsold).

There are two scenarios that business owners want to avoid. In one, you have inventory and nobody willing to buy it at your listed price (for example, airlines, just before takeoff). In the other, you have credit card-wielding customers and no product to sell them (Apple). Microeconomics says that each of these problems can be solved by adjusting the price to bring supply and demand back into sync. Indeed, this guy was willing to pay $900 for the first spot in line at the 5th Avenue Apple Store, and there is a thriving gray market for iPads in Hong Kong. Finally, iPad 2s are still selling at a markup on eBay. So why should Apple leave those profits to bored college students and enterprising international businessmen? Moreover, if the first spot in line is worth $900, shouldn’t the second spot be worth something too?

What might such a pricing system look like? Let’s say that iPad #1 sells at a $900 premium, and that premium decays quadratically (1/x^2) to $0 over the course of the first 1 million iPads sold. Using a simple integral, this comes out to a $300 million profit, or about 30 cents per share. There’s even precedent for this from Apple! The original iPhone went on sale for $599, but was discounted to $399 a just few months later. (Apple offered the early adopters a $100 credit.)

The Case Against Dynamic Pricing

So why won’t Apple be adopting dynamic pricing any time soon? Well, for starters, we don’t live in Economistland. Real-world customers exhibit a divestiture aversion, meaning that they would require more money to give up their spot in line than they would be willing to pay for an equivalent spot in line. In trading lingo, the spread is too wide.

But what about the positive impact of lines? Let’s take a look at one case study, Ike’s Place in San Francisco. For those of you not in the loop with Ike’s, it’s a fantastic sandwich shop, known for notoriously long lines. (Their San Francisco location was recently forced to move after the neighbors complained about noise stemming from their waiting patrons.) Here’s the thing about Ike’s – of course their sandwiches are good, but it’s notable for the lines and not for the  quality of the sandwiches. Moreover, the long lines are a selling point and serve a business purpose. 

If Ike’s wanted to take a traditional approach to improving customer service, they could take steps to reduce their customers’ wait times. For example, by hiring a large staff during peak times, they could make more sandwiches. From my experience at their Stanford location, their ordering process is intentionally slowed down, which creates a buffer effect. By throttling the rate at which customers can order, a larger portion of the total wait is spent waiting in line, which is less painful than waiting at a table for a sandwich (since the line is of definite length and is visibly being serviced).

So what’s the benefit of lines? It’s all psychological.

  • Bandwagon effect: Lines provide social proof (Ike’s sandwiches are good, iPads are magical), which hopefully creates buzz and leads to increased sales.
  • Irrational escalation: If I wait in line and don’t get to buy and iPad, I will be willing to come earlier the next day to justify the time already spent waiting.

In other words, it’s all marketing. It’s all about creating an image, of your company or product as popular, underpriced. It’s all about cultivating a cult-like following for your brand (something Apple has done extremely well). Every company’s dream is to have customers lining up to buy their products.

(Ike’s lines also serve an economic benefit. Think about Ike’s in terms of fixed and variable costs. Ike’s has fixed costs, like space and training, and variable costs, like personnel hours and sandwich materials. If Ike’s were to change their business model to alleviate lines, they would have to service most of their customers between 12 and 1, which would require a large staff for a short time, and a lot of counter space. On the other hand, their current business model allows them to operate with a relatively small staff (over a longer period of time), and out of a small-ish shop, thereby reducing their fixed costs while keeping their variable costs approximately the same.)

In conclusion, pricing is an incredibly complex field. Apple’s corporate image requires simplicity, so they definitely won’t be moving to dynamic price anytime soon, but other businesses in the same situation (massive demand which exceeds an initially limited supply) may choose to charge early adopters a premium.


Giants Baseball: Excellent

October 3, 2010

Giants win the West!

Freddy Sanchez celebrates scoring the Giants' second run on Sunday. (Mercury News)

Giants Baseball: Torture

October 2, 2010

What can you say, except that deep in the back of your mind you knew this was going to happen. Giants enter a three-game series with San Diego to close out the regular season, drive their magic number down to ONE by sweeping the D-Backs, and then drop two in a row to the Padres.

First, the good news. It would still be very difficult for the Giants not to make the playoffs. Between their big lead in the West coming into this weekend and Atlanta’s two straight losses to the Phillies, ESPN’s Hunt for October still has the Giants with a 94% chance of making the playoffs. That itself is not the issue. The issue is, if Jonathan Sanchez and the Giants can’t get it done tomorrow against Mat Latos and the Padres, they have to burn Lincecum against the Padres on Monday in San Diego. And, if they can’t get it done in San Diego, they’re off to Atlanta (assuming the Phillies win again tomorrow) where they burn Matt Cain. In this worst-case scenario, a win in Atlanta on Tuesday sets up a Wednesday matchup between the Giants’ #3 (Bumgarner? Zito?) and NL Cy Young-to-be Roy Halladay. All bad.

The moral of the story is that tomorrow’s regular season finale between Mat Latos and Jonathan Sanchez is a must-win for San Francisco, and it will be a game not to miss. Sure, anything is possible once you reach the playoffs, but the road to the World Series is much more pleasant if the regular season ends tomorrow with a victory.

Go Giants!

Shameless Plug: MentMe

October 1, 2010

Well, I’ve signed up on MentMe, a site that gives high school students access to help from college students are recent grads for guidance through the college admissions process. I think that the site is live now, so if you’re interested in hiring me as a college admissions consultant, my profile is here.

Video of the Night: 2KV Launcher

September 19, 2010

How could you not love a video of a 2KV washer launcher destroying cans?

What’s so great about Google Instant? Hint: Not the Instant

September 19, 2010

Ever since Google’s big announcement about ten days ago about saving a couple billion seconds with their new Instant interface. Presumably they’re using a lot of AJAX to serve up search results as you type in your query. (Formerly they would only suggest other queries, now the results actually update in real-time.) Since the announcement, there have been a bunch of articles both touting kids (including a few of my Stanford classmates) who wrote “Instant” interfaces for other services (most notably YouTube) and lamenting the fact that Bing didn’t introduce this feature before Google did. This is actually a pretty interesting project for learning some AJAX – write a generic Instant interface that works for any site with a searchbox – digg, twitter, etc. I might do it if I have a spare day this quarter.

What they are ignoring here is that the challenge isn’t in the interface. It’s all about the servers. It’s no problem for a 15-year-old or a Stanford kid to build a new interface for YouTube or Bing or whatever. The problem is actually deploying it on the official site and having a couple hundred million users. Imagine that Google serves on the order of 500 million searches a day (quantcast says that about 80 million people visit each day). With the Instant interface, each letter or two becomes its own search. Using everyone’s favorite ‘back of the envelope’ math, turning on Instant takes Google from 500 million to as many as 5 billion queries a day almost immediately (they did stagger the roll-out a bit).

Here’s my point: while it’s great that people are building Instant interfaces for things, for a site like Bing to actually turn Instant on for their main site would require a massive investment in server hardware more than anything. If Google Instant catches on, they may have to do just that, but this isn’t the kind of project that you can just throw an intern at (as much as I’d like it to be otherwise – yes, I am interested). The magic of Google Instant is all about the servers, not the AJAX on the front-end.

EDIT: One very good article (don’t remember which) points out that the server load associated with a 10x increase in search traffic is still minuscule when compared to the cost of serving several hundred million videos on YouTube each day. At the end of the day, it’s all about the ads. CPI has to come down, right? Or not? How can Google count an impression if you only look at the ad between when you type ‘W’ and when you type the ‘i’ in Wikipedia?

Book Review: You Are Here

September 13, 2010

I just finished Colin Ellard’s You Are Here: Why We Can Find Our Way to the Moon, but Get Lost in the Mall [amazon]. It makes a decent read, although I was expecting something a little more like this article over at Slate about wayfinding in the context of architecture or urban planning. Instead I got a highly-interdisciplinary work that started with some incredible stories about navigation in bumblebees (as well as the expected ‘rats in the mazes’). We do get a pretty good discussion about some of the tools that people use to navigate spaces (as well as a long discussion of why we’re so bad at it), but Ellard sounds a bit too much like Al Gore for my taste when he suggests that the solution for our wayfinding problems is to spend more time outdoors.

Overall this was a great extremely well-researched book. It is very evident that Ellard is far and away a subject matter expert in this field. It wasn’t exactly what I was looking for (or what I expected from the picture of a GPS on the front cover), and it’s not going to find its way onto my all-time favorite books list, but it has certainly earned a solid three-star rating.

%d bloggers like this: