Ask Ron Rivest if he’s ever been whisked away by the CIA in the middle of the night, and he laughs-but he doesn’t say no. At Peppercoin, a two-year-old MIT spinoff in Waltham, MA, the renowned cryptographer oversees an operation far less secretive than an intelligence agency but almost as intense: a clearinghouse for electronic “micropayments,” pocket-change transactions that may finally allow magazines, musicians, and a multitude of others to profit from selling their wares online. It’s September, and with only weeks to go until commercial launch, Peppercoin’s software engineers troubleshoot at all hours. Marketing executives shout across the room and over the phone, making deals.
But in the eye of the storm, Rivest is calm and collected. Eyes sparkling, real change jingling in his pocket, he even wears sandals with authority. What Peppercoin is trying to do, he says, is make it easy to “pay as you go” for inexpensive Web content-so you won’t need to pay subscription fees, limit yourself to free content, or share files illegally. With a click of the mouse-and Peppercoin’s software churning away behind the scenes-you can now download a single MP3 from an independent-music site, watch a news video clip, or buy the latest installment of a Web comic from your favorite artist. All for just pennies.
It sounds simple, but it wasn’t possible a few months ago. Most Web merchants still can’t support micropayments-transactions of about a dollar or less-because the processing fees from banks and credit card companies erase any profit. But Peppercoin, the brainchild of Rivest and fellow MIT computer scientist Silvio Micali, is in the vanguard of a new crop of companies-including BitPass of Palo Alto, CA, and Paystone Technologies of Vancouver, British Columbia-that make cash-for-bits transactions superefficient. These companies’ founders are well aware of the string of defunct e-payment companies whose virtual currencies have gone the way of the Confederate dollar. But they’ve got something new up their sleeves: easier-to-use technology that allows Web sites to accept tiny payments by effectively processing them in batches, thereby cutting down on bank fees.
So throw out your current conceptions of Web surfing. Rather than sifting through pop-up ads and subscription offers, imagine dropping a quarter on an independent film, video game, specialized database, or more powerful search engine. If programmers and Web artists could profitably charge a few cents at a time, their businesses could flourish. And with an easy way for users to buy a richer variety of content, experts say, the current deadlock over digital piracy could effectively dissolve, giving way to a multibillion-dollar business stream that rejuvenates the wider entertainment industry the same way video rentals did Hollywood in the 1980s. Down the road, cell phones, personal digital assistants, and smart cards equipped with micropayment technology could even supplement cash in the real world.
“The key is timing and technology,” says Rivest, who thinks Peppercoin has both right. The company’s technical credibility, at least, is not an issue. Rivest coinvented the RSA public-key encryption system, used by Web browsers to make credit card purchases secure. Micali holds more than 20 patents on data security technologies and won the 1993 Gdel Prize, the highest award in theoretical computer science. Their system uses statistics and encryption to overcome profit-erasing transaction fees; the approach is unique and more efficient than its predecessors.
The timing looks good, too-not just for Peppercoin, but for other micropayment companies as well. “One year ago, it was, Will people pay?’ Now it’s, How will they pay?’” says Ian Price, CEO of British Telecommunications’ Click and Buy division, which uses micropayments to sell articles, games, and other Web content to customers in more than 100 countries. And in September, Apple Computer announced that its online music store sold more than 10 million 99-cent songs in its first four months. Apple’s success was the “starting gun for a track meet of companies” planning to roll out pay-per-download services by 2004, says Rob Carney, Peppercoin’s founding vice president of sales and marketing.
Indeed, 40 percent of today’s online companies would sell content they’re currently giving away if they had a viable micropayment system, says Avivah Litan, an analyst at Gartner Research who specializes in Internet commerce. According to Forrester Research, the market for music downloads is expected to grow from $16 million in 2003 to $3 billion in 2008. And a Strategy Analytics report states that mobile-gaming revenues could top $7 billion by 2008. “The market is ready” for micropayments, says Rivest.
Even so, getting the technology to take off won’t be easy. Micropayment companies need to make their systems fully reliable, secure, and easy to use. Just as important, they need to increase demand by working with Web businesses to deliver a broader range of digital products. So on the eve of Peppercoin’s commercial launch, the question is not whether the timing and technology are good. It’s whether they’re good enough.
In Statistics We Trust
Understanding Peppercoin requires a little history. According to old English common law, the smallest unit of payment that could appear in a contract was a peppercorn. Silvio Micali’s wife, a professor of law, suggested that as the name for his startup back in 2001, and it stuck (becoming “Peppercoin” for the sake of clarity). Now, in his office at MIT’s Computer Science and Artificial Intelligence Laboratory, Micali is explaining what makes Peppercoin tick. On hand are technical books and papers in neat piles, should we need them. It’s simple mathematics, says Micali-but don’t believe him.
Micali knows two things: cryptography and coffee. His micropayment analogies involve cappuccinos. There are two standard ways of buying digital content, he says. One is like prepaying for a certain number of cappuccinos, the other like getting a bill at the end of the month for all the cappuccinos you’ve had. That is, the customer either pays up front for a bundle of content-say, 10 archived New York Times articles-or runs a tab that’s settled every so often. The problem with both models is that the seller has to keep track of each customer’s tab, and the buyer is locked into a particular store or site. But in the spring of 2001 came a “very lucky coffee break” when Micali and Rivest, whose office is just down the hall, put their heads together. “We started discussing this problem, and within minutes we had the basic solution,” says Micali. “And we got very excited! First, from the discovery. Second, from the coffee.”
What they discovered was a way to cut the overhead cost of electronic payments by processing only a statistical sample of transactions, like taking a poll. On average, Peppercoin might settle, say, one out of every 100 transactions-but it pays the seller 100 times the amount of that transaction. Given enough transactions, it all evens out, says Micali (see “a Penny for Your Bits: How Peppercoin Works,” below).
Peppercoin’s software makes micropayments efficient and profitable by processing only a statistical sample of all transactions. A customer selects an item and sends digital payment (a Peppercoin “token” worth, say, 10 cents) to a merchant’s Web site. The merchant’s computer verifies the token and delivers digital goods, such as MP3 files, to the customer’s computer.
Peppercoin’s software randomly selects one token out of every 100 or so for processing. In that case, The merchant sends the token to Peppercoin. Each token stores a running total of the customer’s spending at all sites. Peppercoin pays the merchant 100 times the face value of the token, e.g., 100 x 10¢ = $10. Peppercoin bills the customer for the exact amount of his outstanding purchases from all online merchants. The customer pays his Peppercoin bill using a credit card.
It looks simple to the buyer, who only has to click on an icon to charge an item to her Peppercoin account, but the action behind the scenes is pretty complicated. In beta tests, special encryption software runs on both the buyer’s and seller’s computers, protecting their interactions from hackers and eavesdroppers. And encrypted in each transaction is a serial number that says how many purchases the customer has made over time, for how much, and from whom.
Ninety-nine transactions out of a hundred are not fully processed-but they’re still logged by the seller’s computer. One transaction out of a hundred, selected at random, is sent to Peppercoin. After Peppercoin pays the seller 100 times the value of that transaction, it bills the customer for all of her outstanding purchases from all sites that use Peppercoin. Since about one out of a hundred purchases is processed, her last bill will have come, on average, a hundred purchases ago. That’s the trick: by paying the seller and charging the customer in lump sums every 100 purchases or so, Peppercoin avoids paying the fees charged by credit cards-roughly 25 cents per transaction-on the other 99 purchases. “This is fantastic,” says Greg Papadopoulos, chief technology officer at Sun Microsystems and a member of Peppercoin’s technical advisory board. “Ron and Silvio have done what needed to be done-get the cost of transactions down without ripping up the existing infrastructure of credit cards and banks.”
But what’s to keep all this fancy statistical footwork from cheating sellers out of their due? And what’s to keep buyers and sellers both from cheating the system? “That’s the secret sauce,” says Micali.
He’s talking about cryptography, the sweet science of codes and ciphers. Its inner workings are, well, cryptic-paper titles at conferences include things like unimodular matrix groups and polynomial-time algorithms-but it’s used every day to keep communications, documents, and payments secure. Roughly speaking, says Rivest, statistical sampling of transactions makes the system efficient, while cryptography keeps the random selection process fair and secure. So Peppercoin charges users exactly what they owe, and if Peppercoin’s payment to the seller happens to be more or less than the value of the purchases customers actually made, the discrepancy is absorbed by the seller. Over time, this jiggle will become negligible, especially compared to the amount of money Web sites will make that they couldn’t make before.
Think about it for too long, and most people get a headache. But Micali and Rivest have been thinking about this sort of thing for 20 years, so they make a formidable and complementary team: Micali is as animated as Rivest is understated, like fire and ice. “They’ve done brilliant work over the years,” says Martin Hellman, a professor emeritus of electrical engineering at Stanford University and a pioneer in cryptography going back to the 1970s. “Peppercoin has a clever approach.”
But clever mathematics aside, the proof is in the pudding. In the end, Peppercoin’s executives say, their system must be as easy to use as cash. Perry Solomon, Peppercoin’s founding CEO, explains it this way, pulling some change out of his pocket. “I can give you this quarter, and you can look at it quickly and say, Okay, that’s a quarter.’ You don’t need to call the bank to verify it.” Online merchants, however, must check a credit card holder’s identity and available credit before approving a purchase. Going to that trouble makes sense for a $50 sweater or a $4,495 Segway transporter, but not for a 50-cent song. So Peppercoin’s software stamps each transaction with the digital equivalent of e pluribus unum-a guarantee to the seller that it’s Peppercoin handling the transaction, and that payment is forthcoming. The seller can quickly verify this stamp and deliver the goods.
Bootstrapping with Bits
The theory may be impeccable, and the founders’ credentials outstanding, but how does a startup transform a micropayment system into a practical, sellable product? That’s the stuff of late-night whiteboard discussions enhanced by takeout Chinese food and bad TV movies, says Joe Bergeron, Peppercoin’s vice president of technology. Bergeron, a baby-faced programming whiz, has the task of translating Rivest and Micali’s algorithms into software. Like any good engineer at a startup, he has spent many a night under his desk trying to squeeze in a few hours of sleep. “I’m dreaming in Peppercoins now,” he says.
Minting micropayments starts with hardware. A secure data center a few kilometers from company headquarters houses hundreds of thousands of dollars’ worth of computing horsepower and memory. All of Peppercoin’s money transfers flow electronically through these machines. A rack of 20 processors and backups and four levels of hardware security are set up in a special cage walled off by Plexiglas guaranteed to withstand a 90-minute riot; the rental contract even specifies that the cage will repel “small-arms fire and manual tools.”
The Coin-Op Web?
In the 1990s, e-payment startups like DigiCash, Flooz, and Beenz crashed because dot-com companies didn’t think they needed the technology to make money, and because consumers expected Web content to be free. Times have changed, but there are still plenty of skeptics who doubt micropayments will catch on broadly, considering that MP3 listeners and Web-comics fans are the technology’s main U.S. consumers so far. Even those who have made their fortunes in the online-payments world acknowledge that it’s an uphill battle. “It’s quite possible they could fail miserably in this economic climate,” says Max Levchin, cofounder and former chief technology officer of PayPal (see sidebar “The PayPal Precedent”).
But both the supply of digital content and consumers’ willingness to pay for it are increasing, and the micropayment companies’ strategy of signing up Web merchants, one at a time, has promise. “There will be small companies who figure out how to play this chicken-and-egg game,” says Andrew Whinston, director of the Center for Research in Electronic Commerce at the University of Texas at Austin. “The key is to become successful before big companies like Microsoft get into it.”
How a Russian cyberwar in Ukraine could ripple out globally
Soldiers and tanks may care about national borders. Cyber doesn't.
Meet Altos Labs, Silicon Valley’s latest wild bet on living forever
Funders of a deep-pocketed new "rejuvenation" startup are said to include Jeff Bezos and Yuri Milner.
Meta’s new learning algorithm can teach AI to multi-task
The single technique for teaching neural networks multiple skills is a step towards general-purpose AI.
Going bald? Lab-grown hair cells could be on the way
These biotech companies are reprogramming cells to treat baldness, but it’s still early days.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.