Just one difference. In this case, the advertised goods are us. We want to find, and we want to be found. And increasingly, we’re going to have to figure out how to use these statistical profiles of ourselves to create the spark. As our mating rituals migrate from bars and study halls to electronic networks, honing our algorithms may become as important as the smiles, scents, and sideways glances that Shakespeare knew so well.
HAVE YOU EVER noticed the little button on the Google search engine called “I’m Feeling Lucky”? Type in a query and click that button, and it delivers only a single Web page, the one deemed most likely to satisfy your search. “I’m Feeling Lucky” cuts to the chase. Yet it receives almost no attention and accounts for far less than 1 percent of all searches, according to Google. Why is this? For starters, we don’t trust the machine to understand our instructions and deliver the right Web page 100 percent of the time. And what if there’s another Web page that’s just a little bit better? The fact is, we like choices. We like to browse through the possibilities. Imagine if Chemistry.com, like the matchmakers for European royalty in centuries past, lined us up with just one potential mate. We’d feel cheated. And just like the eternally frustrated Henry VIII, we’d naturally wonder if there wasn’t someone with an extra pinch of this or dose of that. Even if science—whether a search engine or an online dating site—had the smarts to give us exactly what we’re looking for, we wouldn’t really be sold until we saw the other possibilities. (Some of us like to keep testing the possibilities long after we’ve made our choice.) The key for these services is to give us a selection of good choices.
And the key for us? Our success in the networked world, whether we’re hoping to land a date or a job, hinges not just on our ability to find, but also to pop up on the first page of other people’s search results. Throughout history, we’ve developed all sorts of ways to be found. We wear perfume, jewelry, tattoos, platform shoes, all of them delivering a message of who we are. We write up résumés, we build big and complex social networks. We crack jokes. Some of us pay to appear in reference books, like Who’s Who. As the Numerati assert their ways, we will be located less by the sights and sounds we produce, or even our friendships, and more by math-based programs churning through our data. The trick, increasingly, will be to help the machines find us and to use machines to locate others.
For companies, this burning need to be found has spawned an entire consulting industry. It’s called search engine optimization, or SEO for short. Let’s say you have a bed-and-break-fast in Tucson. But when potential customers type “Tucson Bed Breakfast” into a search engine, your site doesn’t appear until the fifth page. That spells disaster. Potential customers never find you. So you go to consultants and you pay them to engineer your website so that it shows up near the top of the list. (The Internet is teeming with companies offering this service.)
To optimize your page, they have to understand the search algorithm. How does it define a high-ranked page? Is it a plethora of links to other pages? Loads of traffic? The prominence of certain words? Good consultants test thousands of combinations and figure out what the algorithm is looking for. Then they tweak their Web pages to satisfy it. Engineers at the search engines, meanwhile, tinker with their algorithms to put these manipulators in their place and to keep the most relevant sites at the top of their lists. They get feedback on every click. It’s an eternal battle, not only between the search engines and the consultants, but also among the consultants. Some of them game Google better than others.
This is something we humans have been doing since our early days as bipeds. Gaming systems is our specialty. We figure out how things work. Then we calculate the necessary steps so that they work for us. This is true whether we’re putting together an investment portfolio or angling to win “employee of the year.” Each one involves figuring out the ideal recipe—or algorithm—for the intended result. The dynamic hasn’t changed. But these days, more of the tricks involve automatic systems. Time was, for example, that before applying for jobs we would meticulously lay out our résumés with just the right fonts and on the best watermarked paper, to attract the attention of a human resources manager. Now, according to BusinessWeek, 94 percent of U.S. corporations ask for electronic résumés. They use software to sift through them, picking out a selection of “finalists” for human managers to consider. What does the software look for? That’s what we have to figure out. Some pick out certain words—MBA, Harvard, Excel, fluent Mandarin. Others look for more sophisticated combinations. Plenty of consultants are on call to sell us inside tips. The point is that when we want to be found, whether we’re looking for money or love, we must make ourselves intelligible to machines. We need good page rank. We must fit ourselves to algorithms.
REACH INTO YOUR pocket or purse, and pull out that blinking cell phone. Have a good look at it. In the past decade, we’ve come to take these miraculous pocket computers for granted. But they’re bristling with radio signals, sensors, computing power, and storage. In lots of ways, they’re similar to that tennis-ball-sized packet of hardware bobbing around in the stomach of Norman, the fistulated cow. Now imagine we wanted to emulate Norman. What if we used our phones, like Norman’s in-stomach computer, to record our movements and our interactions, and then enlisted some Numerati to create a mathematical profile of each of us? Could we then perhaps find other people with similar patterns? Could those people become our friends and allies—or lovers?
That’s what Nathan Eagle thinks. A few years ago, Eagle, then a Ph.D. student at MIT’s Media Lab, tried an experiment. He distributed cell phones to 100 grad students, one quarter of them at MIT’s Sloan School of Management, the rest at the Media Lab. These phones, he informed the participants, were equipped with software to record their movements and interactions. Over the course of an entire school year, this data would show the researchers not only where the students went and how they communicated with each other, but also who they circulated with, and even who they were spending the night with. This was a privacy invasion huge enough to agitate a congressional oversight committee. But all of Eagle’s subjects signed lengthy consent forms.
By the time I catch up to Eagle, he’s living on the coast of Kenya, working on an education project. Our Skype connection fails, so we chat online. He tells me that he’s watching turtles swimming in to lay their eggs on the beach below, where they’re hard-pressed to protect them from local poachers. “You can get six shillings for these eggs in town,” he writes. I nudge him toward faraway Cambridge, and he tells me about his experiment. Over the school year, he says, it became easy to see that the two groups of subjects—the business school students and the engineers—moved in different patterns. He could predict with greater than 90 percent certainty which type of student each one was. What’s more, he could look at different types of relationships and figure out which people were friends and which were mere acquaintances. If they met at the water cooler on a Thursday afternoon, that was one kind of relationship. If they were together at a bar on Saturday night in downtown Boston, they were much chummier.
Eagle began to build models of the individuals. He started with basic patterns of cell phone use: whether the grad students were at home or at work, if they kept their phone on or off. Each of these variables was called an “eigenbehavior.” (The prefix eigen is a multiplier of an established trend or direction.) It was easy to calculate the mean for each of these behaviors. The users, when charted, fell into clusters. That’s how Eagle distinguished business students from engineers. Even within those clusters, each individual had a unique combination of behaviors. Some slept past noon on Saturdays. Some kept their phones off Sunday mornings (church?). When Eagle mapped them in colorful charts, each individual’s life looked as orderly as the geometric forms of a Navajo rug. They were so regular, in fact, that he could predict with fair accuracy what each person would do next. He could predict where each one would go, who they would call, what time they would turn in for the night, and whether they would
bother turning off the cell phone when they did.
All kinds of organizations are hungry for such data. Mass transit companies want to predict the movements of commuters. Local advertisers, naturally, would love to hit a phone user with an ad for a bar or a restaurant right as they’re getting ready to carouse. And I don’t have to tell you how useful the Homeland Security Department would find this tracking data.
But Eagle has the idea that we can put this data to use for ourselves. He wants to go into the friendship business. Imagine, he says, that we can switch on our telephones to a “promiscuous” setting. This means we’re open to chance encounters. Our phone works as a beacon, sending out our profile in radio waves to those around us. In the early days, this profile will be like the early days of computer dating. It will include a list of our interests. Swedish movies, say, or bicycle touring, French food. And if those whose paths we cross share these of these interests, our profile will pop up on their phones, and we presumably won’t mind at all when one of them touches our elbow and says, “I had a coq au vin to die for at this little bistro . . .” In the workplace, a similar system could alert us to colleagues in the cafeteria who have mastered the Linux operating system or are knee-deep in the genetics of drosophila flies.
But take this a step further. Our movements with a cell phone can paint an in-depth profile for each of us, each one endlessly more detailed than those forms my wife and I filled out for Chemistry.com. If we give them permission to examine us the way Dan Andresen and his team study their cows, they can scrutinize our movements and social networks. They can map the DNA of our behavior. Why would we give anyone a green light to do this? Imagine that they could use this data to find other people whose profiles match our own. Would they become our next friends? The love of our life? That ocean of mobile data may well be the next frontier for Helen Fisher and the other matchmakers of the world.
Already, companies are amassing loads of this data. Consider that cell phone I asked you to pull out of your pocket a few minutes ago. Your phone carrier can detect that it’s sitting right there, unused. And the company has more than enough information to draw powerful conclusions about you and to make predictions (most of them centered on the chances that you’ll jump to another carrier). It’s a potential gold mine of personal data. Phone companies have all they need to track our movements and our social networks. They could analyze the photos many of us send and the words of our text messages. As we surf the Net and begin to use the phone for e-commerce, they learn even more. If they wanted to (and were ready to face a wave of lawsuits from privacy advocates), they could build entire businesses on this rich data. Or perhaps at some point they could package and sell our own data right back to us. That’s Nathan Eagle’s idea. His scheme, which is only in its infancy, is about empowering us—using our data to make ourselves happier, richer, and surrounded by more friends—or perhaps just to know ourselves better.
THE EARLY RESULTS from Chemistry.com cast my marriage in an ominous light. Neither of us pops up on the other’s Chemistry.com radar. On an otherwise bright Sunday morning, I see the service is lining up my wife with a Negotiator-Director in Rosedale, New York. Calling himself “Working Class Hero,” he writes, “I stop to smell the roses to find the beauty that is in all of us regardless of the soil the roses were planted in.” Perhaps those words resonate for my wife, a horticulturist. The chemistry between the two, according to the computer, also bodes a little too well: “With the spontaneity and creativity of the Explorer and the flexibility and imagination of the Negotiator, you’re both in store for some great adventures and hearty laughs together.” I’m willing at this point to recognize Working Class Hero as a worthy rival. But Rosedale? It’s 40 miles away, on the far side of New York City (and within whiffing distance of JFK Airport). That’s a slog. Google tells me that the drive could take “up to 1 hour and 50 minutes in traffic.” Is this logistical nightmare, for all its potential, really preferable to a connection with an Explorer-Negotiator named Stephen who lives in the very same town (and, as luck would have it, in the same house)?
I recheck my profile to see if there’s some detail that’s keeping us apart. And that’s when I see it. When filling out the form, I had carelessly limited my search to women younger than my wife. Silly me. I was blocking the only connection that mattered to me and practically throwing my wife into the arms of more open-minded rivals. It was data under my own control that betrayed me.
I promptly raise my age limit. Within hours, my wife checks the Chemistry.com page, and there I am. Finally, it’s the photo-less Stephen from Montclair, a fellow Explorer-Negotiator wooing her with the same goofy essay about wearing noise-canceling headphones in cafés. Naturally, the service declares this match “great.” And our Explorer-to-Explorer connection brims with promise. “You’d both be happy jetting off to Paris or Nepal at a moment’s notice. And your life in the bedroom is likely to be exciting too.” At this point, my wife designates Stephen as “sizzling”—and the rest of my rivals as “fizzling.” The algorithms have completed their work. The suitors have been banished to Bernardsville, Rosedale, and beyond.
In truth, we set up something of a farce, and it’s easy to laugh about it. But whether our hormones and the length of our fingers had anything to do with it, the service actually made good on its most important challenge. It allowed us to find each other. What we make of the rip-roaring possibilities ahead is up to us.
Conclusion
* * *
ONE SUMMER long ago, Terry Therneau is telling me, he went to the shores of a lake in northern Minnesota to work as a counselor in a summer camp. Therneau is a leader in quantitative biology at the Mayo Clinic, and this detour from medical data into the forests of his youth takes me by surprise. “One of my goals that summer,” he says, “was to learn the name of every tree in the woods.” It seemed like a simple enough proposition. From what he could see, there were a few dozen to learn. He continues: “As I learned the trees, I began to see more and more of them. Pretty soon, my estimate of the woody plants went up tenfold.” Complexity appeared to grow with his knowledge. Now he sees the same phenomenon as he studies the human body. Millions of proteins, all of them interrelated, swarm through our cells like “clouds of gnats.” The more he learns, the more he sees. When Therneau went back home that summer, he still didn’t know all the trees in the northern woods of Minnesota.
The Numerati too are grappling with towering complexity. They’re looking for patterns in data that describe something almost hopelessly complex: human life and behavior. The audacity of their mission is almost maddening. They’re going to figure out who we’re likely to vote for, who we want to work with, perhaps even who we’re best suited to love, all from the statistical patterns of data? It’s the height of presumption, and it leads to humbling disappointments. Like the trees growing in the forests of Minnesota, we confound those who try to categorize us, and we do it most of the time without even trying. Life is complex.
And yet, bit by bit, the Numerati make progress. No, they don’t truly know us, and they never will. But in each domain, they understand and predict our behavior a bit better today than they did last week. They learn from their mistakes. They haul in more data. They continue to experiment. This is a scientific process, and from the laboratories of advertising to counterterrorism, each of us is laid out as a specimen. In some, we’re rendered in fine detail. In others, it’s bare bones. But there’s no turning back from the trend. In the age we’re entering, our lives will be described, studied, and predicted, every day more, through this statistical analysis.
This will lead to all sorts of frustrations. We’ll be confronted, from time to time, with conclusions that are questionable or even flat-out wrong—but delivered with the certainty of science. Sometimes, this will reduce our options. Already, insurance companies equipped with statistical analyses of costs and survival rates are imposing their mandates on doctors, who were once freer to trust their gut. That trend will only grow. And as t
he numbers proliferate and researchers learn more about our DNA, hospitals, insurers, and government agencies will be instructed by automatic systems to discriminate. Like discerning gatekeepers at exclusive clubs, they’ll wave in certain people before putting a hand up and saying, “Not you.”
How do we fight back? With numbers. For this, we have to understand the methods that produce these analyses, and we must master some of them ourselves. In the past, for example, a worker might make the case for a raise with a pithy paragraph in a year-end review. (That’s been my approach.) Now, increasingly, those of us who can quantify annual achievements on an Excel spreadsheet have the edge. In the most dire cases, we’ll hire lawyers who have mastered the tools of the Numerati and who can debunk faulty and self-serving conclusions drawn from statistical curves and correlations. The battle, whether it’s at work or in the courtroom, revolves around the analysis of data.
As this world takes shape, we’ll have to figure out how much of ourselves to hide. Decades ago, I’m told, my sister-in-law grappled with this question. She was stepping out of the shower in the bathroom of her all-women’s dorm, and she heard the call “Men on the floor!” At many schools, this would have been a non-event, but she was in a highly conservative religious college. She was naked. She had only a small towel to cover herself, and there were men prowling the hallways. She could hear them. She waited, but they didn’t go away. So she began to think about which part of her body to cover with the towel. It barely fit across her bottom or her top. It certainly didn’t cover both. She had to make a choice. Finally, she had an inspired idea. She threw the towel over her head—and then scampered naked to her room. Given the options, it was more important for her to cloak her identity than her body.
The Numerati Page 19