Imagine that your business had a complete log of your customers' wanderings"”every trip to the grocery store, every work commute, every walk with the dog. What could you learn about them‾ Armed with that knowledge, what sorts of goods and services might you try to sell them‾ Just as important, if you made your best pitch"”relevant and timely, of course"”would customers concerned about privacy tell you to get lost‾ This isn't science fiction. A nascent industry extending from the laboratories of Google (GOOG) and Nokia (NOK) to a host of data-fueled startups is wrestling with these very questions.
On a snowy winter evening in New York's SoHo neighborhood, a small team of analysts at a startup called Sense Networks is poring over the movements of nearly 4 million cell-phone users over the course of a year. They have been tracked by global positioning systems, by cell towers that catch their signals, or by local Wi-Fi networks that detect their presence. As far as the Sense analysts can see, these people have no names: They are simply dots moving across the maps on Sense's computers. (The data trove comes from a company New York-based Sense will not name.)
Much can be learned, it turns out, from the patterns of those dots moving across the maps. It's possible to see clusters grow around a popular restaurant or retail store. It's also easy to learn about each dot. Business travelers tend to congregate in certain spots in each city. The newly unemployed often shift from the clocklike routine of work to far more random movements. And Sense can flesh out these digital stick figures with additional data. By noting where dots appear to sleep, the company can assign an average neighborhood income to each one. It then becomes easier to predict whether those spending time near car lots are in the market for luxury brands or economy models.
Laboratory of humanity
After a few weeks of monitoring one dot, the Sense computer usually has enough data to place it into a tribe"”a group of people with common behaviors. One tribe comprises night owls who explore bars and restaurants into the wee hours. Sense founder Greg Skibiski and his chief scientist, Tony Jebara, who is also a Columbia University computer science professor, call them 'Young & Edgy.' Another group seems at first to overlap with the Young & Edgy. These people, though, stay true to a single establishment and return home at more regular hours. 'Barflies,' says Jebara.
With every step they take, the members of Sense's tribes are helping to create a new laboratory of humanity on the go. This emerging phenomenon, powered by Web phones and an explosion of new mobile-software applications, is the long-awaited Next Net. 'The phone in your hand is the bridge between the virtual and real worlds,' says Michael Halbherr, vice-president of Nokia's gate5 mobile Web unit.
Sense, led by the 35-year-old Skibiski, is a mere gnat in this market. It's a services shop powered by five PhDs and a slew of algorithms. Phone companies and advertisers provide Sense with raw data on people's movements and behavior. Sense's mission is to transform mountains of data into intelligence: what individuals will be most likely to buy, or where they'll be when a craving hits. The company is looking for venture money and may struggle in the current rough market. But whatever Sense's fate, the method of its research"”predicting people's preferences by their movements"”could soon work its way into industries from marketing and finance to media.
Marketers have long dreamed of zeroing in on shoppers, whether in a mall or a competitor's store, and hitting them with targeted ads or coupons. (The privacy implications are a big deal, as we'll see.) But the business ramifications of the Next Net stretch beyond marketing.
Mobile data also promise to help researchers fine-tune transit systems, study the spread of crime or disease, and even monitor and optimize the movements of workers. For many businesses, the coming flood of mobile information could bestow a competitive edge. A couple of Sense's primary backers are hedge funds looking to minute-by-minute data on behavior to guide investment decisions.
The Google strategy
To understand what Sense is doing, it's easiest to think about"”what else‾"”Google. The search engine has dissected the behavior of Web surfers, including which sites they visit and which they link to. With this knowledge, Google proceeded to develop a map of the influence and relevance of billions of Web sites. It's the heart of its search and advertising business.
The effort under way now, at Sense and other companies, is aimed at creating a similar map of the physical world. Think of each bar, restaurant, arena, or street corner as a Web page. Each person who goes there has, in a sense, 'voted' for the location. Some spots, like Times Square, are immensely popular"”perhaps the physical world's equivalent of the Yahoo! (YHOO) home page. Others, such as a restaurant in northeastern Vermont, are like an obscure blog. But the people who land in two places at similar times might have certain traits in common.
Eventually these insights may lead to a new kind of cartography. The evolving maps will identify places not by their roads, forests, or mountain ranges but by what kinds of people spend time there and by what they do. If the efforts succeed, these pulsating flows of information are likely to become a vital resource for every company that wants to put up a sign, sell a drink, rent a room, or fill a stadium.
But how to make money from these maps‾ Sense started last summer with a consumer application, Citysense, in San Francisco. Subscribers who downloaded the software to their mobile phones agreed to be tracked and placed into a tribe. They could then locate like-minded people. A Young & Edgy user looking for company at 1 a.m., say, could open a city map on her phone to find which clubs were pulsing with fellow red dots. After a couple of months, though, Skibiski's team concluded a consumer-led business would place Sense in the crosshairs of privacy advocates"”and it wasn't likely to pay the bills.
They focused on crunching mobile data for businesses, including advertisers. Kinetic, the outdoor advertising unit of ad giant WPP, studied Sense's data in San Francisco and saw that one tribe frequented bars in the Marina district where a certain beer promotion (which it will not disclose) did well. They pushed the beer company to extend the promotion to other bars in the city that attracted dots of the same color. The early results look promising. 'We have been experimenting with mobile data for three or four years,' says Steve Ridley, Kinetic's chief executive officer. 'It's now reaching a stage where it genuinely affects our core business.'
The process took a while. A decade ago, during the dot-com and telecom boom, marketers extolled the potential of the mobile Web. They talked of zapping customers with digital coupons just as they passed nightclubs and cafÃ©s. The decline of marketing's traditional demographics"”age, gender, income, and ethnicity"”was inevitable, they predicted, once they could focus on the individual.
This vision fell flat in much of the world, largely for two reasons. First, the early Internet phones offered only rudimentary location-tracking and data service. More important, few customers wanted ads popping up on their phones, especially if they had to pay for them.
The promise of change arrived two years ago in a shiny stainless steel case with a touchscreen. Apple's (AAPL) iPhone, the first computer phone to score a hit among nongeeks, quickly spawned a plethora of me-too devices.
All enticed users to shift much of their computing to mobile machines. The result: an explosion of information about people on the move.
Every bit as important as the phone is the software. Apple invited independent developers to come up with applications for its phone. Everyone from tiny game makers to Bank of America (BAC) could set up shop on the iPhone platform. The App Store now offers more than 8,000 programs, including many that use location to provide services such as recommendations for nearby restaurants. Google, Nokia, Research In Motion (RIM), and Microsoft (MSFT) have followed suit with their own app shops.
'Incredibly rich information'
Every time a user clicks on an application, whether it's to turn a phone into a radio or make a bid on eBay, the time and place of the event zips straight to the company selling the service. Certain phone manufacturers can also peek at this data, depending on the handset. Naturally, the wireless service provider also sees it and can place it into the context of the user's other behavior, from physical movements to calling patterns. While phone companies have long had a line on customer behavior, the applications add crucial perspective by pointing directly to each person's interests and needs. (Barflies, Sense researchers found, spend more time than others playing an alcohol-themed game on their handsets.) 'All of a sudden we have this incredibly rich information on how and where people use their mobile applications,' says Ted Morgan, chief executive of Skyhook Wireless, a provider of tracking technology.
It's also becoming easier to pick up a handset's digital trail. Traditionally, wireless carriers have marked our wanderings only by the nearby cell towers receiving our signals. Each phone, even at rest, stays in touch with those towers so it can send and receive calls. But the towers can miss a person's location by several hundred yards. Satellites are more precise, but they often don't work when people are under roofs. Many of the latest phones, including the iPhone, have Wi-Fi, the radio signals used in home networks. These signals often can pinpoint someone within 33 feet.
On a stroll through a mall, a phone beams its presence to dozens of Wi-Fi networks in stores. Skyhook, which locates millions of wandering customers for companies offering mobile services, has tied together 100 million Wi-Fi access points around the globe. It'll have plenty of machines to track: Some 56 million Wi-Fi-equipped handsets were shipped in 2008, according to the Wi-Fi Alliance.
While the technology pushes ahead, the second issue"”whether people will accept location-specific marketing on their phones"”remains touchy. Sense goes out of its way to avoid personal data, even though it would be a cinch to figure out the identities of many dots. 'It's not worth the risk, and we don't need that information,' says Skibiski. The consensus among marketers is that consumers won't stand for an invasion of targeted ads on their phones unless they have asked for them. Fearful of a privacy backlash (and resulting government regulation), most advertisers are holding off on personalized ads and coupons. Instead, they're focusing on customer service, an area in which mobile technology has a track record.
General Motors (GM) provides an early example. Some 5.5 million subscribers in North America pay for its OnStar (GM) global positioning service.
They're willing to supply data on their whereabouts in exchange for safety and convenience. With a phone, far more possibilities open up. An Amazon.com (AMZN) application lets users take photos with their iPhones of things they'd like to buy. Amazon's computer promptly searches for each item. When it finds a match, it lets the user know the lowest price and nearest location. In the process, Amazon gains data on the movements and habits of its most zealous customers.
In this dawn of geo-tracking, many people are more likely to start out swapping data with friends and families. Already 25 million people have installed Facebook on their handsets. And Google is pushing in the same direction. On Feb. 4 the search engine released Latitude, an app that lets people with high-end mobile phones share location data with friends. Users can check on their buddies stuck in traffic, sneaking out of work early, or catching a ball game. Latitude even provides directions on how to meet up with friends and contacts. (Instructions for finding Skibiski one recent afternoon involved taking the E train from midtown Manhattan to West Fourth Street and walking east on Bleecker Street.)
The privacy implications are considerable. Is it O.K. for a boss to hand an employee a Latitude-loaded BlackBerry and then monitor her whereabouts‾ Companies that operate fleets of trucks have tracked employees for years. But similar technology in cell phones would potentially let all sorts of companies monitor and measure employee movements. Latitude does offer cloaking options. A user can hide from certain people or ask to be located by city, not by street.
Resistance to mobile ads is showing signs of breaking down, at least in some quarters. When surveys ask users if they're interested in receiving the ads, nearly everyone says no, says Greg Sterling, a senior analyst at Opus Research. But when asked recently whether they would welcome messages only from local businesses they select, 43% say they'd be 'very or somewhat interested.' That number, says Sterling, is higher among the data hounds who use the iPhone, BlackBerry, and other high-end phones.
Sense's hedge fund investors"”the company does not name them"”want to study the mobile data to get a leg up on macroeconomic trends. How is traffic at key malls‾ Is that struggling retailer attracting customers with its new marketing campaign‾ 'Usually you get store sales data with a couple weeks' lag,' says one of the investors.
Some companies are rewriting their business plans based on the possibilities for location-based services. Nokia, the leading cell-phone maker, faces a slumping global market, with phone shipments expected to fall 5% to 10% this year. But perhaps no outfit has access to more location data on hundreds of millions of people. By working with Sense to study people's movements, their tribes, the photos they take, and the applications they use, Nokia is betting it will come up with customized suggestions on everything from walking routes to boutiques. All of them would be linked to a map on the phone. Money would come from subscriptions or advertising.
Back at Sense, analysts are busy crunching data for Nokia. At this early stage, says Jebara, you can't always tell whether a person is in a coffee shop or the bar across the street. The geographic read isn't precise enough. So the machines make a guess based on the person's tribe and the time of day. No doubt they sometimes get it wrong. But here's the crucial point: The more we use the mobile Net, the more data we provide. And with more data, the Next Net develops more fully"”and becomes increasingly vital for business.
For a video of Sense Networks and a look at top mobile applications, go to businessweek.com/go/09/futuretech
Baker is a senior writer for BusinessWeek in New York.