IBM's grand plan to save the planet
Here is CEO Sam Palmisano's formula for changing the world: Find problems, throw in billions of dollars in R&D, add consultants and an earnest ad campaign - and watch the profits roll in.http://money.cnn.com/2009/04/20/technology/obrien_ibm.fortune/index.htm
By Jeffrey M. O'Brien, senior editor
Last Updated: April 21, 2009: 3:17 PM ET
CEO Palmisano has shifted IBM from selling tech services to hawking high-tech problem solving.
Gunnar Johansson helps Stockholm reduce car congestion. IBM gear tracks license plates so that the city can charge drivers who use roads at peak hours.
Matiq's Viktor Varan (left) and IBM's Epsen Braathe are trying to use RFID chips to track meat and produce shipments to ensure safe food.
Howard-Yana Shapiro of Mars Inc. is seeking IBM's help in analyzing the cacao genome to ensure the future of the plant, which is being wiped out.
(Fortune Magazine) -- Hurricane Ike rips into southeastern Texas, blowing jagged tree limbs onto power lines all over the Houston metropolitan area. Like firemen responding to an alarm, gumshoe utility workers packing flashlights and rain slickers scour the region for broken branches as 2 million homes sit in the dark. In some areas the search takes weeks. The total cost of the outage runs in the billions.
Half a world away, Stockholm bogs down in rush-hour traffic. A series of bridges connecting Sweden's capital creates bottlenecks that cause gridlock and air pollution, waste millions of gallons of fuel, hamper public transportation, and endanger pedestrians.
Just over the border in Norway, an E. coli outbreak devastates the entire nation. One child dies, nine others are hospitalized, and a country known for the quality of its food now eats in fear. Meanwhile, in West Africa a fungal disease and global warming threaten the world's largest repository of cacao trees, holding the region's primary economic engine hostage - not to mention humanity's sweet tooth.
From one view, this global tour of woe shows our world buckling because of poor planning, too much demand, and strained infrastructure. But there's a more opportunistic way to look at the vignettes. From another angle it becomes clear that they're all propagated by a single culprit: a lack of quality information.
Gridlock is what happens when drivers aren't aware of congestion or don't know how to avoid it. How do we stop E. coli? Cut off the bacteria at its source - if only we had a transparent supply chain. The same goes for power outages. If we could accurately monitor power flow across the electricity grid, the location of a short would become apparent. As for the tree that gives us chocolate, well, with a clear fix on its genome, creating a disease-resistant, heat-tolerant supercacao would be as easy as cross-breeding garden tomatoes.
In the parlance of the information technology industry, these situations all represent "dumb network" problems. The term sounds pejorative, but it simply means that we don't truly understand commuter traffic or electricity flow or the inner workings of the cacao genome, and as a result our highways, utility grids, and cash crops are not managed as effectively as they could be.
The good news is that we now have the technology to convert these analog distribution systems into multidirectional "smart" networks. Readily available sensor technologies like RFID chips and digital video can track movements in granular detail. Cheap data storage, powerful analytics software, and abundant computing capacity give us the ability to warehouse and make sense of all that information. With the knowledge we're gaining, we can remake our world in a more efficient way.
If you've watched a televised sporting event or read a major newspaper, magazine, or website in the past six months, you know that this is precisely what IBM has been up to lately. With the exuberance of a fast-food commercial and all the earnest multiculturalism of a Benetton ad, its "Smarter planet" campaign celebrates Big Blue's ambitious efforts to tackle some of the most vexing dilemmas of our era. What the ads don't say is that IBM is making a killing doing it.
Which may be why the company's president, chairman, and chief executive officer, Sam Palmisano, is in such a cheery mood. Times are tight for everyone, and IBM (IBM, Fortune 500) isn't immune to the severe economic downturn. The company has laid off thousands in recent months and has been criticized for shipping many more jobs to India and elsewhere. But it also handed out bonuses to 386,000 employees this year while announcing blowout financials: In 2008 the company earned $12.3 billion on $103.6 billion in revenue, with a 44.1% gross profit margin - new high-water marks in each category.
Fresh off January appearances with President Obama in Washington, Palmisano is putting the finishing touches on his "Letter From the Chairman" for the annual report. It may be a frigid February morning in Armonk, N.Y., but the note reads like a Palm Springs weather forecast: blue skies and sunshine. When you log three financial records, he says with an exaggerated smile that pushes his eyes into a squint, "it's hard to be depressed."
Credit crunch? Ha! IBM is its own bank. "You're sitting on $14 billion. What are you going to do with it? Two hundred and fifty thousand people are going to get raises in the next couple of months. The executives won't - but that's fine. We make enough money!"
IBM's stellar earnings are rooted in a strategic shift implemented earlier in this decade. Former CEO Lou Gerstner gets credit for saving the company from ruin by steering away from hardware and into services and software in the early '90s. But Palmisano doubled down. In the wake of the dotcom implosion, he sold the ThinkPad division to Lenovo, exited disk drives, and moved aggressively into global consulting and data analytics by acquiring nearly 100 firms, including PricewaterhouseCoopers Consulting and Cognos.
All told, Palmisano has spent $50 billion on acquisitions and R&D in preparation for a seismic shift in IBM's business. Now the pieces are coming together. Roughly 30% of IBM's $6 billion annual R&D budget goes into long-term research, and the department churns out more than 4,000 patents a year. The consultants mine that output for so-called repeatable assets, basically problem-solving technologies that can work in various settings.
"We spend an enormous amount of our resources on exploratory work," says John Kelly, senior VP of IBM Research. "A day doesn't go by that someone in sales doesn't see me and my team. It's pretty hard to find a subject where we don't have really deep expertise." In 2004, 80% of the consulting business in the utilities sector came from the cutthroat and mind-numbing world of enterprise software implementation. Today 80% comes from the type of high-margin, world-changing program born out of the heads of brainiac researchers.
So Palmisano is encouraging his employees to think even bigger, to scout out any dumb network that can be made smarter. Because, as any self-respecting capitalist knows, in great pain lies dormant profit. "We are looking at huge problems that couldn't be solved before. We can solve congestion and pollution. We can make the grids more efficient," he says. "And quite honestly, it creates a big business opportunity."
Darkness comes early on a winter afternoon in Stockholm; so does rush hour. It's almost 4 p.m., and Gunnar Johansson is standing roadside, hands in the pockets of his long trench coat. Vehicles speed by as he gazes quietly at a bank of overhead cameras. A warm smile stretches across his face. Johansson is a transportation economist by training, but to the people of Stockholm he's the magician who made the city's traffic problem disappear.
The average New Yorker or Angeleno might not have noticed anything unusual about the Swedish capital circa 2003, but Swedes pride themselves on an optimum quality of life, so government officials put out a request-for-proposal on a congestion-pricing scheme that would charge vehicles for passing through city limits at various times of day. The goal was to reduce the number of vehicles downtown by 10% to 15% at peak hours. "In many cases you introduce road charges to get money out of the system," says Johansson, a 6-foot 6-inch Swede who lately has been traveling the world (in coach) to speak about the success of the system, which represents the most mature case study in the "Smarter planet" initiative. "Here, that was not the objective. This was about making the transport system work better."
Johansson joined IBM from PricewaterhouseCoopers after the acquisition, just as the Stockholm project was going out for bid. He led IBM's winning proposal, using 18 subcontractors to implement the scheme. To assuage negative public and media sentiment, the government agreed to hold a referendum to determine the program's fate after a six-month test in early 2006. Facing such pressure, officials demanded that the system accurately identify more than 99% of all vehicles. Johansson's team couldn't require drivers to carry FasTrack-style transponders, so they tried overhead cameras using optical character recognition (OCR) software. It's not new technology, but identifying half a million cars a day traveling 60 mph with near-perfect accuracy is a big leap from placing a letter on a flatbed scanner. IBM tested the cameras using a partner's software - without much luck. "We were down to 60% accurate," says Johansson. "We thought, We're never going to make it."
The team's technical architect put in a call to the IBM R&D facility in Haifa, Israel, which had been working on new OCR software, and designed a system to combine it with two commercial products. The team placed two cameras above each lane, facing opposite directions to capture front and back plates. Each camera's vision extended into adjacent lanes for multiple vantage points on every vehicle. Within months IBM had reached 99% accuracy. A small team of humans winnowed misses even further - down to four or five a day out of almost 500,000 vehicles. The information is processed at a data center in Copenhagen, and drivers receive monthly bills in the mail. In the beginning, IBM contracted a call center to handle complaints. The calls never came. "The tax authority even hired 40 lawyers to deal with expected appeals," says Johansson. "They got six appeals."
Of course the objective was not to charge vehicles but rather to decrease traffic, and that's what the system did. Total traffic fell 35% almost immediately and stayed down 22% - and not just at peak times or solely downtown. Emissions also dropped by 14%. The streets became more pedestrian-friendly, and the buses began finishing their routes so quickly that the city had to rewrite the schedules. "For people working and living here, it was a different city. The government had lots of researchers doing surveys to identify the impact," he says. "That wasn't needed."
The system passed in referendum, and the city environmental director who championed it, Gunnar Söderholm, couldn't be more pleased. At a total cost of about $400 million, a third of which went into bolstering public transit, the city improved quality of life. "Some people thought that was too expensive, but this is one of a few examples of infrastructure investment that has immediate payback," says Söderholm at Stockholm City Hall, site of the annual Nobel Prize gala. "We started here with an empty table, designed the scheme, developed a procurement process, had the trial and referendum - all within four years. We've been discussing building a new bypass for four decades, and it's still not decided."
One curious result: Public transit ridership jumped by almost 40,000 users a day, accounting for only a third of the drop in traffic. Some economists worried that fewer inbound trips would mean reduced inner-city commerce, but that hasn't happened. Based on analyzing all the data, Johansson believes he's proved what he's long speculated: Reducing traffic isn't about building more roads, it's about providing better information and choices.
Unlike London's congestion-pricing scheme, which charges roughly $12 to enter the city at any time of day, Stockholm imposes a graduated toll, starting around $1 and rising to around $3 during high demand. The fee schedule makes it obvious when traffic will be the worst, so drivers who trek in during peak hours know they'll pay more for what will probably be a maddening experience. As a result, people seem to be cutting out unnecessary trips: bundling afterschool pickups, say, with visits to the grocery store. In short, Swedes are driving smarter.
Which means that Stockholm provides a lesson for more than just cities like San Francisco that are considering their own congestion-pricing schemes. Johansson imagines one day tracking every car by GPS. Such a system would not only reduce traffic by offering real-time warnings and alerts but could also refine our method of funding road improvements. Rather than an across-the-board gas tax, fees could be levied according to mileage traveled. But in the shorter term, he has some work to do in London. The city is unhappy with the way its own congestion-pricing scheme, implemented by British consultancy Capita Group, is faring. In November, IBM will take over the project.
Last summer Barack Obama campaigned on a promise to digitize the electricity grid. Recently his economic stimulus plan allocated billions to make that promise a reality, and hardware providers, Internet companies, software firms, utilities, and the media have all rallied around the idea. Much of the hype has centered on installing smart meters - basically fancy thermostats that monitor energy use in real time - in our homes.
Sadly, the only information most of us get about our energy consumption comes in the form of a monthly bill, when it's too late to do anything about our behavior. Seeing real-time usage data on the meters, along with options for cheaper times of day to, say, do the laundry, the thinking goes, will encourage us to stop hogging peak kilowatts just as the Swedes curbed unnecessary rush-hour travel. Certainly that's part of the promise of an intelligent grid. But what if the grid were so smart that it could differentiate types of energy, price them accordingly, and heal its own wounds? What if it could actually think for itself?
Utility executives come from around the world to see the demonstration that Don Cortez, a VP of CenterPoint Energy, is performing at the company's technology center just outside Houston. In a typical outage, workers have two choices: to search chaotically for the source of the fault or to wait for customers to complain, and then triangulate the source of the outage based on the geography of the calls. Either way takes a long time.
"We serve a 5,000-square-mile area, which is fairly compact by utility standards," says Cortez, surrounded by a roomful of miniature utility poles, overhead wires, lights, speakers, a pickup truck, and a handful of flat-panel displays that together represent an end-to-end model of a smart grid. "When you have a large thunderstorm, it impacts a large portion of our consumers." In the demo - a flash of light and the familiar crack of a thunderstorm, followed by a mechanical tree branch falling on a wire - the grid tells headquarters which lines have been affected, automatically re-routes power, and dispatches maintenance workers to an exact location. Total outage time: about 10 seconds. In a region plagued by outages, minimizing downtime in this way will save millions.
IBM got involved in the CenterPoint project about three years ago. Cortez was brainstorming possible revenue streams with Michael Valocchi, head of IBM's smart-grid practice, who has been working with utilities for 23 years, the first 17 with PricewaterhouseCoopers. They decided there could be huge money in tackling the outages dilemma. "I knew if we could solve that issue, it was something that I could make commercially viable," says Valocchi, speaking from India, where he's working on another of IBM's 50 smart-grid projects. "It was way above my skill set, but there were a lot of researchers who were interested in it."
At that point IBM invested in the self-healing concept and helped build the demo. Such a partnership is known inside IBM as a "first of a kind" project, or FOAK. According to Valocchi, about half of all FOAKs result in a commercial deployment. In this case he called on R&D for everything from complex algorithms that govern the way power gets re-routed to predictive weather-modeling software. The hope is that the company will sell the self-healing solution to utilities around the world. "I think we can get to a commercialized version," Valocchi says. "We're probably three months away. We'll sell it almost like we sell software - a license with implementation services."
Both IBM and CenterPoint are interested in smart meters. Cortez's team rolled out 5,000 smart meters across the utility's service area in March and plans to deploy 2.1 million more over the next five years. The devices work on open protocols and will report usage information to whomever the consumer allows. The idea is that third parties will build custom applications that enable consumers to regulate usage online, at Google PowerMeter, for example. And it shouldn't be long before an iPhone app links a GPS transmitter to the meter, causing the air conditioning to turn on 20 minutes before the owner walks in the door.
Once fully implemented, the two-way meters will also help CenterPoint divine behavioral patterns. The more electric companies understand about consumer demand, the less they'll be likely to overbuild highly polluting power plants, which cost as much as $600 million each. "A lot of the benefits are really going to be cost-avoidance," says Cortez. "With added intelligence, we won't have to build as many substations."
A digitized grid also makes it possible to upload alternative energies easily, whether from the massive wind farms in West Texas or from all the solar panels popping up on suburban rooftops. Which is nothing but good news for the most polluted city in the country. "Houston doesn't meet the federal air-quality standards or the ozone-level standards. That restricts the ability of the city to grow," says Tom Standish, CenterPoint's group president of regulated operations. "That's not good for us, and it's not good for the economy."
What could be more convoluted than a 100-year-old electromechanical power grid? Try food distribution. A swank restaurant may brag that its 13-ounce rib-eye special, which came from a hand-massaged steer that dined on a grassy hillside in Marin County, traveled to your plate in a van fueled by biodiesel. And if you dine like that at every meal, you'll be fine. But for a more realistic view of what we know about the origins of our nutrition, think about what happens whenever there's a bacteria scare: Stop eating tomatoes from Mexico! No, New Mexico! No, California! Wait, California tomatoes are okay, but the spinach is bad. Or maybe it's the escarole. Stop. Every jar of peanut butter must be yanked from every supermarket shelf in the country. Never mind your concern about whether that avocado is really organic or whether the tuna is truly dolphin-safe. We can't even be completely sure your food won't hurt you.
The mission of IBM's food-traceability initiative is to lift the veil off the entire food chain. The lingo is "farm to fork," but the effort actually extends further in both directions. Researchers and consultants are on one end sequencing the cacao genome to learn how to increase yields, and on the other end coming up with ways to monitor the waste stream to see how our diets are affecting the planet.
One of IBM's most interesting early efforts is in Norway. When a 2006 E. coli outbreak killed one child and hospitalized nine others, ground-beef sales were halted. Weeks later the real culprit turned out to be cured sausage. Even in a country where all cattle and swine are registered, the distribution system remained opaque. Months after the outbreak, the E. coli was finally traced through a slaughterhouse owned by meat producer Glide to the farm where the outbreak originated. "It shattered the illusion of Norwegian food being the safest in the world," says Espen Braathe, an IBM consultant in Oslo, who has begun designing, per government mandate, a transparent nationwide food network. "We worked to see how we can build a completely traceable chain. The aim is to have it up and running by the end of 2010."
As a starting point, Braathe partnered with Norwegian IT consultancy Matiq to outfit Glide's slaughterhouse in Stavanger with RFID chips and readers. All three companies agreed to give me a tour of their early progress. Wearing the type of bunny suits usually associated with semiconductor fabs, we proceed from the back of the production line, where sanitization is most important, toward the front. At various points we cross into a new hygiene zone and change suits.
Along the way Braathe points out the chip readers and demonstrates how the tags, which cost less than a penny each, are melted into the plastic bins that transport the meat throughout the factory and to grocery stores. Eventually we reach the front, where the smell of scorched hair hangs over freshly killed hogs, each with a tag in one ear. Workers with enormous chainsaws split the animals in two from tail to neck and strip out entrails with the nonchalance of college kids folding sweaters at Gap.
Once fully deployed, the system will reveal the origin, delivery date, transportation mode, and destination of any package of meat. The implementation is based on anticounterfeiting schemes that Braathe helped develop for the pharmaceutical industry, and will work much like online checking. Every participant will have control over proprietary data, and a governing body will assess everything in an emergency. In the case of a crisis, this new transparency promises to save time, lives, and a lot of money for innocent companies, which won't be caught up in sweeping recalls.
The chips should also improve productivity for the slaughterhouses. Each workstation has an embedded chip, allowing Glide to track and analyze employee performance. Putting tags in the animals' ears also streamlines the slaughtering. "It reduces the number of people by half. You don't have to have large storage places to sort the animals," says Viktor Varan, a director of marketing with Matiq, with a nervous laugh. "You just take the first one who wants to die."
The next step is enlisting supermarkets and tracking the meat all the way to the fork. But Braathe is already thinking about what other problems he can solve. He talks about the vast environmental damage caused by antibiotics in our waste stream as a result of overuse of prescriptions and antibacterial soap. This is a matter of chemicals, of course, which can be readily monitored and analyzed with various sensor technologies. Meanwhile, his colleagues in the U.S. are moving in the opposite direction, trying to understand our food long before it even makes it to the farm.
Like many monoculture cash crops, cacao has been devastated by blight in parts of the world. The vast majority of remaining cacao is farmed in West Africa. Lately yields have been declining, and experts consider the region highly vulnerable to the same fungus that wiped out South American cacao. So Howard-Yana Shapiro, global director of plant science for Mars and an adjunct professor at the University of California at Davis, decided to sequence the crop's genome in hopes of solving these problems.
"Think about the tragedy that would happen if those diseases or others were to leap the Atlantic and knock out 70% of the world's production," he says at his California home. Shapiro hasn't shaved since 1966 and sports a long white beard that garners attention every December. But sitting among his collection of vintage Indian motorcycles and biology textbooks, he seems more Gandalf the Wise than Santa. "We're the world's largest chocolate company," he adds. "We have a responsibility."
Mars knows chocolate, of course, but doesn't have the requisite supercomputing power to analyze the interplay of the 415 million base pairs that make up the cacao genome. So Shapiro visited legendary IBM researcher Mark Dean, one of the inventors of the original IBM PC, at the company's San Jose R&D research lab. They discussed how sequencing and analyzing the genome could minimize pesticide use, improve the livelihoods of 6.5 million farmers, and even stimulate science around the world. With open access to all the data, researchers in Africa and elsewhere will become Gregor Mendels at hyperspeed. Rather than waiting years to see how cloned trees will perform in shifting environmental conditions, cross-breeding will happen at the genomic level, and growers will know in advance that their trees will, for example, show resistance to a particular fungus or perform better in arid soil. "Imagine if you could get three times as much money from one-third as many trees. You'd have space for more crops and could build a stronger portfolio," Shapiro says. "This would stabilize the economies of West Africa and East Asia."
Dean signed on in a heartbeat. The $10 million project is being funded by Mars, and after only six months the team is well ahead of schedule. "We hope that within five years we'll find enough information to make the plant more sustainable," says Dean. "This is one of those projects that is both humanitarian and economic in nature. It will affect the development of the African region, and it allows us to do some research that will expand our knowledge of how [sequencing is] done."
Its sheer ubiquity has made the "Smarter planet" campaign conspicuous over the past few months. And the ads stick out even more because of their tenor. They're rather preachy. But unlike many highly optimistic tech campaigns (e.g., the old AT&T "You will" ads), they focus on our ability to improve the world now. Every last piece of news, it seems, relates to a looming catastrophe. Meanwhile smiling heterogeneous faces tell us that IBM has already begun fixing what's been broken for far too long. This is no mistake. "Over the past five years we've probably spent $30 billion in R&D and $20 billion buying companies to provide these capabilities," says Palmisano. "Now is the ideal time to invest, oh, hundreds of millions into the marketing. People are willing to do things differently because we have a severe economic situation."
"Smarter planet" is also a very public lobby for economic stimulus money. And IBM is not alone in this regard. GE (GE, Fortune 500), which is working on several smart-meter deployments and wind farm projects, has amped up its "Ecomagination" campaign. More situational opportunists are on the way - from Hewlett-Packard (HPQ, Fortune 500) and Cisco (CSCO, Fortune 500) to Fujitsu and Autodesk (ADSK) - all churning out press releases and marketing materials about how they, too, are using technology and intelligence to improve our world.
This self-promotion is a good thing. Not only are the companies spending at a time when our economy needs it, but they're heralding a new dawn. We've embarked on a phase of technology in which everything has become a computer. Data flow from all directions, and for the first time we can capture them. IBM is the early leader in this new era.
Leadership positions, as the company knows all too well, come and go. But with luck, the tone of "Smarter planet" will remain. The message - that technology can be deployed to greater ends than creating the next fetishized cellphone - is bigger than any single company. And so, too, is Palmisano's epiphany. He deftly led IBM out of the dotcom doldrums. Perhaps more important, he has revealed a model for monetizing scientific research in a way that benefits humanity.
Sure, not everyone can afford $6 billion a year for R&D. But real innovation rarely comes from big, rich companies. With luck, IBM's ad campaign, coupled with its blowout 2008, will call scientists and entrepreneurs to arms. They'll see our archaic global shipping infrastructure, a dilapidated educational system, disappearing honeybees, the fraud on Wall Street, and think, I know how to fix that. And I can make a killing doing it.