India could never have afforded to pay for the bandwidth to connect brainy India with high-tech America, so American shareholders paid for it. Sure, overinvestment can be good. The overinvestment in railroads turned out to be a great boon for the American economy. “But the railroad overinvestment was confined to your own country and so too were the benefits,” said Singh. In the case of the digital railroads, “it was the foreigners who benefited.” India got to ride for free.
It is fun to talk to Indians who were around at precisely the moment when American companies started to discover they could draw on India’s brainpower in India. One of them is Vivek Paul, now the president of Wipro, the Indian software giant. “In many ways the Indian information technology [outsourcing] revolution began with General Electric coming over. We’re talking the late 1980s and early ‘90s. At the time, Texas Instruments was doing some chip design in India. Some of their key designers [in America] were Indians, and they basically let them go back home and work from there [using the rather crude communications networks that existed then to stay in touch.] At that time, I was heading up p. 106 the operations for GE Medical Systems in Bangalore. [GE’s chairman] Jack Welch came to India in 1989 and was completely taken by India as a source of intellectual advantage for GE. Jack would say, ‘India is a developing country with a developed intellectual capability.’ He saw a talent pool that could be leveraged. So he said, ‘We spend a lot of money doing software. Couldn’t we do some work for our IT department here?’ ” Because India had closed its market to foreign technology companies, like IBM, Indian companies had started their own factories to make PCs and servers, and Welch felt that if they could do it for themselves, they could do it for GE.
To pursue the project, Welch sent a team headed by GE’s chief information officer over to India to check out the possibilities. Paul was also filling in as GE’s business development manager for India at the time. “So it was my job to escort the corporate CIO, in early 1990, on his first trip,” he recalled. “They had come with some pilot projects to get the ball rolling. I remember in the middle of the night going to pick them up at the Delhi airport with a caravan of Indian cars, Ambassadors, based on a very dated 1950s Morris Minor design. Everyone in the government drove one. So we had a five-car caravan and we were driving back from the airport to town. I was in the back car, and at one point we heard this loud bang, and I thought, What happened? I shot to the front, and the lead car’s hood had flown off and smashed the windshield—with these GE people inside! So this whole caravan of GE execs pulls over to the side of the road, and I could just hear them saying to themselves, ‘This is the place we’re going to get software from?’ ”
Fortunately for India, the GE team was not discouraged by the poor quality of Indian cars. GE decided to sink roots, starting a joint development project with Wipro. Other companies were trying different models. But this was still pre-fiber-optic days. Simon & Schuster, the book publisher, for instance, would ship its books over to India and pay Indians $50 a month (compared to $1,000 a month in the United States) to type them by hand into computers, converting the books into digitized p. 107 electronic files that could be edited or amended easily in the future—particularly dictionaries, which constantly need updating. In 1991, Manmohan Singh, then India’s finance minister, began opening the Indian economy for foreign investment and introducing competition into the Indian telecom industry to bring down prices. To attract more foreign investment, Singh made it much easier for companies to set up satellite downlink stations in Bangalore, so they could skip over the Indian phone system and connect with their home bases in America, Europe, or Asia. Before then, only Texas Instruments had been willing to brave the Indian bureaucracy, becoming the first multinational to establish a circuit design and development center in India in 1985. TI’s center in Bangalore had its own satellite downlink but had to suffer through having an Indian government official to oversee it—with the right to examine any piece of data going in or out. Singh loosened all those reins post-1991. A short time later, in 1994, HealthScribe India, a company originally funded in part by Indian-American doctors, was set up in Bangalore to do outsourced medical transcription for American doctors and hospitals. Those doctors at the time were taking handwritten notes and then dictating them into a Dictaphone for a secretary or someone else to transcribe, which would usually take days or weeks. HealthScribe set up a system that turned a doctor’s touch-tone phone into a dictation machine. The doctor would punch in a number and simply dictate his notes to a PC with a voice card in it, which would digitize his voice. He could be sitting anywhere when he did it. Thanks to the satellite, a housewife or student in Bangalore could go into a computer and download that doctor’s digitized voice and transcribe it—not in two weeks but in two hours. Then this person would zip it right back by satellite as a text file that could be put into the hospital’s computer system and become part of the billing file. Because of the twelve-hour time difference with India, Indians could do the transcription while the American doctors were sleeping, and the file would be ready and waiting the next morning. This was an important breakthrough for companies, because if you could safely, legally, and securely transcribe from Bangalore medical records, lab reports, and doctors’ diagnoses—in one of the most litigious p. 108 industries in the world—a lot of other industries could think about sending some of their backroom work to be done in India as well. And they did. But it remained limited by what could be handled by satellite, where there was a voice delay. (Ironically, said Gurujot Singh Khalsa, one of the founders of HealthScribe, they initially explored having Indians in Maine—that is, American Indians—do this work, using some of the federal money earmarked for the tribes to get started, but they could never get them interested enough to put the deal together.) The cost of doing the transcription in India was about one-fifth the cost per line of doing it the United States, a difference that got a lot of people’s attention.
By the late 1990s, though, Lady Luck was starting to shine on India from two directions: The fiber-optic bubble was starting to inflate, linking India with the United States, and the Y2K computer crisis—the so-called millennium bug—started gathering on the horizon. As you’ll remember, the Y2K bug was a result of the fact that when computers were built, they came with internal clocks. In order to save memory space, these clocks rendered dates with just six digits—two for the day, two for the month, and, you guessed it, two for the year. That meant they could go up to only 12/31/99. So when the calendar hit January 1, 2000, many older computers were poised to register that not as 01/01/2000 but as 01/01/00, and they would think it was 1900 all over again. It meant that a huge number of existing computers (newer ones were being made with better clocks) needed to have their internal clocks and related systems adjusted; otherwise, it was feared, they would shut down, creating a global crisis, given how many different management systems—from water to air traffic control—were computerized.
This computer remediation work was a huge, tedious job. Who in the world had enough software engineers to do it all? Answer: India, with all the techies from all those IITs and private technical colleges and computer schools.
And so with Y2K bearing down on us, America and India started dating, and that relationship became a huge flattener, because it demonstrated to so many different businesses that the combination of the PC, the Internet, and fiber-optic cable had created the possibility of a whole new form of collaboration and horizontal value creation: outsourcing. p. 109 Any service, call center, business support operation, or knowledge work that could be digitized could be sourced globally to the cheapest, smartest, or most efficient provider. Using fiber-optic cable-connected workstations, Indian techies could get under the hood of your company’s computers and do all the adjustments, even though they were located halfway around the world.
“[Y2K upgrading] was tedious work that was not going to give them an enormous competitive advantage,” said Vivek Paul, the Wipro executive whose company did some outsource
d Y2K drudge work. “So all these Western companies were incredibly challenged to find someone else who would do it and do it for as little money as possible. They said, ‘We just want to get past the damn year 2000!’ So they started to work with Indian [technology] companies who they might not have worked with otherwise.”
To use my parlance, they were ready to go on a blind date with India. They were ready to get “fixed up.” Added Jerry Rao, “Y2K means different things to different people. For Indian industry, it represented the biggest opportunity. India was considered as a place of backward people. Y2K suddenly required that every single computer in the world needed to be reviewed. And the sheer number of people needed to review line-by-line code existed in India. The Indian IT industry got its footprint across the globe because of Y2K. Y2K became our engine of growth, our engine of being known around the world. We never looked back after Y2K.”
By early 2000, the Y2K work started to wind down, but then a whole new driver of business emerged—e-commerce. The dot-com bubble had not yet burst, engineering talent was scarce, and demand from dot-coms was enormous. Said Paul, “People wanted what they felt were mission-critical applications, key to their very existence, to be done and they could go nowhere else. So they turned to the Indian companies, and as they turned to the Indian companies they found that they were getting delivery of complex systems, with great quality, sometimes better than what they were getting from others. That created an enormous respect for Indian IT providers[.] And if [Y2K work] was the acquaintanceship process, this was the falling-in-love process.”
Outsourcing from America to India, as a new form of collaboration, p. 110 exploded. By just stringing a fiber-optic line from a workstation in Bangalore to my company’s mainframe, I could have Indian IT firms like Wipro, Infosys, and Tata Consulting Services managing my e-commerce and mainframe applications.
“Once we’re in the mainframe business and once we’re in e-commerce—now we’re married,” said Paul. But again, India was lucky that it could exploit all that undersea fiber-optic cable. “I had an office very close to the Leela Palace hotel in Bangalore,” Paul added. “I was working with a factory located in the information technology park in Whitefield, a suburb of Bangalore, and I could not get a local telephone line between our office and the factory. Unless you paid a bribe, you could not get a line, and we wouldn’t pay. So my phone call to Whitefield would go from my office in Bangalore to Kentucky, where there was a GE mainframe computer we were working with, and then from Kentucky to Whitefield. We used our own fiber-optic lease line that ran across the ocean—but the one across town required a bribe.”
India didn’t benefit only from the dot-com boom; it benefited even more from the dot-com bust! That is the real irony. The boom laid the cable that connected India to the world, and the bust made the cost of using it virtually free and also vastly increased the number of American companies that would want to use that fiber-optic cable to outsource knowledge work to India.
Y2K led to this mad rush for Indian brainpower to get the programming work done. The Indian companies were good and cheap, but price wasn’t first on customers’ minds—getting the work done was, and India was the only place with the volume of workers to do it. Then the dot-com boom comes along right in the wake of Y2K, and India is one of the few places where you can find surplus English-speaking engineers, at any price, because all of those in America have been scooped up by e-commerce companies. Then the dot-com bubble bursts, the stock market tanks, and the pool of investment capital dries up. American IT companies that survived the boom and venture capital firms that still wanted to fund start-ups had much less cash to spend. Now they needed those Indian p. 111 engineers not just because there were a lot of them, but precisely because they were low-cost. So the relationship between India and the American business community intensified another notch.
One of the great mistakes made by many analysts in the early 2000s was conflating the dot-com boom with globalization, suggesting that both were just fads and hot air. When the dot-com bust came along, these same wrongheaded analysts assumed that globalization was over as well. Exactly the opposite was true. The dot-com bubble was only one aspect of globalization, and when it imploded, rather than imploding globalization, it actually turbocharged it.
Promod Haque, an Indian-American and one of the most prominent venture capitalists in Silicon Valley with his firm Norwest Venture Partners, was in the middle of this transition. “When the bust took place, a lot of these Indian engineers in the U.S. [on temporary work visas] got laid off, so they went back to India,” explained Haque. But as a result of the bust, the IT budgets of virtually every major U.S. firm got slashed. “Every IT manager was told to get the same amount of work or more done with less money. So guess what he does? He says, ‘You remember Vijay from India who used to work here during the boom and then went back home? Let me call him over in Bangalore and see if he will do the work for us for less money than what we would pay an engineer here in the U.S.’ ” And thanks to all that fiber cable laid during the boom, it was easy to find Vijay and put him to work.
The Y2K computer readjustment work was done largely by low-skilled Indian programmers right out of tech schools, said Haque, “but the guys on visas who were coming to America were not trade school guys. They were guys with advanced engineering degrees. So a lot of our companies saw that these guys were good at Java and C++ and architectural design work for computers, and then they got laid off and went back home, and the IT manager back here who is told, ‘I don’t care how you get the job done, just get it done for less money,’ calls Vijay.” Once America and India were dating, the burgeoning Indian IT companies in Bangalore started coming up with their own proposals. The Y2K work had allowed them to interact with some pretty large companies in the United States, and as a result they began to understand the pain points and how to do p. 112 business-process implementation and improvement. So the Indians, who were doing a lot of very specific custom code maintenance to higher-value-add companies, started to develop their own products and transform themselves from maintenance to product companies, offering a range of software services and consulting. This took Indian companies much deeper inside American ones, and business-process outsourcing—letting Indians run your back room—went to a whole new level. “I have an accounts payable department and I could move this whole thing to India under Wipro or Infosys and cut my costs in half,” said Haque. All across America, CEOs were saying, “ ‘Make it work for less,’ ” said Haque. “And the Indian companies were saying, ‘I have taken a look under your hood and I will provide you with a total solution for the lowest price.’ ” In other words, the Indian outsourcing companies said, “Do you remember how I fixed your tires and your pistons during Y2K? Well, I could actually give you a whole lube job if you like. And now that you know me and trust me, you know I can do it.” To their credit, the Indians were not just cheap, they were also hungry and ready to learn anything.
The scarcity of capital after the dot-com bust made venture capital firms see to it that the companies they were investing in were finding the most efficient, high-quality, low-price way to innovate. In the boom times, said Haque, it was not uncommon for a $50 million investment in a start-up to return $500 million once the company went public. After the bust, that same company’s public offering might bring in only $100 million. Therefore, venture firms wanted to risk only $20 million to get that company from start-up to IPO.
“For venture firms,” said Haque, “the big question became, How do I get my entrepreneurs and their new companies to a point where they were breaking even or profitable sooner, so they can stop being a draw on my capital and be sold so our firm can generate good liquidity and returns? The answer many firms came up with was: I better start outsourcing as many functions as I can from the beginning. I have to make money for my investors faster, so what can be outsourced must be outsourced.”
Henry Schacht, who, as noted, was heading Lucent during part of this period, saw t
he whole process from the side of corporate management. p. 113 The business economics, he told me, became “very ugly” for everyone. Everyone found prices flat to declining and markets stagnant, yet they were still spending huge amounts of money running the backroom operations of their companies, which they could no longer afford. “Cost pressures were enormous,” he recalled, “and the flat world was available, [so] economics were forcing people to do things they never thought they would do or could do . . . Globalization got supercharged”—for both knowledge work and manufacturing. Companies found that they could go to MIT and find four incredibly smart Chinese engineers who were ready to go back to China and work for them from there for the same amount that it would cost them to hire one engineer in America. Bell Labs had a research facility at Tsingdao that could connect to Lucent’s computers in America. “They would use our computers overnight,” said Schacht. “Not only was the incremental computing cost close to zero, but so too was the transmission cost, and the computer was idle [at night].”
For all these reasons I believe that Y2K should be a national holiday in India, a second Indian Independence Day, in addition to August 15. As Johns Hopkins foreign policy expert Michael Mandelbaum, who spent part of his youth in India, put it, “Y2K should be called Indian Interdepedence Day,” because it was India’s ability to collaborate with Western companies, thanks to the interdependence created by fiber-optic networks, that really vaulted it forward and gave more Indians than ever some real freedom of choice in how, for whom, and where they worked.
The World Is Flat Page 13