Friday, October 24, 2014


As the IT industry grew in the 1980s and 1990s, it had to compete for programmers with the US industry, which could pay more. So it developed strategies to deal with high turnover. This column was published in Business Standard of 8 April 2003.


The Indian software industry is seen as a global success story. However, it is little realized that throughout most of its rapid growth it had to cope with extreme shortage of programmers. It was at the receiving end in the market for programmers: their wages in the US were far higher, and they tended to take off at the earliest opportunity. Setting up an export industry with such a volatile labour force was one of the industry’s greatest achievements.
India became a source of programmers for an interesting set of reasons. There is a general belief that there were many Indians in the managements of US users of information technology; when the US started to run short of programmers in the late 1980s, these Indians said to their bosses, “Why not try India?” There was no doubt truth in this story. But it is a story told from the American end. There is another story at the Indian end.
In the 1960s, all the large computers in India were IBM 360s. These workhorses of the corporate world needed much programming; each computer was surrounded by an army of programmers, and IBM itself employed many more in India to support the users and solve their problems. Those that it employed were thrown out on the streets when George Fernandes forced IBM out of the country in 1978 for not diluting its stake in its Indian operation. The government set up Computer Maintenance Corporation (CMC) to employ them. The name is significant: the machines needed maintenance, and required a large, trained set of programmers for the purpose.
Very soon after IBM left, computers across the world began to be miniaturized; first there were computers the size of big suitcases, then it shrank until they reached the present dimension of a standard PC. On being deprived of IBMs, the government imported a few Russian computers. The private sector had to get the permission of the Department of Electronics to import a computer. It was generally obstructive and pro-Soviet; so hardly any big computers were imported by the private sector. Instead, it began to use rows of the new mini-computers. Import substitution being in full swing, any old industrialist could get a collaboration, set up an assembly line, and get the government to ban imports. So a large number of “manufacturers” of minis and PCs mushroomed. At that time, these small computers did not have standardized operating systems. So they created many programming problems; to solve them, their manufacturers employed programmers. Thus, programmers were scattered across these manufacturers, users and CMC.
News of the shortage of programmers in the US, especially for old computers like the 360s, reached India through engineers who had returned after studying there, and their friends. They started flying to the US to do brief jobs, and organized teams to export programmers. But at the other end, hundreds of NRI engineers started doing the same thing. They would fly to India, advertise, hold interviews, process the selected applicants’ visa applications and send them to work in US firms. This was the body-shopping business. It was largely in small fellows’ hands, and most of them were NRIs.
Indian employers of programmers suddenly started losing them, and had to do something about it. So a number of them, especially the private computer and PC manufacturers, got into the body-shopping business themselves. They offered to supply entire teams to American clients, and to take them back or replace them on clients’ demand. Thus, they had the advantage of scale. But their programmers, when they got to the States, often found jobs for themselves and left. So attrition was high; but at least the employers made some money out of them before they left.
They could be retained longer if they were not body-shopped – if they were kept in India and their output was exported. In theory it was possible through disks and tapes. But the Indian customs were so cussed about allowing their import that the business never took off.
There was another technology which could enable firms to make disembodied exports – a VSAT link. But it was under the control of another cussed government department – the Department of Telecommunications. Texas Instruments set up a unit in Bangalore in 1985. To get around the DoT’s dilatoriness, TI brought in a VSAT terminal and donated it to DoT for its own use. IBM wanted a link for itself in 1990. When it was not allowed, it threatened to leave. To prevent it from doing so, the Department of Electronics devised a scheme. Its basic form was that of an Export Promotion Zone. It would set up Software Technology Parks (STPs) with VSAT terminals in a building, and rent out rooms with connections to software exporters. Soon it got permission to give wireless connections to exporters who were not in the building. This trick, created to hoodwink the DoT, was hugely successful; the share of software exports handled by STPs grew from 8 per cent in 1992-93 to 68 per cent in 1998-99.
VSAT links did not entirely obviate the need to send programmers abroad; they still had to go and understand clients’ needs, and when software was ready, to go and install it on clients’ computers and make sure it worked. But for the rest of the time, work could be done in India, and programmers could not be enticed away. The Indian firms formed a practice of paying project managers and domain specialists particularly well, for they were the ones who were sent abroad most often.
That brought down the rate of attrition, but it was still high. Indian firms could not compete with US ones on wages, so the only way for them to continue in business was to replace programmers as they left. They could do so by poaching from other Indian firms; a lot of such poaching went on. But it could not solve the industry’s problem.
To replace leaving workers, the firms began to recruit engineers straight from their colleges as they graduated. The graduates might have seen a computer in college; they might even have learnt Word and Excel in the neighbourhood NIIT cybershop. But they were novices in programming. So the firms set up crash programmes to teach them programming.

The worst problem was of programmers leaving in the middle of a job. To deal with it, the firms developed accurate and detailed logs of work done, so detailed that another programmer could pick up the thread without too much disruption. They started getting their documentation systems certified by ISO and CMM. Later, these certifications became a guarantee of timeliness and quality, and helped the Indian firms o sell their services. In this way, worker shortages led them to unique solutions that gave them a brand and a reputation.