Should TCS Be Split Up? Big IT May Now Be Facing "Diseconomies Of Scale"

Is splitting up the way forward for the Indian IT industry?

Some 10 years ago, Wipro Chairman Azim Premji suggested in an interview that running a software services company with employee numbers exceeding 2,00,000 would be a challenge. A professor at Wharton had asked him if it would be tough scaling up people beyond 1,50,000, since there would be natural limits to growth in the services business.

Premji had replied:

I don’t see growing to 150,000 to 200,000 people as an insurmountable challenge….That is doable. After 200,000 people, you might have to think laterally about how to grow beyond that point. You have to think about partnership models — how do you manage partners and still have transparency with your customers?

Well, he is more or less there; as at the end of December, Wipro had a headcount of 1,70,664, while its bigger peers, had already reached the threshold. Tata Consultancy Services (TCS) was well past that point, with 3,44,691, Cognizant with 2,19,300 ad Infosys with 1,93,383.

Is unmanageable headcount one of the major reasons for the sector’s sluggish performance in recent years?

The fact is the sector has gone through several quarters of disappointing results, though not all the companies have done badly simultaneously. TCS seemed to be coasting along for five years while Infosys was struggling till about early 2015. Now, Infosys is doing better and TCS is lagging.

Last week, India’s two biggest IT services companies, reported their third quarter results (September-December 2015). While TCS reported a small drop in dollar income (-0.3 percent from previous quarter), Infosys reported a small rise (+0.6 percent). But Infosys shares went up after the results because the company announced that revenues for the whole year will be better than expected; its full-year revenue growth guidance was raised from 10-12 percent to 12.8-13.2 percent.

Both companies are still adding staff and growing business, but they are also huffing and puffing to retain those whom they recruit.

Both companies are thus likely to slow down recruitment, and automate more of their jobs. But since this can’t be done overnight, they are still adding staff to run existing projects and service new clients. In the quarter ended December, Infosys added 5,407 employees and TCS 9,071 employees.

TCS has been performing less well than Infosys over the last one year, and this could be the result of its larger size, where achieving higher growth on a larger revenue base is tougher. But equally important is the fact that the IT industry is fast reaching a point where managing huge employee numbers is becoming hugely expensive. Employees leave almost as fast as they are recruited.

In the last quarter, for example, TCS added 22,118 employees, but the net addition was only 9,071. This means 13,044 employees left TCS. Put simply, for every five employees TCS recruited, three left. This huge attrition is why TCS had to pay out a huge bonus of Rs 2,628 crore last April, but even this has not helped reduce the rate of employee exit from the company. Even though Infosys’ situation is a little better, despite bonuses and liberal promotions, it still loses a lot of employees every quarter.

The reason for this should be obvious. When your company has three-and-a-half lakh or two lakh employees, managing so many of them is a tough job. Just consider the sheer number of annual appraisals, increments, promotion expectations, and unhappiness with bosses that such large companies have to contend with. The net result is staff exit, especially because young people tend to jump to new jobs whenever the grass appears greener on the other side.

Also consider another problem – management talent. If we assume that you need at least one manager for every 50 employees (actually you need more, since it is tough to manage 50 technical professionals with just one supervising manager), it means a company like TCS will need a minimum of 7,000 managers at various levels.  The higher you go, the more competent the manager needs to be, but the Peter Principle says that each person rises to his level of incompetence.

Good managers are thus in great demand, and hence more likely to get new jobs with promotions. At a 15 percent attrition rate, TCS would lose 1,050 managers every year. Replacing them means spending time and effort to groom new managers.

This was exactly the question posed by K@W to Premji in 2006 – how to find so many of them – but he seemed to have an answer them.

Professor Ravi Aron, who interviewed Premji at that time, asked him:

The problem for Wipro going forward and ramping up in terms of revenues and people in India is not going to be the availability of programmers or technical people; it will be the availability of upper and middle management. If you look at business schools in India such as the Indian Institutes of Management (IIMs) or the Indian School of Business, companies like Wipro, Infosys or TCS are not attracting the top talent from those institutions. The employers of choice for those students seem to be the investment banks and consulting companies. This poses two kinds of challenges for you. First, how do you attract the top talent from campuses? And second, how do you get lots of really sharp people for your middle- and upper-middle management?

Premji’s answer was that he was widening the pool of institutions from where he recruited.

But it is unlikely the quality of the talent was the same. This problem would have been even more acute for the bigger employers like TCS, Infosys and Cognizant.

It is thus time to ask: is size now beginning to impact the big ‘uns of Indian IT? And if the answer is yes, and diseconomies of scale are setting in, isn’t it logical for them to start thinking small?

Maybe, TCS would be better off breaking up into several smaller companies, each with a special focus on specific markets. While in theory, this can be done internally with divisions acting like separate companies, but formal splitting can possibly make the smaller units more entrepreneurial than staying in one large joint family.

In a large company, an underperforming unit is less likely to be killed, especially when the company as a whole is surplussing cash, but as a standalone, it would be trying harder to succeed, and the plug would be pulled more quickly for it to fail.

Is splitting up the way forward for Indian IT? The jury is out, but tepid results tell us it is worth contemplating.