fewer extra costs. Relative to IT, the cost savings for IT-enabled services are much eas-
ier to benchmark and can be higher: IT-enabled services cost savings of 40% and even
up to 80% are reported.39 GE, one of the pioneers of outsourcing service operations to
India, reported annual savings of 340 million USD per year from its Indian operations.
Offshoring IT-enabled services is closely tied to offshoring IT. IT-enabled services and
IT are similar in that they both are driven by declining connectivity costs and both are part
of the new burst of international trade in services. But the ties are stronger still. Many of
the offshore IT providers are also IT-enabled service providers. Many of the most aggres-
sive corporations offshoring IT in the US and Western Europe are also aggressive in off-
shoring IT-enabled services. Finally, many of the countries offering IT services have also
grown IT-enabled services. Major Indian providers such as HCL and TCS have built IT-
enabled services operations. Major global IT providers have grown their IT-enabled serv-
ices: IBM bought Daksh e-services, IndiaÔÇ™s third largest back-of´¬üce services ´¬ürm, in 2004.
A large number of countries are active in IT-enabled services offshoring, and the
choice of locations may be even larger than with IT. Indian is the leader in this area:
according to NASSCOM, there were 210,000 people working in exporting IT-enabled
services in India in 2004, with revenues of 3.6 billion USD. Extraordinary growth fore-
casts call for IT-enabled service revenues in India to reach 64 billion USD by 2012.
Multinationals like Bank of America, British Airways, Swissair, American Express,
ABN Amro Bank, GE, and Citibank have all set up captive facilities in India. GE alone
has 12,000 people working in its many Indian IT-enabled service centers.
30 The fundamentals
The Philippines IT-enabled services exports, at 600 million USD, are now twice as
high as its IT exports. Filipino IT-enabled services centers are used by ChevronTexaco,
American Express, Procter & Gamble, and Accenture. Central European nations that
entered the EU in 2004 are preferred destinations for many continental Europeans. For
example, Poland has IT-enabled services centers for Fiat, Lufthansa, and Philips. Even
African nations tend to be more active in IT-enabled services. These include: South
Africa, Ghana, and the Indian Ocean nation of Mauritius. Interestingly, security cam-
eras in some US car-parks are monitored remotely from the small West African nation
of Cape Verde.40
In some cases the risks are higher when offshoring IT-enabled services relative to
offshoring IT. When offshoring application development work some projects get
delayed because of unclear speci´¬ücations or miscommunication. This can be irritating
or even problematic, but it is usually not critical to the company: business carries on.
This is different with IT-enabled services. When outsourcing revenue-generating
processes offshore, such as contact centers, there is risk to an entire business process.
Revenue can be lost if customers are frustrated with the offshore service provider. A
widely published example is Dell, which decided in 2003 to redirect some customer
service calls to helpdesks in the US, rather than to its call center in Bangalore, India.
Dell brought some activities back to the US when it found that several of its business
customers complained that Indian technical support workers relied too heavily on
scripted answers and could not handle complex computer problems.
IT-enabled services has already become ÔÇťthe next big thingÔÇŁ in offshoring. The
American consultancy Forrester predicted that by 2015, 3.3 million American service
jobs will move offshore, most of these in IT-enabled services.41 We return to the topic
of IT-enabled services when we examine the efforts of nations to build their offshore
IT industry in Chapter 10.
Now that we have introduced the rich landscape of offshoring, let us take you back to
an observation we made earlier in this chapter. Recall that we introduced two evolving
con´¬ügurations of global software activities. The ´¬ürst is that of a global network in
which nodes of software producers are interconnected. The second is a supply chain
structure where each software producer adds value to the software and passes it on the
next phase in the production process.
These offshore con´¬ügurations bring us to the title of this book. As offshoring diffuses,
the software industry will increasingly resemble other industries in which components
and services are sourced globally from high-, medium-, or low-wage nations. Over time
there will be less that is unique in ÔÇťoffshoring.ÔÇŁ Instead, what we label today as ÔÇťoff-
shoringÔÇŁ will become simply ÔÇťglobal sourcing.ÔÇŁ Perhaps our next book may have this title.
2 Offshore economics and offshore risks
Three providers bid on the state of New Mexico unemployment system. Two
American ´¬ürms, IBM and TRW bid 12 million and 18 million USD, respectively.
The winner, Tata Consultancy Services based in India, came in at less than
6 million USD.1
Solidcare Systems, a Security Software Startup in Silicon Valley, has no research
and development (R&D) in Silicon Valley. Instead, to stretch its 5.3 million USD
in venture capital funding, all of its software engineers are in India earning
much lower wages.2
Philips, the Dutch electronics giant, faces strong competition from Asian
producers. It claims to be saving millions of dollars by moving major parts of
its technical and embedded software development to cheaper locations. Its
Indian software center now houses 900 employees and is responsible for 20%
of PhilipsÔÇ™ worldwide software content.3
There are hundreds of such ´¬ürms, large and small, all of which trumpet their offshore
costs savings. Companies in the wealthy industrialized nations are dazzled by offshore
programmersÔÇ™ low wages. Since cost savings are the dominant reason for offshoring,
much of this chapter is devoted to dissecting the controversy regarding the actual costs
and savings. Then, later in the chapter, we turn to the myriad of risks which companies
face when offshoring.
Labor arbitrage: ´¬ünding the lowest wages
A software development manager is shopping for the lowest-cost labor suppliers and is
drawn to the low-wage offshore nations. Thus, he is acting somewhat like an arbitrage
´¬ünancier: he is sourcing labor where it is the cheapest to use it, where it can earn the
greatest return. Since we are dealing with software, in which there are few expensive
assets like factories, the costs of development are driven mostly by the wages of software
labor, from the junior programmers to the seasoned project managers. For example,
variations on Table 2.1, showing comparative wage levels of software professionals,
32 The fundamentals
Table 2.1 Wages for software professionals. Annual, in USD
UK 45,000ÔÇ“99,000(e); 81,000(a)
Ireland 23,000ÔÇ“34,000(f); 35,000; 23,000(h)
South Africa 18,000(f); 30,000(a)
Philippines 5000(h); 6600(d); 10,000(b)
India 5000(h); 5900(d); 9000(a)
Russia 5000(h); 5500ÔÇ“7500(f); 7500(d); 7000ÔÇ“18,000
Poland 4800ÔÇ“8000(f); 9000
China 3000(h); 4700(d); 7000ÔÇ“14,200
Vietnam 1400ÔÇ“6000; 3000(h)
have appeared frequently in the IT business press.4 The wage differentials are striking
when viewed by managers being pressed to reduce costs.
The software engineers in these offshore destinations are not being paid undigni´¬üed
rates. To the contrary, the Indian programmerÔÇ™s basket of goods and services that she can
purchase with an annual income of 9000 USD is roughly equivalent to her European
counterpart. The cost of middle-class housing is much less in Bangalore than in Silicon
Valley. These equivalences represent the notion of purchasing power parity (PPP), a tra-
ditional economic index, re´¬‚ecting different national costs after currency conversion.
Wage data of Table 2.1 need to be treated carefully since they vary considerably ÔÇ“
by as much as 50%. Wages are substantially higher in the principal cities: higher in
Shanghai than in Suzhou, higher in Moscow than in Novosibirsk, higher in Bangalore
than in Ahmadabad, higher in London than in Wales, and higher in Silicon Valley than
in Omaha. Wages also vary considerably depending on the size of the ´¬ürm, with engi-
neers at large ´¬ürms typically earning 25% or more than engineers at small ´¬ürms.
A number of sources try to be consistent across countries by focusing on wage data for
junior programmers, typically fresh out of university, since this is the most objective com-
parison. However, the salary differences (compared to programmers in the US and
Europe) tend to be most pronounced for junior programmers, and the salary gap gets com-
pressed for more experienced engineers and managers. For example, wage differential
between Europe and India is roughly 1:8 (Europe:India) at the junior programmer level
33 Offshore economics and offshore risks
and only roughly 1:3 (Europe:India) at the high managerial levels.6 Thus, many wage
comparisons in the press bias the wage differentials of offshore labor.
Wages are only part of the story. Companies that are managing their own offshore
centers need to be looking at the ÔÇťloadedÔÇŁ wages that professionals receive. Direct
wages re´¬‚ect only part of the labor costs. Bene´¬üts are thick and elaborate in many
nations. In India bene´¬üts include mini-bus pick-up and drop-off every work day. In
Russia, where consumer ´¬ünance is undeveloped, ´¬ürms give loans to buy apartments. In
Israel they include a company car in more than one-third of the cases, as well as some
lunch discounts in about two-thirds of the cases.7 In China bene´¬üts include an extra
13th month payment for the lunar holiday, a housing allowance at 17% of wages, med-
ical insurance at 12% of wages, and more. A leaked IBM internal memo shows that
a Chinese programmerÔÇ™s cost, including salary and bene´¬üts, is 12.50 USD an hour,
still only about 20% of an equivalent American IBM programmer (the comparison was
for a programmer with 3ÔÇ“5 years experience).8
Once other costs are factored, then the fully burdened Indian employee costs about
30,000ÔÇ“40,000 USD per annum. This includes the cost of rent, which is relatively high
in the major Indian destinations (1 USD per square foot in good locations), support,
services, insurance, taxes, and other items. For comparison, loaded cost (including
bene´¬üts and taxes) for an American software engineer, at a large company, is about
120,000 USD and a burdened cost is about 150,000 USD per annum.
We have heard from more than one na├»ve Western manager that Indian wages
will soon be bid up to reach those in the US and Europe. This is misinformed. With its
large pipeline of fresh engineering graduates and its enormous underclass serving as a
drag on wages, Indian software engineering wages will not approach US or British wages
in our lifetime unless there is a catastrophic economic depression in the West that does
not touch Asia (an unlikely delinking). That said, wages in India and China have been bid
up for quite a few years, rising 10ÔÇ“30% per year. One result of the strong demand is that
India is no longer the lowest cost software nation. Instead, some ´¬ürms (including some
Indian ´¬ürms) are turning to China, Vietnam, and others, where wages are lower.
Clients that are not directly issuing payroll checks to their offshore employees are
paying a charge rate to the offshore provider. The charge rate is much higher than the
wage rate and tends to hover around 15ÔÇ“35 USD per hour for India and Russia. This
rate is higher than the wage rate because it includes the providerÔÇ™s overhead, market-
ing, training, and pro´¬üt (these pro´¬üts can be high: the Indian ´¬ürm Infosys is the most
pro´¬ütable major software services company in the world). The charge rate for larger
Indian ´¬ürms is at least twice the wage rates and usually several times than that. How-
ever, these rates tend to ´¬‚uctuate quite a bit, based on general supply-demand condi-
tions and based on the customerÔÇ™s negotiating power. Charge rates below 10 USD per
hour have been unusual for the larger (Tier-1) Indian ´¬ürms, but are more common for
the small ´¬ürms. American providers using offshore resources do their best to compete
with the Indian providers, but generally offer somewhat higher charge rates than their
Indian competitors. Nevertheless, we have heard of very low charge rates offered by
34 The fundamentals
the American providers using offshore resources: for example, an HP bid for market
share in 2004 at a charge rate of 15 USD per hour; a large US ´¬ürm bid for 10 USD per
hour for maintenance work.
Then we come to the onshore charge rate. At least some of the labor needs to be per-
formed onshore to be close to the client. This labor is charged at a different, higher rate,
ranging from 35 to 80 USD per hour. At the upper range this onshore rate is close to
a fully loaded rate for a US-based engineer. The onshore professionals need to be
charged at higher levels, since they incur higher costs and sometimes may need to be
paid the prevailing rate for legal reasons.
For some years the offshore providers were emphasizing the ÔÇť80:20ÔÇŁ ratio in their
marketing pitches: only 20% of the staff need to be onshore charging high rates, while
80% of the staff are offshore charging the more attractive lower rates. In practice, this
ratio varies considerably by phase, project, and client. We know of many clients that
maintain the ratio at about 50:50. Why would clients choose a ´¬ünancially disadvanta-
geous ratio? Some clients are sluggish in transitioning work offshore (as discussed in
change management in Chapter 7) since reaching higher ratios offshore requires proac-
tive policies. Other clients simply viewed their offshore (primarily Indian) providers as
suppliers of labor, only at lower costs. In its extreme form this is body-shopping, the
label that is shunned by all involved in the offshoring business because it is suggestive
Labor arbitrage is most dramatic at the bottom end of the global programming land-
scape. Companies from countries such as Bangladesh, Nepal, or Vietnam are asking
very low rates, sometimes below 5 USD per hour. Another source of very cheap rates
is the assortment of online programming marketplaces that pit programmers against
each other as they compete for small projects. These online marketplaces9 work using
a reverse auction mechanism in which the lowest bidder is more likely to be selected
by the client. In a study of these marketplaces conducted by Carmel and Espinosa, a
key ´¬ünding was that in head-to-head competition between providers from industrial-
ized nations (e.g. the US) and offshore nations (e.g. India), the median winning project
bid amount was only 35 USD for the whole project.10 This is a strong indication that
offshoring is driving down prices, since programmers from high-income nations can-
not survive on proceeds from such tiny projects.
Transactions Costs and Total Savings from Offshore Strategy
In spite of the low offshore wages, total offshore costs can actually be higher than
before offshoring. Why? Because of the economic concept of Transactions Costs.11
This concept helps to explain the economics of offshoring. Classical economics was
based on the theory of ef´¬ücient markets and thus it should always be cheaper to con-
tract out than to hire. Yet, by the 1900s it became evident that this theory was de´¬ücient,
35 Offshore economics and offshore risks
since it could not explain why giant corporations appeared to be quite successful at cre-
ating vertically integrated companies that seemed to do nearly everything within the
boundaries of the ´¬ürm.
Thus, the Theory of Transactions Costs encompasses a powerful idea: that there are
real costs incurred when going out to the marketplace for identifying suppliers, nego-
tiating and contracting with these suppliers, and then, later, policing these suppliers so
that they produce to the desired quality level.
Now we apply this to offshoring. Companies that are acting in their own self-interest
look to the marketplace if their production cost savings of outsourced offshore work
outweigh the additional Transactions Costs incurred when dealing outside the ´¬ürm. In
the case of offshoring, most of the production cost savings are due to the lower wage
costs for the software staff in low-wage nations. Thus, to put it numerically, it is more
ef´¬ücient to offshore when:
Production cost savings Sum of all Transactions Costs
Of course, the dif´¬üculty when offshoring is that the Transactions Costs are dif´¬ücult to
assess without experience or benchmarking. Furthermore, offshore Transactions Costs
tend to be higher when the tasks are dif´¬ücult to de´¬üne, when uncertainty is high, and
when complexity is high.
In order to capture the notion of Transactions Costs, the outsourcing and offshoring
industry coined two useful terms, namely TCO, the Total Cost of Offshoring; and TCE,
the Total Cost of Engagement. Both of these terms attempt to capture all the costs
of offshore activity in order to compare them with ÔÇťonshoreÔÇŁ or ÔÇťnormalÔÇŁ costs. The
problem with TCO and TCE is that many ´¬ürms end up offshoring using the ÔÇťStumble-
and-then-SucceedÔÇŁ form of offshoring. Here ´¬ürms go offshore and encounter problems
and failures. They stick with their offshore decision and eventually make it work.
Such was the case of Silicon-Valley-based ValiCert which later became part of
information security ´¬ürm Tumbleweed.12 Following the contraction of the technology
crash of 2000, ValiCert struggled as a company. In 2001 it hired Indian-based Infosys,
laying-off programmers in the US in this process. After numerous dif´¬üculties includ-
ing its own lack of experience, the company created its own Indian subsidiary reaching
60 employees. During this phase ValiCert continued to experience problems and
strains between the Indian and US of´¬üces. Only in 2003, more than 18 months after
offshoring began, did this small global company begin to properly perform in a dis-
tributed structure. At that point the company attributed its very survival to the Indian
subsidiaryÔÇ™s cost savings. It had computed that the burdened cost of its Indian employ-
ees was roughly one-sixth of the cost of its Silicon Valley-based employees. Thus,
ValiCert practiced ÔÇťStumble-and-then-Succeed.ÔÇŁ
In sum, given the extra Transactions Costs and the many cases of Stumble-and-then-
Succeed, the correct measure for offshoring should be Total Savings of Offshore
36 The fundamentals
Total direct cost savi
Cumulative savings ($)
Financial flows ($/t)
Cu ht ax
Figure 2.1 Offshore bene´¬üts, costs, and savings over time.
By using the notion of TSOS, a portfolio of projects is examined over a certain time
period. Some of these, on an individual basis, will show negative savings, as would any
project portfolio. For new users, the ´¬ürst projects are always hard to manage and cost
savings will only come later. Successful ´¬ürms will show a positive TSOS over time.
Put differently, this is the ÔÇťlearning curveÔÇŁ that companies need to climb up. The word
strategy is used in TSOS because there must be a multi-period offshoring strategy in
order to achieve economic bene´¬üts.
Figure 2.1 depicts a typical cost and bene´¬üt stream over time, with cumulative sav-
ings rising up in a classic inverted-J-shape, also known as a ÔÇťhockey-stick.ÔÇŁ The ´¬ünan-
cial milestone is the breakeven point where the cumulative savings begin to be
positive. This is the TSOS. Of course, the ´¬ürmÔÇ™s time period to achieve a positive
TSOS will vary. Small, well-speci´¬üed projects (out-tasking) can reach positive TSOS
quickly, sometimes immediately. More signi´¬ücant engagements are unlikely to reach
positive TSOS in less than 1 year ÔÇ“ and may often be 2 years and more. And, of course,
in some cases a positive TSOS is never achieved.
Extra offshore costs
Transactions Costs is the more formal term for the grouping of costs which we call
extra offshore costs, and which has also been labeled hidden costs.13 While labor arbi-
trage makes offshoring attractive, these extra cost components tend to muddy the pic-
ture. The extra cost items are listed in Exhibit 2.1. Each of these cost items is discussed
in detail in this section.
37 Offshore economics and offshore risks
Search and Contract
Infrastructure: technology and connectivity
Mitigating risks (e.g. disaster recovery)
Exhibit 2.1 Extra offshore cost items.
The ´¬ürst category, Search and Contract, is usually a one-time cost for the ´¬ürst
engagement. Search and Contract costs for outsourcing revolve around provider selection.
These costs include initial research, consulting fees, legal, contracting, and travel. For
some companies, this item is small because their search is opportunistic; for example,
they contract with a provider they met coincidentally. The duration of this phase is usu-
ally months, and sometimes approaches a year. The overseas trip by one or more exec-
utives has become part of the ritual of offshoring and is a part of the search costs.
Another initial cost is restructuring. This is the euphemism used to signify layoffs,
severance pay and any retention costs: paying the essential people extra so they will
not leave. Retention expenses are critical because ´¬ürms need to keep the key employ-
ees to share their knowledge, facilitating the knowledge transfer (KT) described fur-
ther below. This cost item is more common on the American side of the ocean than in
Europe, where offshore projects start small and layoffs are less common.
The rest of the cost items are mostly ongoing costs, although they change over time.