Figure 1.2 Scenarios of offshore migration.
development activities (low-level) that are migrating offshore at different rates. Offshoring
will reach some plateau (though we cannot say when). The plateau may be at a lower rel-
ative level (as in ÔÇťDevelopment AÔÇŁ) or at a higher relative level (as in ÔÇťDevelopment BÔÇŁ).
While we cannot predict the future, we can draw six lessons from previous offshoring
waves that we detail below.
1 To reiterate, in previous offshoring waves, production shifted to lower-wage nations.
Some of these waves were gradual increases over time, some had an in´¬‚ection point
where offshoring accelerated, others moved through some kind of ÔÇťS curveÔÇŁ
reaching a new plateau (e.g. automobiles), as in the two upper lines of Figure 1.2. In
consumer electronics, production ÔÇ“ and much of the design ÔÇ“ moved ´¬ürst to Japan
and then shifted to the newly industrialized countries (NICs: South Korea, Taiwan,
Singapore, and Hong Kong), and then to other Asian countries (China, Thailand,
and Malaysia) with even lower wages. This migration pattern is referred to by
international economists as the Flying Geese Formation, where the lead goose is
the US, followed by Japan, and so on. These ´¬‚ying geese are beginning to appear
in the software industry, as well. After the US lead, the ´¬ürst geese were the three
Is: India, Israel, and Ireland. The second tier of geese appeared when Indian ´¬ürms
began to move some software work to another tier of low-wage nations: Vietnam
2 In previous offshoring waves, it was not just production that moved offshore but
the know-how about that production, as was the case in the steel industry or,
separately, the semiconductor-manufacturing industry. In software offshoring,
production know-how transfer is most evident in the quality standards such as
Capability Maturity Model (CMM) and International Standards Organization
(ISO) (which are introduced later in this chapter). The Indian organizations, and
later software ´¬ürms in other nations, embraced these standards and are now global
leaders in their application.
7 The offshore landscape
3 Offshoring is a signi´¬ücant industry tremor leading to massive restructuring, namely:
acquisitions, consolidation, job displacement, and the emergence of global giants
with a broad presence in major markets. Since 2000 this tremor and its after-shocks
has been evident in the IT services industry where the distinctions between the
major IT services organizations in the US, Europe, and India have begun to blur.
American ´¬ürms in this segment increasingly resemble Indian ´¬ürms in their
offshore offerings, while the large Indian ´¬ürms are vying for the largest contracts
just like the American ´¬ürms.
4 In previous offshoring waves there was often a corresponding rise in the industryÔÇ™s
productivity in the home countries, due to a rise in R&D investments, automation,
and production ef´¬üciencies. In parallel there were signi´¬ücant changes in the design
approaches used in each industry. Charles Simonyi,4 one of MicrosoftÔÇ™s ´¬ürst
software architects, argues that offshoring is but a prelude to software automation
and mechanization. There is evidence that this is already taking place as software
service companies scramble to automate labor-intensive tasks in data centers,
software customization, translation, web site hosting, and reuse of code.
5 In some industrial offshoring waves there was a split between higher-level design
activities and lower-level production activities, as in the distinction between design
and development of Figure 1.2. Indeed, in the case of software, one of the forces of
offshoring is standardization, allowing some factory-like approaches in software
production. This is a departure from the practice of many decades in which
software was practiced largely as a craft. Standardization is less evident in higher-
level (design) tasks, which are more creative tasks, and which are usually the
sources of a companyÔÇ™s competitive advantage.
6 The political dynamics surrounding previous offshoring waves suggests that
protectionist policies, such as import barriers, can help to slow offshore migration
for some periods, but do not seem to be effective in the long term. This is an
interesting lesson for industrialized nations struggling to deal with the
rami´¬ücations of offshoring (see Chapter 12, Offshore Politics). The political
dimensions have also changed in this offshoring wave. In the case of the US, the
political constituencies of business and labor have diverged. Large corporations
were vocal when the competitive threats came from Japan in the 1980s. However,
in the software offshoring debate of post-2000, US ´¬ürms continue to dominate the
global marketplace. Not only do they not lobby for protection, to the contrary, they
lobby against protectionism. The other political constituency is software labor,
which in the US is largely non-unionized.
We offer a ´¬ünal observation in our look into the future. Offshoring will likely acceler-
ate the formation of two industry con´¬ügurations: networks and supply chains (see Figure
1.3). On the one hand, offshoring has created truly global networks of software activities,
similar to the well-known network structure of the Internet. A network is set of con-
nected nodes with each node connecting to many other nodes. It is not unusual anymore
8 The fundamentals
(a) Network of software activities (b) Supply chain of software activities depicting
embedded software that is used in a typical GM car.
Figure 1.3 Future structures of global software activities: network and supply chain.
to ´¬ünd a network of collaborating teams, as in the case of an EDS project that had such
network collaboration between Mexico, Australia, Egypt, and Brazil.
On the other hand, borrowing from another business area, the software industry is
beginning to resemble the auto industry in that there is a ÔÇťglobal supply chainÔÇŁ of software
producers, where each producer adds value as the software is transformed and then passed
from one phase to the next. We see this illustrated in the auto industry itself, in the embed-
ded software that goes into todayÔÇ™s cars. A typical GM vehicle has about 65 specially
built microprocessors, each with its own embedded software (in fact, together, these 65
microprocessors are now more expensive than the costs of all of the other raw materi-
als that go into the car). GM writes little of this software in-house, with the exception
of the microprocessor for the power train. Instead, it contracts with three major suppliers,
Siemens, IBM, and Motorola, who in turn, source from a network of American, European,
and Asian software centers. In short, a global supply chain of software that goes into
The Offshore Stage Model: progression and diffusion
We now turn to look at companies that are offshoring in order to understand the pro-
gression and diffusion of this phenomenon. The Offshore Stage Model, ´¬ürst described
in an article by Carmel and Agarwal,5 helps us to tell this story.
Companies tend to move through four offshoring stages depicted in Figure 1.4.
Companies that do not offshore are in Stage 1, ÔÇťOffshore Bystander,ÔÇŁ in which they
metaphorically watch the others. In fact, as we later discuss, most companies, whether
large or small, are still in Stage 1.
Stage 2, ÔÇťExperimental,ÔÇŁ is a transition stage in which companies test the offshoring
waters for a year or more. For large corporations this stageÔÇ™s expenditures could be
as large as 10ÔÇ“20 million USD per year. Experimentation is a wise approach for organi-
zational learning and risk reduction because of the many dif´¬üculties in offshoring. Savvy
9 The offshore landscape
Strategy focused on
Domestic sourcing only
Figure 1.4 Offshore Stage Model.
managers experiment to the point where they see measurable, positive results, and only
then do they grow to the next stage. Some call this the ÔÇťStart-SmallÔÇŁ strategy and,
according to one study, 63% of companies are using this approach.6
In Stage 3, ÔÇťCost Strategy,ÔÇŁ companies begin to experience signi´¬ücant and consistent
cost savings in their IT work. By this stage, ´¬ürms have corrected some early missteps and
have expanded their offshore activity as measured by number of projects, staff, or budget.
There have been hundreds of ´¬ürms, if not thousands, large and small, which claim cost
savings in their software related activities driven by the low wages in offshore nations.
Various studies have tried to determine just how much offshoring saves. The composite of
studies indicate that the cost savings ranges from 15% to 40%7 for companies offshoring
at least a year (this is discussed in Chapter 2, Offshore Economics and Offshore Risks).
Experienced companies move to Stage 4, the highest stage, where they truly leverage
offshoring. In this stage companies move beyond mere cost savings derived from wage
differentials and bene´¬üt from other strategic advantages. Here, offshoring is used to drive
innovation, speed, ´¬‚exibility, and new revenues.
The Offshore Stage Model is also useful to measure offshoring diffusion. Since it
was introduced in 2002, it has been used to estimate the ratio of large companies at
each stage of the offshore progression, as shown in Table 1.1. The rough estimates in this
table, made by two American research companies, indicate that only 10% of the largest
US corporations were active in offshoring in 2003ÔÇ“2004 (i.e. they were in either Stage
3 or 4). Furthermore, about half of the largest American ´¬ürms do not offshore at all.
In spite of the enormous attention to offshoring in the US in the early 2000s, offshoring
was still rather limited.
10 The fundamentals
Table 1.1 Offshore stages of US Fortune 1000 ´¬ürms
Percent of 1000 largest US ´¬ürms
Percent of all software work
in this stage (2003ÔÇ“2004)
which is offshored for a typical
Stages Meta Group (%) Forrester (%) ´¬ürm in this stage (%)
Stage 1 55 50ÔÇ“60 0
Stage 2 33 25ÔÇ“30 5
Stage 3 8 5ÔÇ“10 10ÔÇ“30
Stage 4 4 5 40ÔÇ“50
Source: Estimates by Meta Group8 and Forrester9
The stages can also be used to anticipate offshore diffusion for large ´¬ürms. If, assuming
conservatively, only 20 ´¬ürms a year move out of Stage 1 and into Stage 2, and the annual
advance from stage to stage is just 10% of the ´¬ürms in that category, then by 2010, nearly
one-third of US ÔÇťFortune 1000ÔÇŁ ´¬ürms will be active offshore users in Stages 3 and 4.
IT offshoring has been driven primarily by executivesÔÇ™ desire to lower operational cost.
This is the Cost Strategy of the Offshore Stage Model. Lowering operational costs does
not necessarily translate into a companyÔÇ™s strategic advantage, just as saving money on
a new of´¬üce lease is not a strategic advantage, but merely the relentless day-to-day
effort of any company to reduce its operating costs.
However, in some industries, IT offshoring is beginning to be viewed as a strategic
necessity. Some call it ÔÇťoffshore or die.ÔÇŁ When one companyÔÇ™s cost ef´¬üciencies allow
it to lower prices or expand its competitive options, then other companies must match
their competitorÔÇ™s strategy, or fail. Offshoring is becoming part of the larger context of
hyper-competition: companies are swept into faster and faster cycles of competitive
responses and reactions in order to remain ´¬ünancially viable and cost competitive. Not
offshoring may well become a strategic peril. Such was the case of one of AmericaÔÇ™s
largest television manufacturers, Zenith Electronics, which resisted offshoring for
decades, while slowly shrinking, before it disappeared completely.
While cost reduction is the primary strategic focus of most companies that are off-
shoring, it is not the only strategic advantage to offshoring. The fourth and ´¬ünal stage
in the Offshore Stage Model is labeled ÔÇťLeveraging Offshore.ÔÇŁ As we saw in the estimates
of Table 1.1, there are relatively few companies that have reached this stage. Those that
have progressed to this stage have moved beyond mere cost reduction and bene´¬üt from
innovation, speed, ´¬‚exibility, and new revenues. We discuss these bene´¬üts in greater
detail in Chapter 5, Offshore Strategy. Here we introduce the two most important of
these additional strategic goals: attaining speed and accessing talented labor.
11 The offshore landscape
The ´¬ürst strategic lever is the increase in speed, agility, and ´¬‚exibility. This means
that companies that offshore can rapidly ramp-up (by reducing the time to get the
project started) and reduce project duration (time-to-completion). The abundant sup-
ply of labor offshore gives companies greater agility: to assign a large number of engi-
neers to a problem; to forge ahead in several directions instead of just one; to ramp-up
(scale-up) and respond to a business need within days instead of months.
Companies that develop software products bene´¬üt from the second strategic advan-
tage: accessing talent. For these companies, their success stems from innovation and their
innovation capabilities come from their talent ÔÇ“ their most brilliant and creative engineers.
Firms that expand abroad to tap this talent are called ÔÇťknowledge seekersÔÇŁ 10 and tend
to behave somewhat differently than those seeking mainly lower wage rates. In previ-
ous decades technology companies would tap foreign talent by going to other high-
wage, industrialized nations. In the 1990s, they began turning to Israel, India, and later
to China. For example, by 2003, 77 global software product ´¬ürms established direct
R&D subsidiaries in India.11 Many others perform contract R&D on an out-tasking
basis in India.
Stories about offshoring often mention follow-the-sun, also known as round-the-clock.
Along with low costs, follow-the-sun is another allure of offshoring. It is often men-
tioned by those who seek to make offshoring sound unique. Follow-the-sun, as the
name hints at, exploits time zone differences to speed up project work. For example, a
team in America can hand off its work at the end of its day to team members in India
or China, who can then continue the work while the US team members sleep.
This has undeniable appeal. If software work can be coordinated properly, then project
duration can be reduced by a factor of two. Moreover, if three teams are correctly posi-
tioned across time zones, then a theoretical threefold duration reduction is possible.
This is much like a factory running three shifts, 24 hours per day, producing three times
the volume. Using follow-the-sun development, a company may be able to save months
from the development cycle and release a product earlier, thus giving it a competitive
advantage. This is an enormous potential bene´¬üt of offshoring.
However, coordination in follow-the-sun must be ´¬‚awless in order to reduce project
duration. One miscommunication can delay the entire dayÔÇ™s worth of work. In practice,
few globally dispersed software efforts have been able to fully capitalize on the theo-
retical advantages of follow-the-sun. Daily follow-the-sun coordination is simply too
dif´¬ücult for software teams. An IBM team, described in CarmelÔÇ™s 1999 book,12 was set
up to capitalize on follow-the-sun. However, fairly quickly, the global team discovered
that daily handoffs were too dif´¬ücult to coordinate.
Nevertheless, follow-the-sun can be effective for some activities and for certain phases
in software work. Startups in Silicon Valley have been excited about rapid prototyping
12 The fundamentals
of new software products in which the coding is done in India, and then sent back to
the US for comments and re´¬ünement. Activities, such as bug-´¬üxing (in the maintenance
phase), or call-centers (e.g. technical support), are better suited to follow-the-sun,
because they are usually small tasks (low granularity), of low complexity, and can be
routinized between the time-separated sites.
Recall that roughly half of AmericaÔÇ™s largest corporations are not offshoring at all. The
percentage is higher in Europe, and higher still once small- and medium-sized ´¬ürms are
Why is this? Why is it that relatively so few American, European, and Japanese ´¬ürms
are offshoring? All this is in a business environment in which offshoring is one of the
accepted, if not expected, strategies. For example, the US strategic consulting ´¬ürm BCG
issued a report in 2004 that practically shouted, ÔÇťthe real question now is not whether
to go global, but how much and how fast you can move.ÔÇŁ
There are many reasons for the relatively small participation in offshoring, but we
begin with a simple one: it is more dif´¬ücult to work with people far away than those close
by. It is more dif´¬ücult, because of ´¬üve factors introduced here (and covered in more detail
in Chapter 8, Overcoming Distance and Time):
1 Communication breakdown. We human beings communicate best when we are
close. Yet, offshoring is all about working with people far away, with whom
communication is conducted via ÔÇťnarrowÔÇŁ channels such as e-mail or telephone.
A software engineer would always want to conduct a dif´¬ücult design session face-
to-face. Why? Because people communicate with more than mere text or words.
The way the words are delivered (via tone of voice, the pauses in speech, the body
language, the gesturing at the whiteboard) are all vital. Some say that 80% of
the messages we convey go beyond the plain text. The all-too-frequent result of
communication over distance is that dreaded word: miscommunication.
2 Coordination breakdown. Software is a complex task that requires many small and
large adjustments. People who work on a common task coordinate via countless
adjustments: a question, a request for clari´¬ücation, a small improvement, an
ad hoc solution resulting from a 1-minute chat while standing in line at the cafe.
So much coordination comes from spontaneous, face-to-face conversation. When
offshoring, all of these small adjustments do not take place, certainly not easily.
When coordination slows or breaks down, several dynamics occur. Problem
solving gets delayed again and again, or the project goes down the wrong track
until it becomes very expensive to ´¬üx.
3 Control breakdown. Successful management control takes place when managers
can roam around to see, observe, and dialogue with their staff. Hence,
management by walking around (MBWA). When a team leader or project manager
13 The offshore landscape
is supervising software developers many kilometers away, roaming around and
getting a ÔÇťfeelÔÇŁ for whatÔÇ™s happening becomes an unusual event. Sometimes it
never happens at all. And, when managers cannot roam, they have to rely on
collecting information and imposing their will by means of technology: telephones
and e-mail. This is less effective than face-to-face.
4 Cohesion barriers. Groups that are close together jell and bond. People get to like
each other, trust each other, help each other, and work harder for each other.
Offshoring introduces a situation in which the group of dispersed individuals is
unlikely to form these tight social bonds.
5 Culture clash. Offshoring means going to far-away lands and working with foreign
cultures. Each culture has different principles, values, beliefs, communication
norms, and behaviors that are embedded deep in our minds. In fact, we now
understand that our respective cultures are ÔÇťprogrammedÔÇŁ into our minds by age 10.
The result of all these deeply engrained differences is that in any cross-cultural
communication, the receiver is more likely to misinterpret messages or cues.
Hence, the familiar complaint of miscommunication across cultures. We devote
Chapter 9 to the problems and solutions of cross-cultural communications.
These ´¬üve factors represent only some of the dif´¬üculties that make offshoring dif´¬ü-
cult. These and other offshore challenges translate into the extra costs of offshoring.
These extra costs can sometimes offset wage advantages making offshoring a losing
proposition. The extra costs of offshoring are covered in Chapter 2.
What is done offshore?
ÔÇťÔÇ¦ not a single activity is immune ÔÇ¦ [ to offshore].ÔÇŁ
Findings from a 2003 survey
by the American industry magazine
ÔÇťOffshore [workforce] is less innovative ÔÇ¦ the technological innovation stays
here [in the US].ÔÇŁ
Director of offshore development,
US embedded software company
ÔÇťEverything can be done offshore except for when youÔÇ™re developing new
hardware and the software for it ÔÇ“ at the same time.ÔÇŁ
Director of an offshore development center,
There is no consensus on the question of whether all software activities can now be done
offshore. There is consensus, however, that there are certain activities that are a better ´¬üt
at offshore locations while others are better to leave in-house and in-country ÔÇ“ ÔÇťonshore.ÔÇŁ
14 The fundamentals