[image: McKinsey Quarterly] <http://www.mckinseyquarterly.com/>
 [image: management]
  Interview
* Building a global IT organization: * An interview with DPWN's managing
director for IT

Stephen McGuckin explains how the company consolidated its IT operations and
developed a unique model for outsourcing and offshoring.

Michael Bloch and Marcus Schaper

Web exclusive, May 2006

Few companies have as much on-the-ground familiarity with remote locations
around the globe as Deutsche Post World Net, the corporate parent of
Deutsche Post and DHL, a logistics and parcel delivery business.

So perhaps it was natural that DPWN would emerge as a leader in the creation
of a globally connected IT-development and operations model that has
consolidated many of the company's once-fragmented application-development
and data-processing locations into three supercenters, in Cyberjaya,
Malaysia; Prague; and Scottsdale, Arizona, in the United States. Along with
some outsourced help in India, these facilities keep DPWN's IT operations
and development running around the clock. They are complemented by another
application-development center, in Bonn, Germany.

Getting to such a high degree of consolidation wasn't easy—or fully planned.
DPWN's managing director for IT services, Stephen McGuckin, who over the
past decade has held a series of senior IT positions at DHL and its parent,
didn't set out to create a 24-hour IT-development and operations program
when he began consolidating the company's Asian IT centers, back in 1997.

His goal then was to reduce costs and to improve the productivity of a host
of IT-development locations for specific countries. After a thorough study
of the possibilities, he chose Malaysia as the most central and economical.
Working with the Indian IT-outsourcing company Infosys Technologies, he
consolidated much of DHL's IT staff in Asia at the new location and
outsourced many activities previously undertaken in house.

The lessons learned from that first offshoring experience, and its
demonstrable success in terms of both cost savings and improved
productivity, were essential when the program was replicated in other
continents.

The consolidation of the IT centers led to the current model, with its three
large IT centers. In the process, McGuckin overcame deep skepticism from
many of the company's other top executives, maintained an unwavering focus
on not only the cost but also the business benefits to be gained by
consolidation and outsourcing, and learned how best to manage an outsourcing
vendor in a complex environment.

McGuckin discusses the lessons he learned along the company's path to
consolidation and outsourcing—from the first moves to locate an
application-development center in Asia to the current global-development
model—in an interview with Michael Bloch, a principal in McKinsey's IT
practice in Paris, and Marcus Schaper, an associate principal in the IT
practice in Hamburg.

The *Quarterly*: Can you explain why IT is so important for the logistics
industry and for DHL and Deutsche Post in particular?

Stephen McGuckin: It's very simply summarized in one sentence: the
information is more important than the package. Customers understand that
the physical network can break down. You can have planes delayed, trucks
tied up in traffic, or bad weather that delays the physical shipment. But
customers won't accept it if the information system breaks down. They won't
accept it if you don't tell them that something's gone wrong and give them
the opportunity to take some corrective or preventive action. They may still
not be happy, but they'll be completely unhappy if they're not informed. So
the physical network can work with less than 100 percent accuracy, but the
information system has to work all the time.

Your javascript is turned off. Javascript is required to view exhibits.

The *information is more important* than the package

The *Quarterly*: People might argue that IT is easy for DHL because you run
a very simple business model where everything is a package. That's not
really the case, is it?

Stephen McGuckin: Unlike, say, financial services, where the information is
the transaction, we have two things—the physical consignment and the
electronic consignment. It's critical to get those synchronized at the
pickup and keep them synchronized through delivery. Consider what happens at
a hub, where packages get sorted. When the package is scanned, we have less
than two seconds to retrieve the delivery address for the package from the
data record so that the sorting equipment can route the package; otherwise
it has to be handled manually. If that data's not there, the system won't
work. The planes will fly empty.

It's tough enough for a domestic shipment. It's harder if it's
international, because now we've got to get it through outbound and inbound
customs, making sure that we've got the right information across more than
228 different countries, some with multiple customs authorities or points of
entry. That all has to be synchronized, and we've got to pay all the duties,
and then we've got to prepare the invoices to collect that duty back from
the customer.

In fact, in some cases that's the simple model. Often we're doing
spare-parts logistics and we may get a call telling us that a high-tech
piece of medical equipment is broken and they're not sure which part is
necessary to fix it. The machine might be in Taiwan and the spare parts in a
warehouse in Singapore. So we send a number of parts to Taiwan and clear
customs.

We pre-alert the engineer at the hospital to say that the courier will be
there at a certain time, so that he can meet the courier, look at the parts,
and pick the right one. Then we have to reexport the remaining parts,
reclaim the customs duty that we paid on the inbound side, and put them back
into the warehouse in Singapore. So if people think it's a simple process,
they've got it wrong. The same IT intensity is true in our logistics
business unit, where we're managing complex supply chains—for instance,
managing inventories on behalf of our customers or handling end-user
returns.

The *Quarterly*: How did you develop the IT service centers in your
round-the-clock development and operations model?

Stephen McGuckin: The genesis was back in 1996, when I was managing IT for
DHL in the Middle East. We had different organizations in the Far East,
Southeast Asia, the rest of Asia, and the Middle East—each with its own
central IT unit responsible for development and deployment in the country
data centers and for managing the network that linked those country centers
with central databases and switching.

The *Quarterly*: The data equivalent of the "last mile"?

Stephen McGuckin: Yes, interfaces with local customs, whatever was necessary
for meeting local financial legislation, and so on. I was asked to take over
as CIO for the Asia-Pacific and the Middle East regions, where we had
regional teams in Singapore, Hong Kong, and Bahrain. By the way, the company
also had regional teams in London and San Francisco, so IT was headquartered
in four of the top ten most expensive places to live in the world. In the
Asia-Pacific region, we were also losing about 20 percent of our
productivity because people were traveling around the region for internal
meetings. The CEO and I decided to consolidate these IT activities, even
though the rest of the management team was skeptical.

The *Quarterly*: What was the reason for the skepticism?

Stephen McGuckin: They were worried that their projects would be delayed or
the quality would go down. Really, it was just a general resistance to
change.

We researched the best place to locate a consolidated IT center where we
could centralize much of the application development. We looked at the
Philippines, Perth, Brisbane, Thailand, Malaysia, and India, and we also
included, for the sake of our analysis, Singapore and Hong Kong. Malaysia
came out on top. There was a business-friendly government, there were
incentives, and the labor was skilled and affordable.

The government there was just getting started with the Multimedia Super
Corridor, which I can best describe as an IT-enabled business-development
zone with a set of guarantees to businesses that set up there. Two of these
were important to me. One was that telecommunications costs would be
benchmarked and prices kept competitive with Western rates. Second, there
was a guaranteed five-day turnaround to obtain a work permit for a qualified
person from anywhere in the world.

In some ways, Malaysia was still a difficult sell at that time. The
infrastructure for the Multimedia Super Corridor was still being built, and
at times there was rationing of water and power. But we managed to bring 30
good people from Hong Kong and Singapore; some of them are still there. And
success begets success, so when people saw the good work we were doing there
and the infrastructure improving quickly, we found it easier to attract
additional specialists.

The *Quarterly*: Beyond cost cutting, what did you hope to get from
consolidating IT in one offshore center?

Stephen McGuckin: While I felt that offshoring the noncritical development
activities was a smart thing to do, equally important was the business
transformation role that offshoring would facilitate. The challenge with our
model was the extended time to deploy new products and services. We could
develop a new piece of end-user software, but any requirement to deploy it
in 50 different country data centers before we could get it operational
would take us 18 months. If we had a single data center and could deploy the
software once and make it available to 50 centers overnight, then we could
reduce our time to market.

The *Quarterly*: DHL has combined offshoring with a significant effort to
outsource a number of commodity activities, mostly along the lines of
application development. What were you looking for in an outsourcing
partner?

Stephen McGuckin: One thing that was very important to me was to choose a
partner that had a process for transferring knowledge. Many of the companies
in India said they could do it, but when I asked for details they were
vague. That didn't give me a lot of confidence. But the company that we
eventually chose—Infosys—actually had a detailed, documented process, and
that was one reason we chose them. Three months after they started, they had
picked up all of the knowledge required.

I was also looking for a company with low attrition rates, because it was
important to me that, once we had done the knowledge transfer, the knowledge
should remain there rather than be lost. Infosys again made the short list
because it was the only company, at that time, offering stock options to
employees, and it had one of the lowest attrition rates in India.

A third criterion was that I wanted to be a reasonable part of the company's
revenue, so I could get its attention, but I didn't want it to be totally
dependent on me. I was looking for an outsourced vendor where the volume of
business that we would offer would be between approximately 5 and 10 percent
of the whole.

A fourth requirement was that the outsourcing vendor should be flexible in
meeting my business model requirements. There are essentially two models for
outsourcing. One is time and materials: you hire a team and then it's up to
you to keep them productive. The other model is fixed price: you give a
detailed specification and get a quote for delivering that project in a
particular time frame for a particular price. I wanted both: a stable team
but one that would bid for every piece of work, so that the vendor had to
optimize the team, not me. We had long discussions about this with Infosys
and eventually came up with a unique model that offers a minimum guarantee,
which allows them to keep a team in place. They bid for every piece of work
we ask them to do. We agree to the amount, and then the minimum guarantee is
decremented. We now have a large team working for us at Infosys, and we've
expanded our work there to become one of their largest customers.

Within a year of the transition, *we had cut IT costs* by 40 percent

The *Quarterly*: Can you tell us about the transition from managing
employees to managing a vendor's staff?

Stephen McGuckin: In my opinion, most outsourcing initiatives fail because
the customer does not know how to manage the vendor. Managers are skilled at
managing teams of technicians, not an external vendor. They must be
retrained. For example, in the beginning there was conflict between my
people and Infosys because we would disagree on whether something was a
change or just an amplification of a requirement. My people would come to me
saying, "This is never going to work. It's hopeless." It took about six to
nine months and lots of education before the model really settled down.

The *Quarterly*: How did it settle down—with more rigor on the side of DHL
in determining its specifications, more flexibility from the vendor, or a
little bit of both?

Stephen McGuckin: I think a bit of both, but certainly there was more change
on the DHL side. Our people learned that they were now managing a vendor and
not programmers. We introduced the rigor that we needed in our
specifications in order to manage the vendor rather than the informality
that surrounds an internal IT department. We also had to learn the right
level of outsourcing. We had outsourced everything from high-level design
all the way through to integration testing.

But after about six months, my team told me we had outsourced too much and
no longer had enough visibility into the design of applications. This
limited our ability to judge the merits of the vendor's bids. We decided to
leave the detailed design with the vendor, but we brought the high-level
design and architecture back in house. Thereafter, the team felt it could
control things and could really understand whether the outsourced partner
was overbidding or underbidding.

The *Quarterly*: How did you manage your relationship with the business
leaders who opposed consolidation and outsourcing?

Stephen McGuckin: We migrated the first services to Malaysia in August 1997
and by October 1997 had our official opening. For that, we invited the CEO
and his management team so they could appreciate that there were people
there and there was a service delivery capability. We didn't have any loss
of service during this transition while we closed down Hong Kong and
Singapore and moved those operations.

We had some loss of productivity—some projects were delivered later, and
there might have been quality problems with them. But when the senior
managers started to see the Malaysian operation, their attitudes began to
change. It was probably another 18 months before there was complete
satisfaction that this move had been a good thing. Before the move, the CEO
had seen his IT costs escalating by 15 to 20 percent a year. Within a year
of the transition, we had cut IT costs by 40 percent while returning quality
and delivery to the levels before the move. So he was pretty satisfied that
he was getting value for money. The working model was now starting to show
real dividends, quality product was coming out, and we were ready to take
things to the next stage.

The *Quarterly*: When you became the CIO for the United States, you were not
looking for the lowest costs there, right?

Stephen McGuckin: I would say that the required skills involved change
management. It was not a well-run environment. It was impossible to get a
standard process in place, because people would just do things their own
way. They were isolated in their little fiefdoms, and you could never
actually get to them. There were eight different IT centers.

The *Quarterly*: All sharing application development?

Stephen McGuckin: All not sharing. All trying to avoid sharing. We had to
choose the right location to consolidate these operations.

We looked at a number of locations, and Scottsdale came out the best; it had
a qualified pool of people and was relatively free from natural disasters.
It wasn't, perhaps, the lowest-cost location, but it was reasonably priced.
We had considered South America and Costa Rica, but questions about the
maturity of the commercial infrastructure and the high price of
telecommunications ruled these out.

Scottsdale, judged by all the criteria, was best. Unlike Malaysia, where we
started from scratch, in Scottsdale we had a small head start because we had
about 60 people there; we had some seeds that we could develop. Again, speed
of execution was extremely important, so we announced the move in November
2001 and we had the operation up and running by the end of May 2002.

The *Quarterly*: Was Infosys working with you on Scottsdale?

Stephen McGuckin: Yes, we just executed the same program again—brought
Infosys in, got them to do the knowledge transfer, and then we undertook the
migration of the higher-value-added activities. Now, this migration was a
little bit more complicated because the US was a country site as well as
providing support for the Americas region. This was a higher-risk
proposition than the one we initially had in Malaysia, but we now had a
little bit more experience and brought over some members of the team who had
done that sort of work before.

The *Quarterly*: Your European IT center, in Prague, is the newest and
largest of the three, with more than 1,000 people. How was the European
migration different?

Stephen McGuckin: In Europe, it was even tougher because we had three big
companies merging, each with €6 billion or €7 billion in revenue a year and
different heritages and standards. Each believed that its own processes,
standards, and solutions were optimal, but no one location was big enough to
support the data-processing requirements of the combined and growing DHL
business. We looked at finding the European location as a procurement
exercise. And Prague was the optimum—like Scottsdale, not the cheapest, but
the best in terms of English-language speakers, air links, the capabilities
of the local workforce, and the receptiveness of the Czech government.

The *Quarterly*: Some people said that when DHL went into Prague and hired
1,000 IT people, it would dry up the labor market and wage rates would
rocket.

Stephen McGuckin: This was obviously a concern, particularly since the Czech
Republic was due to join the EU, which many people expected to result in
increased labor costs in the region. However, I didn't see how an increase
in the supply of labor in the EU would result in a price rise. It seemed
pretty safe to me.

The *Quarterly*: Still, hiring 1,000 people in a few months cannot have been
easy.

Stephen McGuckin: When we announced our plans for a large IT-development
center in Prague, we had a big official signing. We blitzed the papers and
had advertisements in the trams, so we got the maximum publicity that we
could just before we started recruiting. We had job fairs in Prague, in
Ostrava, Brno, Bratislava, and a few other places. Actually, we have more
than 45 different nationalities working in Prague. We were able to attract a
significant number of Czech nationals who were working abroad and looking to
come back home and find highly challenging jobs.

The *Quarterly*: Do you see any limits to the model?

Stephen McGuckin: No, I don't see any limits. I think we won't hop about as
frequently as companies do in the manufacturing world. I think there may be
a time when we move again, but I look at it as a ten-year horizon.

I think you want to get in early, get in fast. You can't do this without an
element of risk, so you have to be able to identify and manage the risks.
But if you're not prepared to identify and manage those risks, you'll never
move, because moving will just seem too big a job.
 About the Authors

*Michael Bloch* is a principal in McKinsey's global IT practice. He is a
leader of the outsourcing and offshoring practice and is based in
Paris. *Marcus
Schaper*, an associate principal, specializes in managing the performance of
IT, particularly software and applications. He is based in Hamburg.

This article was first published in the Summer 2006 issue of *McKinsey on IT
*.

Site Map <http://www.mckinseyquarterly.com/site_map.aspx> | Terms of
Use<http://www.mckinseyquarterly.com/terms.aspx>| Updated:
Privacy Policy <http://www.mckinseyquarterly.com/privacy.aspx> |
mckinsey.com

Copyright (c) 1992-2006 McKinsey & Company, Inc.


[Non-text portions of this message have been removed]



 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/Indo-Linux/

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



Kirim email ke