Cisco - Collaborating On New Product Introduction

Management Information
Systems
Cisco Systems, Inc.: Collaborating On New
Product Introduction
Overview
On November 13, 2007, more than 100 employees of Cisco Systems,
Inc. assembled in classic Cisco fashion: they dialed in from multiple
locations around the world for an important meeting.
The purpose of the gathering was to get the green light from senior
management to manufacture a new high-end router that would make
the giant networking company more competitive in an age of surging
Internet traffic.
The project’s code name, Viking, said it all.

Overview
The router for broadband service providers would break ground in
power and speed, reminiscent of the Norse warriors and explorers of
Europe during the eighth to eleventh centuries.
The meeting represented a culmination of several years of

development work by a cross-functional, global team of Cisco
specialists in engineering, manufacturing, marketing and other areas.
Just months earlier, in mid-2007, Cisco overhauled the project by

sharply boosting the router’s speed and capacity.
Overview
This would allow the company to leapfrog competitors and offer a low-
cost, powerful new router platform for the next 10 to 15 years.
That day in November, the Viking team was seeking an “execution

commit” from senior management in manufacturing.
If it got the go-ahead, Cisco would be ready to commit the resources to

launch the new product.
But the Cisco team knew it faced many challenges.

Overview
The Viking project would be one of the company’s most complex new
product introductions ever.
First, even though the project had been essentially re-started in mid-
2007, Cisco was still aiming to announce the machine in November
2008.
That would give it just a year to line up manufacturing, supply chain and
marketing arrangements—an unusually accelerated schedule.
Overview
Second, Cisco, which outsourced virtually all its manufacturing, wanted
to start making the high-end router immediately in a low-cost location:
China.
This differed from Cisco’s past practice of outsourcing a complex new

product in the United States first and later shifting overseas once
production matured.
Third, Cisco proposed to use one of its contract manufacturers,

Foxconn Technology Group, to build the product in Shenzhen, China,
and to give Foxconn a broad role overseeing manufacturing and the
supply chain.
Overview
But Foxconn had never made such a complex product for Cisco.
Could Foxconn handle the technical complexity while producing the

router from Day One in China?
Could Cisco find ways to engage closely with Foxconn and mitigate the
risks?
Finally, Viking would test Cisco’s well-honed new product introduction,

or NPI, expertise.
Overview
The project would require tremendous global collaboration among far-
flung teams within Cisco, and close coordination with Foxconn’s
manufacturing site half a world away from Cisco’s San Jose
headquarters and labs.
Cisco’s Track Record and Competencies –
Company Background
Cisco was a classic Silicon Valley story of success, rising from geeky
start-up to world-class corporation.
The company was founded in 1984 by Len Bosack and Sandy Lerner, a
husband-wife duo of Stanford University computer specialists who
experimented with networking different buildings on campus.
Their work led to the creation of the “multi-protocol” router that enabled
disparate computer networks to talk to each other, in much the same
way as different telephone networks were linked.
Company Background
The company became the leader in selling networking gear to
corporations; it quickly expanded into sales to telecommunications and
broadband service providers and, later, to consumers.
In the 1990s, Cisco emerged as a leader in networking technology for

the Internet age, respected for both its technology and management
practices.
At the helm was John T. Chambers, who became chief executive officer
in 1995 and went on to become one of the world’s most visible CEOs,
considered a tech visionary and management guru.
Company Background
Cisco enjoyed meteoric growth during the 1980s and 1990s.
It went public in February 1990.
Just eight and a half years later, its market value topped $100 billion,
reaching that mark faster than any company in history.
In March 2000, for a brief time during the Internet boom, Cisco became
the world’s most valuable company, with a market capitalization of more
than $500 billion.
Company Background
Even after the dot-com crash, the company grew at a healthy clip.
Revenue increased at a compound annual rate of 13.1 percent

between 2002 and 2008.
The company earned $7.3 billion on sales of $34.9 billion in fiscal 2007,
which ended July 28, 2007.
It made a profit of $8.05 billion on revenue of $39.5 billion in fiscal

2008, which ended July 26, 2008.
Collaboration and Globalization
By 2007, nearly a quarter century after its founding, Cisco operated in
more than 120 countries and employed 61,535 people.
The business was organized into more than 40 business units, defined
roughly by product spaces, such as core routing, edge routing, access
routing or wireless networking.
Business units belonged to a half dozen broader groups such as

service provider; data center, switching and services; access
networking and services; software; or consumer and small business.
The broader groups made up Cisco’s engineering or development
organization, which was run by a “development council” rather than a
single executive.
In addition, business units had access to company-wide resources

such as sales, services (known as customer advocacy within Cisco),
marketing or manufacturing.
In effect, the company was organized both vertically and horizontally—

or in a series of overlapping circles, as some described it—in order to
promote multi-functional teamwork, a Cisco hallmark.
In December 2006, the company announced the selection of India—
where it already employed more than 2,000—as the site for its “Cisco
Globalization Center East,” the first in a series of globalization centers
envisioned.
The Bangalore center would develop new businesses and tap into
India’s technical brainpower. Cisco dispatched a senior executive from
San Jose to head the center in Bangalore and to serve as “chief
globalization officer” for the company. It was a step toward Chambers’
goal of locating 20 percent of Cisco’s top leadership outside the United
States by 2013, shifting the company toward globally distributed,
collaborative management.
Creating a Global, Flexible Supply Chain
By the early twenty-first century, virtually all of Cisco’s production was
done by contract manufacturers in their network of factories around the
world.
Cisco believed that outsourcing enabled it to tap the most cost-effective

manufacturing resources worldwide and to leverage its supply chain
partners.
Cisco itself would add value by managing the supply chain and
focusing on product design and development.
As the outsourced model became more sophisticated, Cisco’s contract
manufacturing partners took on increased responsibility for components
planning and procurement, order scheduling, designing manufacturing
processes, and overall supply chain management.
Angel Mendez, Cisco’s senior vice president of worldwide

manufacturing, said, “I think what we are doing, which is somewhat
unique, is driving an adaptive supply chain in a very large outsourced
model across a very large spectrum of products and geographies.
That combination is an interesting thing.”

In the early twenty-first century, Cisco consolidated its base of contract
manufacturers and suppliers.
It cut the number of contract manufacturers from 13 to 4 major ones by

late 2006.
This enabled it to leverage spending while working more closely with

them.
In winnowing the field, Cisco examined numerous aspects of its
contract manufacturers: capacity to build in large volumes in multiple
locations, ability to support a broad range of Cisco products, as well as
design knowledge and new product introduction capabilities.
Similarly, Cisco sharply reduced the number of vendors in its extended

supply network.
In early 2001, Cisco had close to 1,500 suppliers and 80 percent of its
spending went to about 200 of them.
By late 2006, it had about 600 suppliers and 90 percent of spending
was with just 95 of them.
The changes made it less costly and simpler to manage suppliers and
also resulted in major cost savings on components.
Many of Cisco’s manufacturing and supply chain improvements were

made under its Manufacturing Excellence or “MX” initiative, launched in
2005.
The aim was to promote general excellence in manufacturing by
emphasizing major “aspirational” improvements rather than incremental
changes, according to Mendez.
In early 2006, the company began shifting formally to a manufacturing

model called Cisco Lean, intended to boost efficiency and flexibility.
The lean process, developed with Cisco’s contract manufacturers and

suppliers, was completed in mid-2007.
The goal was to convert Cisco and its extended supply chain to a
system in which product was built only after a customer had actually
ordered it.
Explaining Cisco Lean, Mendez said: Traditional manufacturing

operates utilizing a ‘push’ model.
In other words, a company builds product based on what it forecasts

customer demand to be.
With a push model, the extended supply chain including contract
manufacturers and suppliers are dependent on the accuracy of
forecasts, which can vary.
Lean manufacturing is a “pull” model, which means that product is not

built until the customer has already placed the order.
This is also known in the industry as “just in time” manufacturing. The

benefits of lean manufacturing included reduced inventory across the
extended supply chain, more predictability in lead times and on-time
shipment, and simplified processes, according to Mendez.
Cisco was a leader in using technology to create information links
throughout its extended supply chain.
Its Autotest system, for instance, captured real-time data from the
facilities of contract manufacturers, providing a “one window” view of
production lines around the world.
In addition, according to Mendez, Cisco used information networks to

collaborate with customers, contract manufacturers and suppliers in
areas such as demand management and planning, product quality
improvement, and lean manufacturing.
Innovating in New Product Introduction
Cisco’s business model called for boosting growth in part through rapid,
aggressive introduction of a wide range of communications products.
Indeed, the company brought more than 250 new products to market in
fiscal 2008.
The company had a well-oiled machine for new product introduction, or

NPI.
The effort was led by a business unit’s product development teams,

with coordination and support from many other functions and groups
within Cisco, as well as customers and supply chain partners.
New product development required both technical expertise and
management talent to understand the market, translate market needs
into a product, and bring the product to market quickly.
At the outset, marketing and engineering would talk to customers and

define the product features.
Then engineering and manufacturing would work together to ensure

that the product design met market requirements and could be cost-
effectively produced.
Employees from finance would help with budgeting and calculating a
project’s return on investment.
The supply chain organization would work closely with the NPI team to
influence which technologies and suppliers Cisco would use in the new
product.
Manufacturing specialists would ensure that the supply chain

arrangements would facilitate a smooth ramp-up to commercial
production and allow for lower manufacturing costs later in a product’s
life cycle, when profit margins would shrink.
Cisco’s NPI process involved three major stages: strategy and
planning; execution; and deployment.
The process included a series of checkpoints, or “gateways,” between

each stage (Exhibit 1).
The “concept commit” gateway came at the beginning of the first stage,
after designers had brainstormed design ideas.
This checkpoint ensured that a cross-functional team had approved a

“product requirement document” and an associated business plan.
This gateway also ensured that sufficient resources would be
committed to get to the product design point.
Once a Project cleared the “concept commit” checkpoint, the design

would be fully defined and verified.
By the end of the first stage, designers would have developed a

definitive design specification for the product.
Next came the “execute commit” checkpoint, which ensured that the
cross-functional team agreed on design specifications and was
committed to dedicating the resources needed to ship the product on a
particular date.
The project then entered the execution phase in which engineering

worked with manufacturing to develop and test prototypes.
Manufacturing would do a thorough “design for manufacturability”

review early in the prototype development and would help engineering
develop a product that was easily testable.
Normally, Cisco aimed to do two rounds of prototyping, although
additional prototyping sometimes was needed.
A final “technical readiness review,” or TRR, would go through results of

the prototyping, such as manufacturing yields and quality metrics.
After prototyping, there would be a “pilot” build of the product at the

manufacturing site.
At this point, primary responsibility for the new product would shift to
manufacturing from engineering.
Near the end of the execution phase came the “orderability” checkpoint,
which ensured that Cisco could hit the targeted ship date, meet the
expected ramp-up in demand, and that suppliers were lined up and
ready.
Following the orderability gateway, the product would be added to

Cisco’s price list and entered into the deployment stage.
It was ready for release, meaning it could ship in the volumes and at
the quality levels required.
This stage included “first customer shipment,” or FCS, when the
product was released to one or more lead customers that typically had
given input on defining product features.
After FCS, there was another gateway, known as “time to quality and
volume,” or TTQV.
This checkpoint, which included analysis on production yields and

costs, would ensure that Cisco’s contract manufacturing sites were
achieving six-sigma quality goals and could make the product cost-
effectively in high volumes.
The TTQV check typically was done two to three months after
production had begun.
Finally, there was a “post-production assessment” to identify and

capture lessons learned from the project.
The team would discuss what had gone well or not so well, and how to
incorporate the successes and avoid the mistakes in future
development projects.
Viking Product and Market
The Viking router was four years and $200 million in the making. In late
2004, Cisco had started work on a router that would let broadband
service providers consolidate their data, voice, video and mobile traffic
in a single box.
But as Cisco worked with a few lead customers to define the product, it
became apparent that it would not meet the escalating needs of service
providers.
By fall 2007, Cisco completely redesigned the proposed router for

increased speed and capacity.
In November 2007, Cisco management committed to producing a much
more powerful router than originally envisioned.
But the company did not change the original target launch date of
November 2008.
Only a year later, on November 11, 2008, Cisco unveiled a key addition
to its product portfolio:
the ASR 9000 router, or Aggregration Services Router.

The ASR 9000 would allow carriers to consolidate data streams from
disparate networks based on Ethernet, a popular networking protocol.
When fully loaded, the machine could transmit 6.4 terabits of data per
second, enough capacity to deliver high-definition video streams to all
1.2 million homes in Los Angeles simultaneously.
This would achieve a new high point among routers of its class.
Viking Product and Market – Market Needs
The ASR 9000 represented a new generation of “edge” router.
The device was aimed at easing bottlenecks at the network “edge,”

where service providers’ long-haul networks met end users’ local
networks.
The router would consolidate voice, data, video and mobile traffic into
larger data streams to feed into the core of service providers’ networks.
This was akin to a postal system with local post offices feeding central
sorting facilities.
Carriers were seeking technology to help them cope with the flood of
traffic and to move Internet content to the network edge, where it would
be closer to users and thus more easily and quickly accessed.
Cisco expected the edge router market to grow significantly between

2007 and 2011.
The ASR 9000 would be a critical asset for Cisco in the charge to
expand the company’s presence and market share in this important
market.
Cisco saw a major market transition happening as network traffic
soared, fueled by greater broadband access and increased use of
video technologies such as YouTube, high-definition television, movie
downloads and mobile video.
The company predicted that global Internet Protocol-based traffic would

grow at a compound annual rate of 46 percent from 2007 to 2012,
nearly doubling every two years.
As a result, bandwidth demand would reach approximately 522

exabytes, or more than half a zettabyte, in 2012.
This demand would be equivalent to downloading 125 billion DVD
movies a month.
Broadband carriers needed infrastructure for what Cisco called “visual

networking in the zettabyte era.”
They needed equipment designed specifically to handle the demands

of video, which used far more bandwidth than other types of data.
Ron Westhauser, senior director of product operations at Cisco, noted,

“Video had a huge and rather sudden impact on the amount of network
traffic.”
The Viking’s “speeds and feeds” would be major competitive factors,
according to Henrik Jensen, director of product operations.
Service providers wanted high speeds and high capacity as well as a

compact footprint, power efficiency, low costs per connection, and
importantly, the ability to scale up the router’s capacity over a number
of years by adding denser circuit cards.
Carriers also wanted hardware and software features in the router that
would let them better manage their networks and offer customers
revenue-generating services, such as ad insertion.
Viking Product and Market – Product
Features and Positioning
The ASR 9000 could be configured and upgraded according to a
carrier’s needs.
There would be two “chassis” choices—a box 93.35 centimeters (36.75

inches) high weighing up to 170.45 kilograms (375 pounds) or a box
44.45 centimeters (17.5 inches) high weighing up to 140.55 kilograms
(230 pounds).
The box would have up to eight slots inside for “line cards” or circuit
boards.
Each card would have “ports,” or connections, for plugging in cables
and handling data traffic.
Initially, each card could handle traffic at 80 gigabits per second (or 160
gigabits per second of full duplex traffic, which counted data sent and
received at the same time).
A card could be upgraded, with more ports, to process up to 400

gigabits per second of traffic (or 800 gigabits per second of full duplex
traffic).
So the machine, when fully loaded and upgraded, would have a total
capacity of 6.4 terabits per second.
That would give carriers lots of room to expand in order to meet the
explosion in network traffic.
Cisco viewed the ASR 9000, which carried an entry-level price of

$80,000, as an edge router platform for the next 10 to 15 years (Exhibit
2).
The router would allow Cisco to leapfrog competitors.
Its total capacity of 6.4 terabits per second would be about six times
greater than that of competing routers.
In late 2007, Juniper Networks Inc. and Alcatel-Lucent had newer edge
routers on the market than Cisco had.
Juniper had launched its MX960 machine in early 2007 and Alcatel had
introduced its 7750 router, the result of an acquisition, in 2003.
Cisco’s previous high-end edge router, the 7600, had been released in
2001.
Juniper’s MX960 could be scaled up to 1.1 terabits per second of

capacity over 11 high-density circuit cards, according to an internal
Cisco assessment.
Cisco’s 7600 router could process 720 gigabytes per second of traffic.
The ASR 9000 would be a successor to the 7600 router; it could be
used as either a replacement for it or a complement to it. It also could
help Cisco pull further ahead of Juniper in the edge router market.
While Cisco accounted for 54 percent of the service provider edge

router market in the third quarter of 2007 and commanded an 84
percent share of the enterprise router market, according to market
researcher Dell’Oro Group, Juniper had chipped away steadily in the
carrier edge router market, holding a 16 percent share.
Cisco knew it had to exceed the Juniper MX960’s capacity.

Viking Product and Market – Time-to-Market
Pressure
By November 2007, the Viking team also realized it had no time to
spare.
Cisco wanted to jump on the market transition to video networking.
Service providers needed bandwidth at lower prices.
If Cisco did not meet that demand within 12 months, it could lose
market share.
Pressure
Recalling the urgency, Westhauser said: The explosion of video
happened faster than anyone could have expected due to the explosion
in traffic from YouTube, iPods and iTunes, and the extension of video-
on-demand downloads to the living room.
A product that could provide lots of bandwidth at a lower price point—it

was just what the market wanted….
It’s part of Cisco’s culture to catch market transitions.

Pressure
It was a huge market transition happening with a lot of customers
willing to spend a lot of Money buying new equipment.
The upshot was that, after the Viking project was “reset” in November
2007, Cisco would have only one year to launch the beefed-up router.
This was much faster than the typical three to five years it took to
develop a high-end router for service providers.
Viking Product and Market – Cost Pressure
The Viking team faced huge cost pressures.
Bandwidth prices were constantly falling and customers expected

continuous improvements in price-performance on their equipment.
In the aggregation router space, competitor discounting was intense.
The Viking router would have to be very advanced but very cost-
effective to replace service provider equipment that already was fully
depreciated.
“It was a very cost-challenged space,” according to Westhauser.
There were two specific challenges: keeping the machine’s cost per
port low and ensuring that carriers could upgrade by using line cards
with higher port count.
Early on, Cisco set targets for both cost per port and port density.
These targets would be achieved through sound product design and
optimal supply chain arrangements.
Indeed, the Viking team had a clear target cost to meet.
Leticia Jensen, senior manager of product operations, said, “We knew

we had to implement the most cost effective-supply chain at launch.”
Viking Product and Market – Other
Considerations
The Viking router’s technical complexity would be immense.
Inside the metal chassis were 10 multilayered printed circuit cards,

including eight cards for handling data traffic and two cards for overall
processing.
There were processors, specialized ASICS chips, pins, connectors, 65

custom parts, up to six power modules and two fan trays.
In all, the router contained about 300,000 components, about 30 times

more than in a small business router.
Considerations
Cisco also faced the challenges inherent to outsourcing production of
such a complex machine.
The contract manufacturer would have to put all the pieces together
with the highest quality, reliability and on-time performance required in
the demanding service provider market.
Cisco would have to work closely with the contractor to reduce

production and supply chain risks.
Considerations
Finally, Cisco needed to ensure that the router would be attractive to
service providers worldwide.
Emerging markets were the fastest-growing part of Cisco’s business,

so keeping the router’s costs down was important to its global success.
Facing The Viking Challenge
The Viking team, which at one point involved at least 300 Cisco
employees, faced a series of decisions in carrying out the complex
program.
Its objectives were to: 1) identify the market needs and define the
product features, including the price-performance balance; 2) produce
a design that met customer requirements; 3) design and set up the
optimal supply chain for cost-effective manufacturing; 4) take steps to
ensure success, such as doing extensive prototyping, testing and
debugging during the development process; 5) set up a marketing and
pricing plan that would ensure sales success and help meet financial
objectives for the product.
Throughout the process, close collaboration was imperative, both
inside and outside Cisco.
The project followed Cisco’s CPDM (Cisco Product Development

Methodology) and NPI methodology that included extensive checklists
and “gateways” to ensure discipline and rigor as decisions were made.
Each step would involve numerous tasks to be completed before

development could proceed to the next step.
Cisco employees used a web-based tool called “NPI Metrics” to monitor
and track progress.
The NPI Metrics “to-do list” for Viking contained up to 325 tasks to be
completed over 12 major phases—from “concept commit” in June 2007
through “time to quality and volume” in May 2009.
Each task was assigned an “owner,” a due date, and a definition.
Yet, Viking was not by any means Cisco’s largest development project.
That distinction belonged to Cisco’s CRS-1, a massive 92-terabit-per-
second core router launched in 2004 that took more than four years,
$500 million and 500 engineers to develop.
The CRS-1 was a mega-project that represented an “engineering

marvel,” according to Sri Hosakote, vice president of engineering.
By contrast, Viking was “a very complex execution marvel,” he said,

since it involved many different geographies, close cooperation among
numerous teams within Cisco, and an accelerated development
schedule.
Facing The Viking Challenge – Early
Customer Engagement
From the outset, Cisco engaged telecom customers closely in order to
define the product’s features.
In fact, it was the early feedback from several lead customers that had
prompted Cisco in mid-2007 to revamp the project by sharply
increasing the proposed router’s capacity and speed.
Noting the importance of establishing a two-way street with customers,

Brendan Gibbs, senior director of product marketing, recalled:
Customer Engagement
This was a big strategic decision very early on.
We started working with some very, very large strategic customers of

ours.
We agreed to modify the product definition and the road map based
upon what their needs were, the idea being that if we partnered very
early, we'd not only get very candid and good feedback to product
development, but it would also help these customers and our mutual
relationship. And we really based the whole program of development on
that.
Customer Engagement
Later, Cisco continued the tight relationship by giving key customers
sneak peeks at initial prototypes, as early as nine months before first
customer shipments were scheduled.
The lead customers also tried out and gave feedback on Viking
prototypes.
Facing The Viking Challenge – Building in a
Low-Cost Location
It became clear early on that the router would have to be built in a low-
cost, overseas location.
But offshoring production from the get-go would be a departure from

the way Cisco previously had launched complex products.
In the past, a new high-end router would initially be produced in a

contract manufacturer’s U.S. facility.
Later, after production had matured and stabilized, the contract

manufacturer would shift manufacturing offshore, typically to Thailand
or Malaysia.
Low-Cost Location
Manufacturing of the 7600 router, introduced in 2001, had been
launched in Florida and moved offshore starting in 2004.
By November 2007, the Viking team was proposing that manufacturing

be launched in China, which had an ever-expanding and ever-
improving electronics manufacturing base.
The team had weighed the benefits and risks of launching in China.
Low-Cost Location
Cisco would get the benefit of low-cost production right away. Avoiding
a transition to another manufacturing site later on would save money,
time and effort.
Cisco would open up the long-term possibility of using China as a

manufacturing base for other complex products.
The risk was that the Chinese factory might prove unable to handle the
router’s technical complexity or would fall short on the quality, reliability
and on-time delivery needed for a carrier-class product on which there
was virtually no room for error.
Facing The Viking Challenge – Selecting
Foxconn
In mid-November 2007, Cisco made a bold decision: it awarded the
Viking manufacturing job to Foxconn, a fast-growing contract
manufacturer with extensive operations in China.
One of four global contract manufacturers that Cisco regularly used,

Foxconn had never built such a complex product for the big networking
company.
In the past, Cisco’s high-end routers had been produced by its three
other major contract manufacturers—Flextronics, Jabil and Celestica.
Foxconn
Foxconn had produced simpler, high-volume items for Cisco, such as
Voice-over-IP phones, desktop switches and wireless network routers.
Foxconn had been progressing well in meeting Cisco’s expectations.
It had won “Transformation Partner of the Year” awards from Cisco

twice in the previous three years.
In August 2007, the company had just received its “Level 3”

qualification from Cisco, after passing tests qualifying it to build the
most complex products for the networking company.
Foxconn
Cisco rated and qualified its contract manufacturers using three levels,
with Level 1 indicating the simplest products and manufacturing
processes, Level 2 indicating a middle ground, and Level 3 indicating
the most complex products and processes.
Flextronics, Jabil and Celestica previously had qualified at Level 3.
Before officially awarding the Viking job, though, Cisco did a full
technical assessment of Foxconn and asked it to build and run dozens
of test circuit boards through its soldering, assembly and other key
processes to validate full capability.
Foxconn
The result was that the Viking router would be built at Foxconn’s
massive, walled manufacturing site in Shenzhen, China, where less
complex Cisco products were made.
Foxconn had won an opportunity to move up the food chain.
Leticia Jensen said, “Over the assessment period, we noted the high
executive level commitment, excellent team, and a demonstration of
processes and technical capability.
Foxconn proved they were committed and ready to take on this

challenge.”
Foxconn
For Cisco, choosing Foxconn was a high-risk, high-reward decision.
The upside would come from validating this partner’s ability to

successfully make a complex but extremely cost-sensitive product.
But there would be a downside if Foxconn proved unable to handle the

Viking router’s technical complexity with the quality, reliability, speed
and low costs that Cisco needed.
Foxconn
The question for the Viking team was: Should Cisco take the technical
risk or the cost risk? “Option A” would be to pick a proven, technically
savvy partner and work with it to reduce technical risk and drive down
costs.
“Option B” would be to select a low-cost provider with less experience

in this technology and complexity and work through the learning curve
with it, according to Westhauser.
Foxconn
He said, “When we did the analysis, we decided to go with low cost.
You find when you design processes, it’s harder to turn a Mercedes into
an entry-level Toyota than to turn an entry Toyota into a Lexus.
Mercedes has never been able to put out a low-cost car, but Toyota
was able to migrate from entry-level cars up to Lexus.”
A major argument for Option B was that low-cost manufacturing

required certain fundamental disciplines that were hard to teach.
Foxconn
Westhauser observed: There are disciplines and mindsets in achieving
low costs.
One of the mindsets is a passion for eliminating and finding waste.
It's a mindset of do it right the first time.
It's a mindset of do it as simply as possible.
We thought we would better be able to teach someone with a low-cost

mindset and low-cost disciplines to handle the complexity.
Foxconn
There were also long-term benefits to Cisco from developing a low-cost
supplier—even if it had to spend more time, money and effort in the
short term to go through the extensive validation that was required.
If Foxconn succeeded with Viking, Cisco would gain flexibility in its

choice of contract manufacturers for future generations of high-end
products.
Leticia Jensen said, “If Foxconn proved that they could manufacture
high-quality, high-end products, they would be another choice in our
portfolio. In the ideal supply chain, all our partners are capable of
manufacturing all our products.”
Facing The Viking Challenge – Viking Supply
Chain
But the trump card was Foxconn’s vertical integration.
Cisco liked the idea that many different pieces of the Viking
manufacturing could be done on a single campus in China, sometimes
in adjacent buildings.
Foxconn’s massive site in Shenzhen could handle a wide range of

tasks.
Chain
These included making tooling and molds; performing sheet metal and
plastics work; using surface mount technology to load components onto
circuit boards; doing product assembly; making key parts such as the
metal chassis; and doing testing and integration work.
In addition, Foxconn, as the world’s largest electronics contract

manufacturer, had deep experience in procuring components, an area
where it enjoyed extensive buying power and economies of scale.
Cisco carved out a broad role for Foxconn.

Chain
The contract manufacturer would be responsible for all major
subassemblies, except power modules, fans, and optical technology
plug-ins, which would come from other vendors.
Foxconn would assemble the router’s chassis, “backplane,” circuit

boards, power tray, and fantrays in Shenzhen.
These subassemblies would be trucked to Foxconn’s “direct fulfillment”

site in Hong Kong, about a two-hour drive away.
Chain
The Hong Kong site would do final assembly, testing and customer
configuration of the product, which would be shipped out by a third-
party logistics company located at Foxconn-Hong Kong.
There would be about a dozen major suppliers, although Foxconn

would buy components from a few hundred suppliers in all (Exhibits 3
and 4).
The Viking supply chain was tightly focused on a single company.

Chain
Chain
Chain
In the past, by contrast, Cisco might bid out work on the chassis,
backplane, printed circuit board assembly, system integration, and
direct fulfillment to separate suppliers to be done at different sites.
In Viking’s case, Leticia Jensen said, “We were looking to streamline

the whole supply chain to a mega-site where we had all the different
elements being manufactured real-time and coming together very
quickly.
A single site being accountable for all the major pieces of the supply
chain created an agile structure.
Chain
The ability to react to market demand shifts with speed was very big for
us.”
In addition, the extended supply chain was concentrated in Asia, to be

near Foxconn and to tap the region’s strength in low-cost production.
Foxconn would be purchasing electronic parts from a list of suppliers

qualified by Cisco’s global supply chain management group.
Chain
Most of the components would come from suppliers in China, Malaysia
or elsewhere in Asia.
Monica Patel, new lead product program manager observed, “What

they’re trying to do here is satisfy end-to-end as much as they can in
Asia long-term.
It makes it much easier and more cost-effective.”

Facing The Viking Challenge – Foxconn’s
Vertical Integration
The Viking team saw numerous benefits in Foxconn’s high degree of
vertical integration.
It could make manufacturing more flexible, responsive and efficient,

resulting in lower costs, easier logistics, shorter lead times, simpler
information flow, and easier coordination.
Leticia Jensen noted, “Foxconn could provide us with a fully integrated

solution that not only provided a responsive and cost-efficient supply
chain, but it also improved the ease of doing business for our team.”
(Exhibit 5)
With Foxconn handling many aspects of manufacturing in Shenzhen
and overseeing final assembly in Hong Kong, most of the Viking
production could be located within a two-hour drive.
For instance, printed circuit boards could be trucked to the direct

fulfillment site in Hong Kong rather than being transported by ship
between, say, Malaysia and Mexico.
This could sharply cut transportation costs and save days of shipping
time.
It could also reduce inventory needs and save money tied up in holding
inventory.
Cisco could react more quickly to fluctuations in customer orders; the

Hong Kong fulfillment center, for instance, could quickly replenish its
inventory of circuit boards, since they were assembled in Shenzhen,
just north of Hong Kong.
The arrangements could result in shorter lead times to get product to

customers.
Foxconn’s vertical integration also meant Cisco could eliminate price
markups that normally came with having numerous layers of suppliers.
Kevin Sin, manager of program management for Foxconn, said, “One

of the reasons we have a lower cost is that we are so integrated.”
The “one-stop shopping” that Foxconn offered could make it simpler

and easier for Cisco to manage Viking manufacturing, since it was
dealing mostly with a single contractor.
Relying on a single supplier to handle many manufacturing tasks on a
single campus also could simplify information flow throughout the
supply chain.
Sin noted, “Everything can be done internally (to Foxconn).
It’s the same campus. If there’s a problem, I can ask them to come over
to talk to us.
Nothing can beat face-to-face.”

Yet, there were tradeoffs to Foxconn’s vertical integration.
Cisco ran the risk of being overly dependent on a single supplier and
whatever financial and operational constraints it had.
This could present a higher risk for this product than if the supply chain
were fragmented.
Leticia Jensen acknowledged, “You’re basically putting all your eggs in

one basket and making a bet on your partner’s long-term success.”
The tight integration also meant that Cisco could miss out on the
chance to use suppliers that might be even more efficient or skilled
than Foxconn at individual steps in manufacturing.
And if there were a natural disaster in southern China or a catastrophe

or poor performance at Foxconn’s manufacturing site, a tightly focused
supply chain would suffer a more devastating impact than if production
resources were more dispersed.
Facing The Viking Challenge – Incentives in
the Partnerships
Both Cisco and Foxconn had strong motivations to work together
closely.
Leticia Jensen noted, “The key mitigation to risk in this project was the
fact that both partners had a lot at stake.
We had very strong incentives to succeed together.”

the Partnerships
For Cisco, engaging closely with the contract manufacturer would help
ensure the success of a key new router platform and reduce the risks of
using Foxconn in the untested situation of making a sophisticated
carrier-class product.
In addition, Cisco had long-term incentive to develop the contract

manufacturer.
If Foxconn performed well in making a high-end router in a low-cost

manufacturing environment, Cisco would have more flexibility for future
products.
the Partnerships
For Foxconn, a major incentive was the chance to prove itself on a
more complex product, thus opening up new business opportunities.
Foxconn’s Sin said, “We always wanted to move to high-end products.
High-end products have more value-added content and better market

diversification.
If we perform well, it is pretty obvious that we will get more projects.”

(For Foxconn’s perspective, see Appendix 1.)
the Partnerships
The two companies also had contractual provisions and built-in
processes to help manage their relationship and the risks to both sides.
Contracts between Cisco and its contract manufacturers typically

established cost targets, quality goals, pricing, materials markups and
timelines.
But they were reviewed quarterly, with the possibility of terms being
renegotiated. In addition, Cisco did quarterly evaluations of its contract
manufacturers and their individual production sites, informing them of
how they scored relative to other contract manufacturers.
Facing The Viking Challenge – Ensuring
Success
Cisco’s NPI process required that steps be taken to reduce the risks
inherent in new product development.
Because Viking was a major and complex program, the team knew that
“risk mitigation” measures would be more crucial than ever.
Cisco engaged its Viking supply chain partners early on to help simplify
product design and manufacturing processes.
The goal was to maximize the ease and efficiency of manufacturing, a

concept known as “design for manufacturability.”
Success
Early on, for example, it contacted suppliers of the 65 custom parts that
would be in the router, and incorporated their feedback when building
prototypes.
Compared with the 7600 router, the Viking machine had 16 percent
fewer mechanical parts, a greater amount of reusable hardware, and
shorter assembly times.
Getting Foxconn closely involved early in development was another

way of lowering risk and ensuring success.
Success
In early 2008, Foxconn sent three of its Chinese engineers to work
alongside Cisco engineers in the laboratories of Cisco’s Building 20.
They stayed for several months working on early prototypes.
Leticia Jensen said, “The Foxconn engineering team was involved in

bringing up proto units, setting up test equipment, helping with debug,
and putting together documentation very early on.”
Success
Westhauser added, “We never before had our partner’s engineers with
cubicles in the Cisco design buildings and benches in the laboratories.
They've been right there with us and hands-on as equal team

members.”
Later on, Cisco sent eight engineers to Shenzhen as consultants, but

with Foxconn engineers responsible in early manufacturing, according
to Westhauser.
Success
The two companies cooperated intensely on prototyping in order to
reduce the risk of technical failure.
The first two batches of Viking prototypes were built at Foxconn’s

prototype facility in San Jose, only a few miles away from Cisco
headquarters.
The third batch was made in Shenzhen, marking an unusually early

shift from lab to manufacturing site.
Success
The goal was to transfer knowledge and experience as early as
possible from Cisco’s engineering labs in San Jose to Foxconn’s
manufacturing floor in Shenzhen.
Cisco and Foxconn adhered to a strict schedule for developing,

producing and testing prototypes.
More than 100 prototype chassis were built and about 1,000 prototype
line cards were produced. The unusually large number of prototypes
was needed for distribution among the far-flung software and hardware
engineering teams that were working on Viking.
Facing The Viking Challenge – Collaborating
with Technology
The Viking project represented a major exercise in global, cross-
functional teamwork, supported by a host of technology tools.
More than 300 Cisco employees worked on the Viking project,

estimated Hosakote.
Just within the engineering organization, five sites were involved: San
Jose; Research Triangle Park, North Carolina; Petaluma, California;
Boxborough, Massachusetts; and Bangalore, India.
with Technology
To keep development going on a compressed schedule, prototype
hardware had to be shipped continually by Federal Express among the
sites for engineers to work on.
The “sheer logistics” were immense, said Hosakote.
In addition, there were four different sites working on software for the
router, presenting a challenge of getting “culture aligned,” he said.
For instance, an engineer in Massachusetts might be called on to

influence and motivate developers in North Carolina—via Cisco’s
WebEx webconferencing or its TelePresence teleconferencing system.
with Technology
Hosakote commented, “We needed to listen to each other and help
each other no matter where we were.
How do you get people in San Jose helping someone in North Carolina
that might be falling behind?
The project gets done no matter where you are.”
Viking team members used Cisco’s NPI Metrics, a web-based tool that
provided a single view of timelines and tasks for the far-flung team.
with Technology
Once a week, project managers in San Jose ran meetings with multiple
Cisco sites participating by dialing in or videoconferencing.
In the last three months before product launch, the Viking team used
the difference in time zones across the globe to keep the work going
24/7 among seven locations—California, Massachusetts, North
Carolina, Bangalore, Shanghai, Malaysia and Shenzhen.
They held dial-in meetings three times a day, at 8 a.m., 2 p.m., and 8
p.m. California time, so that multiple sites could join in.
with Technology
They used a “wiki” web site to share and update information and to
hand off work daily between California and China. Westhauser
commented, “This was probably the most globally developed product
that Cisco has ever had.”
Technology links between Cisco and Foxconn also helped ensure

smooth collaboration.
For instance, Cisco engineers could log in remotely to Foxconn in San

Jose or Shenzhen to test, diagnose and troubleshoot prototype boards
and routers.
with Technology
The contract manufacturer had access to Cisco’s ERP, or enterprise
resources planning, software system and to its software tools for
creating the bills of materials needed in procurement.
Facing The Viking Challenge – Marketing
Decisions
The Viking team faced a series of decisions on how to position and
launch the product.
Overall, the router had to be competitive on two fronts—the cost per

port needed to be low and the machine needed to be upgradeable to a
high density of ports on each circuit card.
Early on, Cisco set targets for both. In addition, Cisco benchmarked
against prices competitors were charging and looked at historical
pricing data on its previous routers.
Decisions
Cisco also wanted to ensure that the ASR 9000’s pricing would not
disrupt the company’s preexisting business selling the 7600 router.
Gibbs explained, “We have a very large, established business with the
(7600) platform.
We wanted to price at parity with the 7600.
That gave us a well-established price per port target.”

Decisions
Viking was a major new platform, so Cisco was determined to make a
broad splash with the marketing launch.
A key goal was to portray the router as a next-generation edge router

platform that could carry service providers through the next 10 to 15
years.
Cisco orchestrated the product launch in a major new way—by using

the latest Web 2.0 technologies and pitching the service provider router
like a consumer product.
Decisions
To build buzz in the months before the launch, the company created a
fictitious blog featuring humorous video clips of a bumbling tech
reporter, Ira Pumfkin, chasing down Cisco executives, including CEO
Chambers, to ferret out details of a big, upcoming product
announcement.
There was a downloadable video showing off the router’s speed and
power.
Developed jointly with ad agency Ogilvy & Mather, the video evoked a
television ad for a luxury sports car.
Decisions
Both the reporter’s blog and marketing video were posted on YouTube.
To market as well as entertain customers, Cisco created an online

video game about piloting a rocket ship while managing data packets
on a high-speed network.
In addition, the company briefed 155 industry analysts and reached out
to business and trade journalists, resulting in more than 100 articles
following the ASR 9000 launch on November 11, 2008.
Decisions
Commenting on the new-style marketing launch, Gibbs said, “We
realized there are a lot of different ways people get their information
now.
The Viking introduction took it to a whole new level.
This was significantly more complex (than past product introductions).”

Conclusion
In late 2007, as the Viking team looked ahead to launching the major
new router, it hoped to set a new bar for sophistication in Cisco’s
product development efforts.
The Viking program would test Cisco’s NPI expertise as never before.
Cisco had to launch the product extremely quickly.
It wanted to break ground by outsourcing manufacturing of a high-end

product to a low-cost foreign site from Day One.
Conclusion
It proposed to elevate its contract manufacturer, Foxconn, to a new
level of product and process complexity.
The Viking program would require an unprecedented amount of global,

cross-functional collaboration within Cisco and with its business
partners.
How could Cisco best meet these challenges?

Cisco - Collaborating On New Product Introduction

Uploaded by

Copyright:

Available Formats

Cisco - Collaborating On New Product Introduction

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cisco - Collaborating On New Product Introduction

Uploaded by

Copyright:

Available Formats

Management Information

The project’s code name, Viking, said it all.

The meeting represented a culmination of several years of

Just months earlier, in mid-2007, Cisco overhauled the project by

That day in November, the Viking team was seeking an “execution

If it got the go-ahead, Cisco would be ready to commit the resources to

But the Cisco team knew it faced many challenges.

This differed from Cisco’s past practice of outsourcing a complex new

Third, Cisco proposed to use one of its contract manufacturers,

Could Foxconn handle the technical complexity while producing the

Finally, Viking would test Cisco’s well-honed new product introduction,

In the 1990s, Cisco emerged as a leader in networking technology for

It went public in February 1990.

Revenue increased at a compound annual rate of 13.1 percent

It made a profit of $8.05 billion on revenue of $39.5 billion in fiscal

Business units belonged to a half dozen broader groups such as

In addition, business units had access to company-wide resources

In effect, the company was organized both vertically and horizontally—

Cisco believed that outsourcing enabled it to tap the most cost-effective

Angel Mendez, Cisco’s senior vice president of worldwide

That combination is an interesting thing.”

It cut the number of contract manufacturers from 13 to 4 major ones by

This enabled it to leverage spending while working more closely with

Similarly, Cisco sharply reduced the number of vendors in its extended

Many of Cisco’s manufacturing and supply chain improvements were

In early 2006, the company began shifting formally to a manufacturing

The lean process, developed with Cisco’s contract manufacturers and

Explaining Cisco Lean, Mendez said: Traditional manufacturing

In other words, a company builds product based on what it forecasts

Lean manufacturing is a “pull” model, which means that product is not

This is also known in the industry as “just in time” manufacturing. The

In addition, according to Mendez, Cisco used information networks to

The company had a well-oiled machine for new product introduction, or

The effort was led by a business unit’s product development teams,

At the outset, marketing and engineering would talk to customers and

Then engineering and manufacturing would work together to ensure

Manufacturing specialists would ensure that the supply chain

The process included a series of checkpoints, or “gateways,” between

This checkpoint ensured that a cross-functional team had approved a

Once a Project cleared the “concept commit” checkpoint, the design

By the end of the first stage, designers would have developed a

The project then entered the execution phase in which engineering

Manufacturing would do a thorough “design for manufacturability”

A final “technical readiness review,” or TRR, would go through results of

After prototyping, there would be a “pilot” build of the product at the

Following the orderability gateway, the product would be added to

This checkpoint, which included analysis on production yields and

Finally, there was a “post-production assessment” to identify and

By fall 2007, Cisco completely redesigned the proposed router for

the ASR 9000 router, or Aggregration Services Router.

The device was aimed at easing bottlenecks at the network “edge,”

Cisco expected the edge router market to grow significantly between

The company predicted that global Internet Protocol-based traffic would

As a result, bandwidth demand would reach approximately 522

Broadband carriers needed infrastructure for what Cisco called “visual

They needed equipment designed specifically to handle the demands

Ron Westhauser, senior director of product operations at Cisco, noted,

Service providers wanted high speeds and high capacity as well as a

There would be two “chassis” choices—a box 93.35 centimeters (36.75