In a recent talk to executives from MIT's Sloan Fellows program, HP Labs Director Dick Lampman describes a new type of research that includes working directly with customers and focusing on business impact.
Lampman, who is also senior vice president for research at HP, says he believes that industrial research should be at the intersection of science and technology, focused on solving real problems.
Read the full text below or download the transcript in PDF format.
» Contributing to strategy and other roles
» The big challenges
»
Reinventing the economics of IT
» Printing and imaging growth
» Solutions for telecommunications
» Transforming services
» Looking at "what's next"
» Collaborating with customers
» Post-boom research funding
» Successful technology transfer
HP has from its very beginning been a company defined by innovation.
Change has been good to us. Imagine the original Hewlett-Packard
company in 1939 – it was a very specialized company selling
electronic test equipment to a relatively small audience – and
think about how it's grown, along with the whole electronics
and high-tech industries, to what we have today, a Fortune
11 company.
For FY05 we reported $86 billion in revenue and we generated
$7 billion worth of organic growth. That means not counting
mergers and acquisitions, just growing the core. We have about
150,000 employees doing business in 170 countries directly,
and we do business in other countries indirectly through distributors.
HP spends about $3.5 billion a year on research and development.
That’s about five percent of our revenue. Of that $3.5
billion, roughly five percent goes to research; it’s
reserved to look ahead of where the businesses are.
HP Labs was founded by Bill Hewlett and Dave Packard, who knew
that thinking about what customers might need in the future
can be very difficult to do in a business. The pressure on
any business is always about allocating scarce resources. The
firestorm of competition that every business manages is something
that never changes.
Because of that, Hewlett and Packard chose to reserve a certain
amount of the R&D budget solely for looking ahead of the
market, looking at emerging technologies, at emerging opportunities,
at industry changes, and at changing customer needs. At HP
Labs, we wake up every day all around the world thinking about
those types of questions.
Now, a bit about our mission. Our number one role
is contributing to the strategy of the company. Most people
don’t naturally think that’s what research
does but, as a high-tech company, having a forward-looking
view of technology is critical to setting the direction
of the company.
We engage at every level in the company, from the key
technical staff all the way up to senior
management.
That engagement is bi-directional. We’re trying to
understand what’s going on in the business, what’s
going on in the customer base and what’s going on
in the marketplace, along with understanding where new
technology
is emerging and where we can drive new technology to change
things.
In some sense, we're trying
to create a new type of research and a new type of research
person, which is somebody who is expert in their field
but who also understands what matters, what technologies
are likely to have important impact. That's quite different
from the traditional model for research, where it is
science led. Our belief is that it has to be an intersection
of science and technology, focused on solving real problems.
Another key role for HP Labs is to help HP win in its existing
businesses. We’re always trying to create new value
for customers. If we can do that with new technologies,
we believe that will create value for the company.
A third role focuses on going beyond the existing company
strategy. We're constantly thinking about where the company
is going to have new opportunities. Where can we take the assets
we have as a company and move into new areas?
HP started out building electronic test equipment to
be used by electrical engineers in manufacturing and
engineering. Now we're providing systems for multimedia,
telecommunications and high-security banking applications,
to name a few. We didn’t get there by just staring
at the same customers and the same technology all the
time. In fact, the history of HP is really one where
we’ve been very forward-thinking in recognizing new opportunities.
So HP today, and clearly after the Agilent split, is
an IT company.
We moved into IT because people realized that computing
technology was becoming cheap enough that it could be
used to control HP's test and measurement instruments.
Not long after that, the people who were working on the
computers thought, "gee, there’s a lot of
other things we could do."
And here we are a few billion dollars later. We’ve
gone through that pattern over and over again. Many times
that movement has been led by the Labs. HP’s first
computer architecture came out of HP Labs. All three
HP printing technologies – laser, inkjet or our
new commercial printing technology – were either
led by Labs or jointly developed with the businesses.
So across the board, HP Labs has helped drive a lot of
that change.
Finally, one of our roles encompasses what everybody
expects to find in a research lab. We have deep investments
in science and technology, which are very much like a cutting-edge
academic program, really at the forefront of the field.
We’ve picked a few areas where we wanted to have
people inside HP whose expertise was equal to that of anybody
in the world because we believed that competence in these
areas could be leveraged into areas that are important
for HP. » Back to top
Earlier, I talked about strategy in terms of the company.
What we’ve done in HP Labs is apply our thinking
about what’s going on in the world – whether
it’s in technology, in markets, or in the industries
in which we operate – to devise a strategy. It may
not be the strategy you think of when you think about a
research lab, because a lot of times people talk about
pervasive computing or large-scale computing or maybe specific
technical disciplines like artificial intelligence. The
way we formed our strategy was thinking about what the
big challenges are going forward that are important to
HP.
I’ll start with strategy number one: We call it Reinventing the Economics
of IT. We chose that because our belief was that the IT
industry (and that includes HP) was going through
a transformation. During the 1970s, 1980s and the early
1990s, there was a huge push in terms of price performance.
People got excited when you brought out a new microprocessor.
We saw a couple things going on. One of these was a move toward industry-standard building
blocks. From a customer’s standpoint, this was
great because you’re seeing lower prices and you’re
seeing more standardization, which makes it easier to
manage IT.
From our point of view that we could reduce our R&D spending in
that area and shift it into other areas that were becoming
more important to our customers. At the same time, our
customers realized that most of their spending was not
on ‘boxes’ or servers – it was on the
people and the software to run all this stuff. Specifically,
the cost that keeps growing is people.
» Back to top
So in the late 1990s, we shifted our focus to
looking at how we could start attacking what is truly the
major cost of IT for our customers and, at the same time,
make IT more valuable to customers. The sheer complexity
of IT, certainly in the enterprise, but even for consumers,
is becoming a bigger problem than the technology itself.
Our research program involves working on our next-generation
data center. It's all about management software – software
to help automate IT operations and take a lot of the labor
out of IT, and, in the process, make it more robust
and reliable. Most of the outages you see in IT today are
actually caused by people. It's not because people aren’t
diligent. It just that the sheer complexity of IT means
there are so many potential ways to cause problems.
This is also reflected in the difficulty of delivering IT
projects. The complexity these projects
means they proceed slowly because there’s
a tremendous amount of testing required.
We’re attacking those issues in a very broad way.
We’re working with some of HP’s largest customers
to get a detailed understanding of what problems we need
to solve, and we're working with HP's internal IT. Although
we have delivered some technology, this is a decade-long
project to get to that ‘lights out, 24/7 data
center’ which runs reliably with a small number of
people.
When we started this program, one of our team made a joke
that our goal was to get to a data center with one person
and one dog in it. And the reason for that is the dog guards
the data center to keep people out. The person’s job
is to keep feeding the dog. While I don't expect we will
ever see that outcome, that vision encourages the team to
think boldly.
Part of that vision involves storage. Storage has become the fastest-growing component of IT. It has all the same issues in terms of complexity as management. How do we automate it? How do we build models? How do we virtualize things? Instead of managing individual computers, we're now managing whole data centers as a single object. It's a tremendous push and one in which we've had a lot of success.
There's also a grid element of this, because the grid, while
originally a scientific computing idea, is migrating
into the idea of how do you tie multiple data centers together
using grid-like technologies. And so we’ve been working
with the grid community to bring commercial-scale management
practices to the grid so that it could be used to manage
multiple data centers.
» Back to top
Another key research priority involves printing and
imaging growth. I’m sure you’re all aware that
HP is very successful in the printing business. Our success
has been built on the desktop – in the office and at
home. But years
ago we were looking at the total print opportunity for HP
and realized that, for all of our success, we
were addressing only four percent of the available print
market.
So we started looking at areas where we could bring the same
sort of digital revolution that we brought to the desktop
into other segments of printing. Specifically, we looked
at commercial printing – typically, the kind of printing
you’d see in a magazine or catalog. It's high-quality
printing with a high visual impact, but it is currently
limited because it is a mass-production technology. Basically,
you cannot change information on a per-page basis.
We saw an opportunity with something called liquid electrophotography,
which is a bit of hybrid technology. It uses a laser writing
system like a laser printer, but it uses real ink, or actually
a very special kind of ink, so you get the kind of visual
quality you expect in commercial printing. HP is building
a very large business around this technology now. We have
a long-term program of continuing to make that technology
available at higher performance and lower cost.
» Back to top
A third key area we're working in is delivering solutions for specific industries. Our largest single area is mobile rich media. We’ve been working in telecommunications for about 15 years, originally focused on how to manage the wired voice telecom network.
The telecom network has been through several big transformations. Originally it was, basically, wires connecting your phone to somebody else through a switch. Then we had the intelligent network where there was a second channel alongside your voice, which essentially is a digital computer network that is controlling what’s going on. Capabilities like toll-free numbers, credit-card calls and caller ID came out of the creation of this parallel network.
HP is at the core of this telecom network. The majority of the world’s phone calls go through systems that run our software to help manage the network.
» Back to top
The next big transformation was in mobility. Most of that development,
in terms of digital mobile phones, took place outside the
United States. So our labs in Europe and Japan took a lead role in helping develop technologies to prepare
us for that big shift to digital, which is just now starting
to happen in the U.S., but occurred in Western Europe and
Asia-Pacific first.
The next step is moving from digital mobile to digital mobile
with media streaming and interactive services. Our team has
played a lead role in the company in creating the technology
that’s required to do that. We've worked with some
of the largest mobile operators in the world to understand
what drives them so we could develop technology for
their future needs.
This is, again, a new way of working in research because
in these complex areas you can’t just sit in your office
and look at it from a science standpoint. You can’t
just look at what’s going on in the market today.
You have to get beyond that and find the people who are reshaping
industries and get inside those organizations, get inside
their thinking, so you can understand what they’re
going to need in the future.
There’s an adjacency to the telecommunications industry
in terms of media and entertainment. Digital multimedia is
becoming a huge industry. All of it, whether it's radio, television or film, is being transformed by digital
technology.
We're looking at the future structure of that industry so
we can build the IT infrastructure they’re going to
need. In digital entertainment, we participate on both sides.
We make many of the devices people use to consume digital
entertainment. But at the same time, we're working with people
who own and create content, and deliver that content to consumers.
» Back to top
Our fourth big research area is in services, which are becoming
a much larger part of the IT industry. A lot of our customers
are asking HP to take responsibility for certain parts of
their operations.
Traditionally, people have thought about services as a sort
of 'pipe,' a means by which we deliver technology
to the customer. About five or six years ago we began thinking
that maybe we could make that pipe better. Our model was
based on considering what happened in manufacturing in the 1980s. Manufacturing used to be about labor, capital
and material. Today if you talk about manufacturing without
talking about IT, and the ability to control inventory and
just-in-time materials, people would think you were living
in the stone age.
So our question was, how could we apply that same intellectual
discipline to the services practice, shifting it from something
that today is driven by labor and labor cost – labor
arbitrage is the hot topic right now – and start thinking
about how technology could make labor more efficient and
change that whole equation.
Over the past five or six years we’ve started to
find powerful ways to impact the services business. Using
technology to transform the services business has become
core to HP’s strategy. It's something our team is very
excited about. We now have a wide range of engagements that
look at the whole business process from end to end, at how
services are acquired and delivered.
» Back to top
Our fifth strategy is the not-strategy. It’s really
about what’s next. In this area, we're looking at devices,
software and markets. In devices, our most prominent work
is in nanotechnology – specifically, molecular electronics.
When we started this program about ten years ago, it was
pretty clear to us that what has really been the foundation
for the whole IT industry will end – that is, Moore’s
Law does expire. We are reaching those limits probably sometime
in the next ten years.
We are working very aggressively in that area. I believe
we have one of the strongest programs in the world.
In software, we are looking at questions about how Web technologies
are going to evolve. The Web has evolved from the basic html
or http technologies to XML and higher-level description
languages and Web-based services.
The problem we’ve been focused on in recent years is
really something called the Semantic Web, which recognizes
that eventually all of these things have to connect. It's
really an unaddressed problem. We have a very strong team,
one of the leaders in the world, working on that.
In the area of markets, we're considering how the IT industry
is going to grow from a geographic perspective. Specifically,
we're looking at emerging economies around the world.
Sometimes we describe that program as aimed at HP’s
next billion customers. We currently serve an available market
of a little under a billion customers. But then the potential
market opportunity falls off dramatically because
of issues around the affordability of technology.
In many of these regions, people's incomes are rising considerably. Meanwhile, the cost of the technology we deliver
continues to go down. The combination of those forces means
we have many more opportunities around the world.
Our question was, are there non-financial obstacles that
will limit the adoption of IT? For example, one issue is
language. Once you get beyond the core languages the IT industry
addresses, there are many more languages in the world
that aren't addressed at all. There's also the problem of
infrastructure – or the lack of it – in many
developing countries.
In some cases, even ownership models are worth exploring.
We tend to think about individual ownership, either on your
desk at work or personal ownership. These are models that
are not going to work in large parts of the world.
We have a program driven out of our lab in India that’s looking at questions like these. In fact, we
just announced a keyboard we developed for phonetically
based languages like Arabic and many Indian languages, which
are quite difficult to work through a standard keyboard.
We've come up with a simple, low-cost way to do data entry
into those languages.
» Back to top
Now that I've discussed our overall strategy, let me tell
you more about our work on next-generation data centers.
We're looking at automation and virtualization technologies
to reduce the number of people required to support IT. Some
of the impetus for this comes from our experience in the
telecom industry. That industry is more mature in terms of
automation; a very complex system can be run by a few people.
More automation also means fewer errors.
Another key element of the next-generation data center is
security. More centralized IT potentially creates more vulnerability.
How do we manage that? HP is going beyond the traditional
security model in which people are creating viruses and worms
on one side, and IT companies are constantly trying to catch
up, and customers are constantly having to apply updates
and patches. Our approach is proactive security, where the
security system has more of a biological model. Your body
can recognize diseases it’s never seen. We’ve
been working on an IT security system that can sense the
symptoms of an invasion of a new type of worm or virus and
locally contain that, which gives the human operators time
to respond.
Another area I want to tell you more about is our work in
mobility. One of the problems we’re investigating right
now is this convergence question – looking at an architecture
that allows us to unify and interoperate across the now-separate
worlds of mobile telephones, the Internet and videoconferencing. » Back to top
In commercial printing, what we're doing
is no less revolutionary than what we did on the desktop.
Today, people think the minimum production run is thousands
of copies of a document or booklet or magazine. With our
technology, we can make every single page different at no
extra cost.
That's in part due to the digital-press technology itself.
But a lot of what we’ve been doing is working with
graphic artists and designers, the people who do page layouts
and designs, learning the rules they use so that we can develop
an automated workflow.
Imagine a news magazine you receive at home customized
to what you’re interested in, with maybe a few
random things that you might want to get exposed to. There’s
a lot of turmoil in newspapers, magazines and advertising
right now, much of it driven by the Internet. The very same
technology that's causing a disruption in those industries
could actually create a whole new publishing model.
» Back to top
It’s clear to us that the visual impact of
paper is still quite unique. We continue to look at alternatives,
but as long as we're using paper – and we think that’s
going to go on indefinitely – the question is,
how do you transform printing with digital technology the
same way we’ve transformed the office? That's what
we’re setting out to do with the Indigo presses.
At the same time, we’re investigating whether we can
build an alternative that has the kind of visual impact and
portability and robustness of paper. In this area, we've
developed a plastic display. The kind of technology it takes
to manufacture this is not what you see in today’s
computer flat panel screens or large-scale TVs, which have
precision glass and electronic assemblies.
This display is literally built from pieces of plastic that
can be made on something that looks like a printing press.
It’s a startling technology. We still have a ways to
go, but our thinking is that if something is ever going to
replace one of our businesses, we’d like it to be done
by us. » Back to top
Earlier I mentioned our work in nanotechnology. Electronics
have shrunk from their first form, the first computers, by
five orders of magnitude. We think there’s that kind
of room to go, where computers could be as small as grains
of sand or dust. We’ve gotten a lot of recognition
for our work in this area from places like MIT and Scientific
American.
We're also looking at the question of what happens when you shrink to
the dimensions where you can see the individual atoms, which
is where we are. You’re not going to be able to think
about developing integrated circuits in the same way you think today. At that scale, computing has to operate like other
things in nature, which means it will have to tolerate errors.
We started our work based on the idea that, unlike traditional
electronics where the focus is always about being perfect,
this has to be more like something like DNA, or other life
processes, which has the ability to correct itself. And so
we talk about coding theory as ways to build things that
can suffer statistical errors, random errors and still function
perfectly. We’ve demonstrated that.
Another piece of what we do in HP Labs involves our work
in open source and our collaborations with universities.
One collaboration we've been involved in is a program we refer to as Gelato. In Gelato, we are
developing scientific computing applications for Linux on
the new Itanium chip, which we co-developed with Intel. HP
provided the leadership for it, but Gelato really is operating
as an open-source community.
The design objective for that chip was not just to
create an industry-standard chip, but to create one that replaces what’s done today with proprietary
RISC chips. This chip was designed, from the beginning, for
high availability and ultra-high performance. We’re
just at the beginning of that lifecycle, and many universities
have signed on to help us drive that forward.
» Back to top
One of the things I talked about earlier was how we’re
working with customers. In the traditional model, HP Labs
worked with the product businesses of HP to either sell an
idea or to listen to a problem they had. That idea
wored its way through the R&D community and eventually
got to the business managers. Ultimately
the technology became something HP sold, ofte in the form of
something delivered through services to our customers.
That model worked for most of the last 40 years. In the last
five years, we started looking more at software businesses,
where the pace of change is much more rapid than the old
serial process can accommodate.
The model we’ve been using increasingly over the last
couple years is working directly with our lead customers.
Earlier I talked about trying to understand our customer
base, what companies are leading their industries. Who are
the people that are really the cutting edge?
Doing this is beneficial for our customers because they get
early access to technology. But it’s beneficial to
us, too, because we get early insight into what that industry
is going to be needing going forward. Frequently
we then develop a ‘first-of-a-kind’ system. We'll work
with the HP Services organization to deliver this unique
offering.
If we’re right, we’ll do it more than once and
we’ll have repeat business. And if we are really right,
eventually our solution will find its way into a standard
product. We’ve gone down this path a few times in the
last few years. Increasingly, we believe that in fast-moving
areas, this is a much better way to develop technology.
That’s the end of the formal remarks. At this
point I’d be happy to take questions about anything.
» Back to top
How has the recent change in leadership affected the vision
and the strategy of the company in general and the labs?
In terms of the labs, our strategy has been in place
for about six years and actually has not changed. What
has changed
is that our top two programs – future data centers
and commercial printing – have become the
core for HP’s two big growth initiatives.
So we made some good bets on behalf of the company, but that’s
our job – to try and look ahead.
In terms of CEO Mark Hurd, he's put priority on two things. One of these does
not affect HP Labs directly, and that has to do with coverage – meaning,
do we have enough people out in the world talking to customers about what we
have to sell? The other big priority for Mark is R&D. He understands that
R&D is how a company creates new value, creates things that customers care
about. » Back to top
HP's top management has been under tremendous pressure in
the marketplace to improve near-term performance. That
apparently would translate into the policies and actions
of the CEO. I wondered whether you sense a friction or difference
of opinion or vision in terms of allocating strategic funds
into long-term innovations which may not translate into near-term
performance, either top line or bottom line. And then I wonder
whether you enjoy a certain level of autonomy so that you
could probably manage this huge budget in a spectrum fashion
or a portfolio fashion so that it will sustain the company
going forward rather than respond to the relatively short
manageable horizon of the CEO or top management?
» Back to top
In the past five years since the bubble economy vanished
and the IT industry has been under all this pressure, it
would be very natural for the resource balance to shift more
toward short-term R&D and away from something like HP
Labs. We actually track that, looking at HP Labs as a percentage
of overall R&D.
If the company really wanted to shift to a shorter-term focus, our percentage
would go down and the rest of R&D would go up. But, in fact, that number
has tracked within a tenth of a percentage point over the last six years. There is certainly a natural pressure,
but it’s one that HP has resisted. Mark (Hurd) certainly continues on that
path; he’s a very strong supporter.
As for the second component, which is the level of independence, the way research is managed in different companies is quite different. Sometimes research
receives negotiated funding from individual business units.
At HP Labs, we get one check every year and we’re told to spend it wisely. When we allocate
funds, we have a tremendous engagement with the company trying to understand
the best way to do that, but we’re held responsible for making
the decisions about allocation. So we retain quite a bit of independence. Obviously
we don’t ignore what’s going on in the company, but ultimately we’re
held responsible for the decision.
» Back to top
I want to get a sense of how you think about failure in
R&D. What is your overall reaction to failure? Is there
an acceptable level of failure? And do you quantify that
in any sense over the course of the year? » Back to top
As Mark Hurd likes to say, our job is to be out in front
of the road map. So failure is something we have to
deal with. The way we think about it is if you don’t
fail sometimes, you’re not trying very hard.
But it’s not failure in the sense most people
think about. It’s not just a question of whether researchers
can prove their idea works; it’s a question of whether
it’s important. Most of the time the ’failure’ is
actually a team redirecting itself.
We do a lot of reviews and it’s not a hands-off environment. But the recognition
that we’re on the wrong course usually comes from the teams.
We have a set of people in the Labs worldwide who are very much thinking about
both achieving technical success and having impact. They’re much closer
to that and see it much earlier than the management. That’s
where the culture of the organization and the strength of the team comes in;
it’s rare we fly right into the wall.
» Back to top
I'd imagine that there's a lot of challenges in taking a
one-of-a-kind solution and getting that into the more production-oriented
R&D teams. Could you tell us a little bit about the processes
and team structures you use to make that successful? » Back to top
The truth is, there is no standard process. It’s different
every time. In some cases, the right thing is for it to stay
in services where it tends to be a semi-custom offering.
Other times, if we’ve had a lot of success with it,
typically we’re also engaging with the potential business
host to help create a standard product.
More than process, the key to technology transfer is building the human side
of the equation. We work with people from the businesses at every level – the
business general manager, the R&D manager, the marketing manager. Occasionally
we have projects we want to move along more quickly at the corporate level, so
we’ll put in some incubation funding and scale up the business. That might
occur when there is still a pretty high risk level involved in the work.
The question in any business is, how do you balance top-down initiatives, which
allow you to move more quickly, with the more entrepreneurial, bottom-up activity. That’s one of the reasons I say these things
move in different pathways. Some of the work we’ve done in digital media got CEO-level recognition, so there was a big infusion of
money to jump-start that. Then it migrated into a business much more quickly.
In any real business, along with all of the organizational-chart considerations,
there's the individual leadership, trust and confidence you have to build
in the organization to make the innovation process work. If we wait until we’re
asked, or if businesses wait until it’s clear that there is a market, it’s
already too late because somebody else has occupied that position.
The thing that is not written about the innovation process is the importance
of the human fabric, and how you create a risk-taking culture – not foolish
risk, but measured risks, so you pursue new opportunities as they emerge
and not be late.
» Back to top
Many times when you ask companies whether they’re
being innovative enough, they’ll reply with the number
of patents they filed last year, and certainly HP has an
amazing patenting engine. But that doesn’t necessarily
translate into commercialization. As the director of the
labs here, how do you measure whether you’re being
innovative enough?
» Back to top
The patent numbers war is something I have mixed feelings
about because, at the end of the day, patents actually are
a very weak indicator. Whether you have seminal patents – things
that redefine a field – or incremental patents, both
count as one patent. We have a strict program for
monitoring what kind of patents we’re filing out of
the labs.
For us, the real measure is impact. We look at transformational things we’ve
done, things that have changed the industry and HP. You can’t measure that
by putting a dollar value on it. We can’t claim credit for a $20 billion
business based on something we did back in the 1980s. But most people in that
business would say that we wouldn’t be in it if we hadn’t done that
initial work, and if we hadn’t followed it up for a couple of decades of
continued refinement. Yet there are thousands of other people who deserve credit
as well, which is why we go for the qualitative measure rather than saying every
dollar there belongs to us.
We engage enough with the businesses so they can see what we’re doing.
Sometimes it’s obvious because there’s good alignment; sometimes
they’re skeptical. Sometimes we invest in things that are not viewed as
strategic by the company. But that’s part of our responsibility. Enough
of those have become strategic that we get a little bit of room to do that without
criticism.
I’ll give you the most spectacular example of that. The inkjet technology,
which powers a substantial part of HP, was viewed as counter-strategic within
HP. We got fairly intense criticism for working on it. At that time, the company
was betting on laser technology as the engine for digital printing, personal
digital printing and office printing. It turned out that thermal inkjet was a
good bet and, in the end, we found people in the businesses who shared that belief
and something very nice happened. » Back to top |