Written evidence submitted by Professor
Andy Hopper (TIC 27)
1 There are many ways to innovate ranging from
corporate research supporting internal business units to outsourcing
of research to universities. However over the last decade there
has been a reduction in the number of corporate research laboratories
and attempts at technology transfer from universities have not
made their mark considering the level of investment in "third
stream" funding.
2 Many efforts have been made to improve the
situation but it may be that little more can be done within existing
structures.
In industry it is often difficult for business units
to absorb innovative ideas and products. This is because big things
start small and large organisations are most effective once products
are at scale.
In academia there is pressure to publish which in
my view is a distraction to innovation, and peer review of research
grants is not a method of maximising economic impact. Most importantly
teaching has to be delivered and the training of well qualified
people is the most effective economic output for an academic.
3 Nevertheless knowledge and opportunities are
being generated for new markets and ventures. Our universities
are producing an excellent feedstock of well educated and talented
people. Investment in technology, engineering, and science is
the underlying wealth creator of our society. The global marketplace
is ready and waiting.
4 A gap exists for institutions whose context
and mode of operation is focused on economic output. If this "institutional
gap" is filled then the well known "funding gap"
is likely to be filled as well. To do this the new institutions
have to produce fundable or acquirable businesses. For the most
successful outcomes this can only be done if they are able to
attract the most talented people.
5 The TICs present an opportunity to do something
new but the positioning and scope of these centres is crucial
to success. An important consideration will be the ability to
attract top talent.
One positioning would be as middle ground entities
integrating the work of industry with that of universities. This
would tend to emphasise a service mode for the university and
a customer mode for the industrial partner. To proceed in this
way would require more realism than exists today of the value
different stakeholders bring to commercial ventures.
However another positioning would be for industry
and universities to fuel a TIC which aims to produce really new
businesses based on very innovative work carried out within the
TIC. For the industrial participants this disruptive approach
presents a threat and an opportunity. The threat is that the output
may undermine the industrial business model but the opportunity
is to be aware of it and use it. For universities this positioning
is straightforward and based on supplying qualified employees,
a very wide intellectual context, and only occasionally intellectual
property. This in turn is best achieved by academics conducting
world-class research judged in its own terms.
6 Below is a description of the Olivetti Research
Laboratory (ORL), an industrial research lab which existed in
Cambridge for a period of 16 years and innovated in the ICT sector.
I hope the ORL experience provides a useful template for how a
TIC might be structured. The model may not generalise or translate
directly to other sectors, but does present a specific example
from which lessons can be learned.
ORL was a major wealth creator in the UK, a source
of inward investment, and one of its legacies is the continuing
existence and evolution of numerous businesses which have attracted
investment of over £100M.
7 ORL was set up by Olivetti in 1986 as a subsidiary
company based in Cambridge with a charter of changing the world
of computing. Over its existence ORL was initially funded by Olivetti,
subsequently by Olivetti and Digital Equipment Corporation, then
by Oracle and Olivetti, and finally it was bought and funded by
AT&T. The approach was to keep the number of sponsors small.
8 ORL had a symbiotic relationship with universities,
funded a large number of PhD students, and sponsored research.
However it was not a downstream organisation and used universities
for recruitment and to reach more distant parts of the worldwide
research community. The specific research conducted within ORL
was equal to if not more significant than that carried out at
any university. Sponsored university research was expected to
be world beating in its own terms.
9 The leadership of ORL chose the research directions
with full backing of the most senior people in the sponsoring
organisations. It was agreed from the outset that strategic and
operational decisions would be made in the local context and not
by head office. In other words the model was to back leaders on
a sustained basis.
10 The research directions were chosen to take
into account technology, business models, and the availability
of venture, angel, and corporate investment capital. Direct links
were maintained to all these sectors. The contractual arrangements
with sponsors permitted full visibility of the global marketplace
within the sponsoring companies. Projects were targeted for impact
on timescales from 3-10 years. The innovation model was to buck
trends and construct real systems.
11 The strategic positioning of the laboratory
alongside but independent of a university was attractive to a
particular group of very talented people. These were individuals
primarily incentivised by real-world impact and in many cases
financial gain, rather than publication and teaching, or the corporate
world.
The laboratory employed up to about 50 people at
any one time. At that size it was possible to have little internal
hierarchy and while a command line was available it was used judiciously
so as to empower talented employees. Indeed 20% of time was available
for individuals to work on any project of their choosing (an approach
now used by Google). Thus barriers to contribution and innovation
were reduced as much as possible.
12 As projects showed promise additional personnel
were recruited with the intention of forming teams and businesses
which were fundable or acquirable. Thus the model was personnel
transfer rather than solely technology transfer.
13 The commercialisation model was to offer first
option on output to sponsoring companies but on the basis of "use
it or lose it". However "lose it" meant lose control
but maintain an appropriate share of subsequent success. Thus
commercialisation could not be held back.
14 A spin-out company was the next option in
the waterfall of commercialisation, followed by licensing, and
ultimately open sourcing. In this way commercial output was maximised,
and apparently unwanted output was kept alive with the possibility
of reuse as the industrial landscape changed.
15 Each spin-out consisted of the core team,
the seed funding, the initial business plan, and early customers.
Intellectual property was assigned exclusively to the new venture.
The core team always left the research lab and became founders
of the spin-out with appropriate stakes. This approach was also
well suited for potential transfer to a sponsor or acquisition
by another company. Thus the ORL model was to spin-out the team,
and the technology, and move on to a new research challenge.
16 In essence over a period of 3-5 years ORL
developed very strong groups around promising projects which were
disruptive in the market place. Individual incentives were constructed
to maximise output while the clear and commercial arrangement
with sponsors made for an admirable amount of flexibility and
support from them.
17 A number of companies were spun out. The most
successful Virata was in the DSL chip business and at its peak
had a NASDAQ market capitalisation of over $3Bn, $500M in cash,
and some 35% of the worldwide consumer DSL chip market.
18 The most successful open source output (VNC)
became a de-facto worldwide standard for remote control of computers
with by now several hundred million downloads.
19 In the long term ORL sponsors were offered
returns which outweighed the costs of running the laboratory.
Furthermore personal wealth was also created and recycled in to
the UK economy.
20 The laboratory was suddenly shut by AT&T
in 2002. An interesting observation can be made on what has happened
since then. Almost all the teams that constituted the ongoing
projects reformed as start-up companies. Where required these
companies have subsequently attracted investment capital (totalling
over £100M) and a number are doing very well. Indeed some
of these companies have themselves spawned new businesses thus
the cascade of innovation has continued.
21 The companies spun out directly from ORL were
Adaptive Broadband, IPV, and Virata. The phoenix of subsequent
companies includes Adventiq, Apasphere, Apoideas, Cambridge Broadband,
Camvine, Camvit, Cotares, CRFS, DisplayLink, Imense, Ndiyo, RealVNC,
Solarflare, and Ubisense.
I would argue that the success rate of these ventures
is much higher than for conventional start-ups and in particular
spin-outs from universities. This is sometimes called the "soft
company" model and is behind some of the UKs biggest successes
such as ARM plc.
22 "TICs" have come at a time when
governments around the world are fulfilling their role in promoting
and stimulating innovation. However not all are sophisticated
in their approach. I recommend that "TICs" should be
used for creating entities with forward-looking innovation models
and not for the creation of conventional "integration centres"
which cover middle ground between universities and industry. More
sophisticated innovation models which back people and not processes,
maximise incentives, and minimise barriers, are the most likely
to be successful. Such approaches are also best suited to the
UK mindset and UK strengths.
I declare an interest having been Managing Director
of ORL during the whole time of its existence. Hermann Hauser
was the other co-founder and on the Board. All views expressed
in this note are my own.
Andy Hopper CBE FIET FREng
FRS
Professor of Computer Technology and Head of Computer Laboratory,
University of Cambridge
Chairman RealVNC
Chairman Ubisense
1 December 2010
| 
