Information Management for the Intelligent Organization

Chapter 5: Environmental Scanning in Action

"The future was predictable; but hardly anyone predicted it."
(Attributed to Alan Kay, Apple Fellow)

"Therefore those who do not know the plans of competitors cannot prepare
alliances. Those who do not know the lay of the land cannot manoeuver their
forces. Those who do not use local guides cannot take advantage of the ground.
The military of an effective rulership must know all these things."

(Sun Tzu, ca. 500 BC, The Art of War)


Perspectives from Neurobiology

Environmental scanning is about seeing, perceiving, and making sense, and it is
only natural that we turn towards the most sophisticated perception system that we
know, the human visual system, to understand how nature has solved this rather
daunting problem. In a sense, environmental scanning is the visual and perception
system that feeds information about the outside world to the strategic planning and
management functions that constitute the brain of the organization. There are
many parallels between how the human visual system processes information from
external scenes and how environmental scanning processes information about the
external milieu. First and most basic, humans use the information from the visual
system in order to perform a number of tasks that are essential for survival, such as
foraging for food, avoiding dangers, planning movement from point to point, and
so on. For an organization, having timely, accurate information about external
change is just as critical to its survival in much the same ways: securing resources,
spotting threats and opportunities, setting long term plans, and so on. Second, the
visual information that enters our eyes is vast and ambiguous. We are easily
deceived by our own visual systems - each of us experience frequent errors in
perceiving objects or recognizing people, and ingenious optical puzzles have been
devised to show up how the visual system copes with inconsistent or incomplete
visual cues. In fact, relying on only any one aspect of visual information from the
eyes usually leads to ambiguity in trying to understand an external visual scene -
the brain needs to synthesize visual information on many features of the visual
scene (shape, color, motion, texture, etc) in order to come to an interpretation. For
an organization, the flood of information about the outside environment is so
marked by equivocality, inaccuracy, and incompleteness that settling on an
acceptable interpretation becomes an organization-wide challenge. Third, and this
follows logically from the last observation, seeing is the result of an active
construction process involving many levels and subsystems in the visual system.
The eye and retina do not function at all like a camera that passively registers the
effects of incoming light signals which are then decoded in the visual cortex, a view
that held sway until the mid-1970s. Instead, interpretation is an inseparable part of
seeing, and this interpretation is achieved through a complex division of labor in a
crosslinked network of discrete cortical areas and subregions that specialize in
particular visual functions. In the last chapter, we presented the symmetrical point
of view that organizations behave very much like interpretation systems, where
scanning/perception is intimately linked with interpretation as part of the
organization's overall learning process.

Over the last few years, we have made great leaps in our knowledge and
understanding of how mammalian visual systems function, including the
anatomical structures and neuronal processes that make up the human visual
system (Gazzaniga 1995, Koch and Davis 1994). Light entering the eye is focused by
the lens onto the retina. Nerve cells in the retina, called ganglion cells, receive
signals from other neurons in the retina and send signals to the brain. Each of the
retinal ganglion cells will fire most rapidly to a small spot of light in just the right
position but will only respond weakly to uniform illumination over that general
area. In this way, the retina is already processing information by sending to the brain
interesting parts of the visual field, where the light distribution is not uniform. The
early parts of the visual system are therefore highly parallel, and the visual
information is mainly relayed through a small part of the thalamus (the lateral
geniculate nucleus) to the visual cortex. Five visual areas have been found in the
visual cortex (V1 to V5), each specialized in function: the neurons in V1 and V2
operate together as a kind of post office that parcels out signals to the other visual
areas for further independent but parallel processing; the neurons in V3 specialize
in form; V4 in color; and V5 in visual motion. Altogether, four parallel but distinct
perceptual pathways connecting these five visual areas have been identified, one for
motion, one for color, and two for form (Zeki 1992, 1993), shown schematically in
Figure 5.1 below. The four perceptual pathways follow a similar sequence, signals
from the retina and geniculate nucleus are projected onto V1, which together with
V2, segregates and distributes the signals to the specialized visual areas for
independent, parallel processing. For example, in the motion pathway, the main
visual area is V5. Input originates from the retina, is relayed through the geniculate
nucleus, and projected on to V1. From here signals pass to V5, both directly and
through V2. Based on research of the macaque monkey, Crick (1994) suggests that it
is possible to represent the five visual areas and their connections in a quasi-
hierarchy of visual information processing. Starting with retinal ganglion cells at
the lowest level, signals are relayed through the lateral geniculate nucleus at the
next level, then projected onto the V1 level. V2 at the next level works with V1 to
distribute signals to V3, V4, and V5 at higher levels. As we move up the hierarchy,
the features to which the neurons respond become more complex, and the sizes of
the receptive fields increase, so that they cover a progressively larger part of the
visual field. The arrangement is quasi-hierarchical because connections can and do
skip levels, nearly all connections are reciprocal, and there are many cross-
connections between cortical areas at the same level.

Despite this extensive division of labor in the visual cortex, none of the
specialization and subdivision is evident at the perceptual level. How does the brain
produce a unified picture of the seen world? Anatomically, there is no master area
in the visual cortex to which all the other areas converge and where a unified image
might be generated. Instead, the specialized visual areas connect with each other,
either directly or indirectly through other areas, forming a massively parallel and
reciprocal network of links between the four parallel perceptual systems at every
level. Zeki (1993) hypothesizes that signal integration is achieved not in a single step
but at different stages, including the stages where functional specialization is first
established in the visual cortex (V1, V2). This multi-stage integration depends on
the ubiquitous re-entrant connections between the visual areas. Re-entrant
connections allow information to flow both ways between visual areas in a kind of
two-way feedback that resolves conflicts between differently specialized cells
responding to the same stimulus, and correlates neuronal events across different
visual areas without the need for a supervisory function (Edelman 1992). According
to this theory, the integrated visual image is an emergent property of the complex
network of activated cortical structures and neuronal connections:

integrated visual image in the brain is the product of the simultaneous
activity of several visual areas and pathways, including areas such as
V1 and V2 which receive their input from the LGN, distribute them to
the specialized visual areas and are re-entrantly linked with the latter.
As the re-entrant pathways are diffuse, non-modular and not easily
localizable, and the opportunities for integration are many and
integration itself is a multi-stage process, it follows that integration is a
diffuse and not easily localizable process. (Zeki 1993, p.334)

Fig. 5.1 Perceptual Pathways in the Human Visual System

In summary, the brain's visual perception system employs a number of information
processing strategies to turn the tide of incoming visual signals into understandable
visual scenes: functional specialization; parallel but quasi-hierarchical pathways;
and multi-stage integration through re-entrant connections. It is tempting to
consider to what extent these information processing strategies might be applied in
organizations to improve the effectiveness of environmental scanning.

Functional specialization in environmental scanning makes good sense in today's
environment where the interpretation of complex, ambiguous events and
developments necessitate the use of in-depth (tacit) knowledge. The tracking of
customer issues should involve the sales personnel, competitor issues the
marketing staff, technology issues the production and R&D departments, and so on.
Already some organizations have chosen to outsource the monitoring of critical
issues or sectors to outside specialists usually because there is no in-house expertise
to do so or because a more thorough analysis is desired. In the human visual
system, this functional specialization begins very early, in the first two visual areas
of the visual cortex. This suggests that specialized processing of the incoming
scanning information should begin as close to the source as possible in order to
grasp the incoming information in all its original richness and detail.

The human visual system is so rich in connections that it behaves both as a parallel
and as a hierarchical information processing system. Thus, the first visual area (V1)
separates and distributes signals on different attributes of the visual scene to the
other visual areas for further, independent processing that is done in parallel. At the
same time, information also flows up serial connections between visual areas so
that progressively more complex representations are built up. The implication for
environmental scanning seems to be that an organization can benefit by
simultaneoulsy balancing the advantages of the parallel and hierarchical processing
of scanning information. Parallel processing would tolerate a certain amount of
positive redundancy among groups in the organization analyzing scanning
information, and encourage them to produce multiple, competing hypotheses about
the incoming data. Hierarchical processing requires cross connections between these
groups as well as with upper-level management to attempt to progressively broaden
interpretations by combining insights from different groups.

The notion of multi-stage integration suggests that the brain has solved the problem
of integration in a unique way - there is no master area in the visual cortex, no
"Cartesian Theatre" (Dennett 1991), where a unified visual image is staged. Instead,
the brain has been able to gain flexibility and speed of response by distributing the
information integration process across multiple stages of parallel perceptual
pathways. The suggestion for environmental scanning could be that an organization
should not attempt to integrate its scanning information at a single locus, such as in
the mind of the president or chief executive. Integration and interpretation of
scanning information should be a shared activity, distributed across multiple levels
of the organization, and taking advantage of specialized knowledge wherever they
may be found in the organization. Where a high-level committee exists to translate
scanning information into plans, the group should include knowledgeable staff
from various functional areas and organizational levels, co-opt experts with special
insights, introduce outside opinions, and so on. Re-entrant connections are the
mechanism by which the brain achieves multi-stage integration. In organizations,
this suggests that the flow and sharing of scanning information should be two-way:
A sends a message to B, B processes the message and sends a return message back to
A. This form of bi-directional feedback and general information sharing could be
facilitated by the use of electronic bulletin board systems (see discussion on Usenet
newsgroups in Chapter 2), and the new generation of workgroup application
systems exemplified by products such as Lotus Notes. Overall, the ultimate goal for
organizational scanning is to emulate nature's trick for the biological visual system
by analogously mobilizing, networking, and synthesizing the information gathering
and processing capabilities of the organization to actively construct a unified
interpretation of the external environment to guide action and planning.

Environmental Scanning in Action in US Corporations

Are there real organizations practising scanning and interpretation in ways that
approach the design principles we have observed in the human visual system:
distributed information processing, functional specialization, and multi-stage
integration? Consider the experiences of firms like AT&T, Motorola, Kraft, General
Motors, and Eastman Kodak.

AT&T

...

Motorola

...

Fortune 500 Companies

...

Environmental Scanning in UK Corporations

...

Environmental Scanning in Swedish Corporations

...

Environmental Scanning in Japanese Corporations

...

Scanning for Future Learning

...

Close-up View of Five Canadian CEOs

So far, we have been describing the scanning systems and procedures of large
corporations and multinationals. Thanks to their high profiles in the media, and to
publicly available information from their former employees and consultants, we
have a fair idea (though not always the most up to date picture) of their business
intelligence systems. Although we have been discussing mostly scanning systems
and structures, it is individuals in the organization who gather the information and
use it in consequential ways. To better appreciate environmental scanning in action,
we need to see how individual managers and others have actually obtained and
used information. In this section, we move from macroscopic descriptions to a
close-up view of how Canadian CEOs do scanning on a daily basis. Environmental
scanning, of course, is not limited to large organizations, and here we will
concentrate on the scanning by CEOs of small to medium-size companies which do
not have formal scanning systems. We present below five first-person accounts
based on personal interviews that were conducted recently with 13 Canadian CEOs
in the publishing and telecommunications industries (Auster and Choo 1994; Choo
1993, 1994).

...

Designing an Environmental Scanning System

As is evident from our survey of scanning practices, there is no universal model, no
textbook solution, to the problem of gathering and using business intelligence.
Perhaps the most important consideration in the planning and design of scanning
practices is to take into account the existing culture of the organization, particularly
the people's attitudes and beliefs about information creation, access, and sharing. In
some cases, it may be necessary to change the organization's information culture
and politics before scanning can become effective. Although there is no single recipe
for success, our survey of scanning practices and the opinions of many experienced
practitioners and consultants in the field seem to converge on four general
principles, which in a sense represent the summation of the insights of the experts
and professionals (Bernhardt 1994; Dedijer 1987; Fuld 1988, 1985; Gilad and Gilad
1988; Gilad 1994; Gordon 1989; Herring 1988, 1991; Meyer 1987; Prescott and Smith
1988; Stanat 1990; Sutton 1988; Tyson 1990). Thus, the state of our knowledge about
scanning seems to suggest that: Go to Book Home Page.