READY FOR TODAY – EVOLVING FOR TOMORROW
KNOW YOUR AUDIENCE: THE
VALUE OF ANALYSING SENTIMENT
DURING NATO OPERATIONS
Lt Col Martin Browning (GBR-A) and Maj Chris Bell (GBR-A)
Tracking population sentiment in a theatre of operations has long been a fundamental aim
for coalition forces – as emphasised by recent experiences in Afghanistan and Iraq. As
audience interconnectivity increases, the need to influence the Information Environment
(IE) pan-domain grows exponentially. Sophisticated algorithms employed with inter-agent
computer modelling can now identify those sections of the population who may support our
narrative, be open to persuasion or significantly oppose it. Thus an intrinsic understanding
of such sentiment will further enable better planning, facilitate pertinent responses to
seismic events, and enhance our ability to set the conditions for mission success.
It is an emerging
concept, which
will require further
considerable
investment and
resources to earn
its place as a
fundamental tool
in the information
operations armoury.
Sandtable Modelling in the
Information Environment
Through a progressive programme of
experimentation, the Headquarters Allied
Rapid Reaction Corps (ARRC) have
been working closely with a company
called Sandtable to explore sentiment
analysis. Sandtable aims to provide
realistic opinion dynamics based on the
Relative Disagreement Model 1 and Social
Judgement Theory 2 . With a high-fidelity
population database at its heart, the data
supporting the model is drawn from real-
world statistics in recent elections and
contemporary opinion surveys.
Employing an agent based system,
individual members of the population
are represented according to age,
ethnicity, social and political background
and geographical location. This allows
the impact on various target audiences
within the population to be interrogated
in great detail. The agents are free to
have independent interactions with other
agents, key influencers and competing
Information
Operations
narratives.
Further validation of the social dynamic
aspect of the model is planned in future
phases of the programme.
This tool was tested on Ex ARRCADE
FUSION - the Corps Warfighting
Exercise - to exhibit sentiment changes
in the context of a peer-to-peer fictitious
scenario. Specific target audiences,
key influences and sentiment shifts
were subsequently identified as military
operations unfolded. For the Joint Fires
and Influence Branch (JFIB) within
HQ ARRC, this was one of several
key training objectives they wished to
develop.
Throughout the exercise, employment
of the tool enabled commanders to build
up a detailed understanding of sentiment
at town, city, regional and country wide
levels. Projected onto a map of the area
of operations, commanders could identify
the changing sentiment related to specific
military actions, and the effect that the
Info Ops campaign was having (see
Figure 1). The commander could then
“test” several distinct courses of action,
and the model would suggest which
would be most successful dependant on
the population make up, mood, narrative
susceptibility and – importantly – the
method of delivery.
This was enabled by the sophisticated
simulated media environments that are
represented within the Sandtable Target
Audience Analysis Tool (STAAT). These
included accurate reach models for TV,
Social Media, Radio and Newspapers.
Furthermore, the key influencers across
these mediums and social groups could
be identified and targeted with tailored
influence activities.
Shown within their networks, the
changing attitudes of the key influencers
could also be tracked on a shade-shift
model, displaying the magnitude of
both their influence and their positioning
within the battle of the narratives. This
facilitated basic network exploitation and
visually describes channels of influence.
With such comprehensive information, a
commander may then use the STAAT to
measure and analyse current sentiment,
identify key influencers, select pertinent
mediums for exploitation and plan the
most effective course of action required
to achieve an outcome.
ALLIED RAPID REACTION CORPS
47