Sometimes catastrophic events cannot be anticipated and
the best an organisation can do is be prepared. The explosion at
the Buncefield fuel depot in Hemel Hempstead last December shows
how a disaster can strike unexpectedly and highlights the
importance of preparedness.
But are organisations doing all that they can to identify
developing problems?
For several years now, businesses large and small have been
applying sophisticated techniques to look for patterns that
indicate certain behaviours. To date, the focus has been primarily
on trying to understand customers' propensity to buy certain goods
and services, to determine who future customers might be, or to
look for potentially fraudulent behaviour.
To do this, organisations have been applying business
intelligence tools to the vast volumes of data captured by their
underlying business systems. Moore's Law has meant that the
computing power necessary to find the metaphorical needles in the
haystack is available for minimal investment. The continual
development of business intelligence tools has allowed the analysis
to be undertaken from the desktop by knowledgeable users, with
little need for deep technology skills.
Could such tools and approaches be used to identify patterns in
non-customer or non-financial data? Could they show patterns of
behaviour that may identify cultural or process issues which, if
addressed, could reduce the likelihood of a future major
incident?
Some businesses clearly think so. At least one major oil and gas
company has applied business intelligence techniques to look for
patterns of behaviour in its health and safety data. So why is it
not more widespread?
To date, business intelligence has been applied most prominently
in the finance and marketing functions of an organisation. The data
underlying these business areas is awash with numerical facts that
business intelligence aficionados love to get stuck in to. Using
business intelligence on data which is intrinsically non-numerical
in business areas such as health and safety simply hasn't been the
norm. But there is no reason why this needs to continue.
There is no doubt the underlying data exists. In the
construction, chemical and energy industries, to name but a few,
there are copious regulations covering everything from monitoring
toxic substances and "spillages", to compensation and response
regimes following incidents.
In the UK, for example, the Integrated Pollution Prevention and
Control (IPPC) regulations and Control of Major Accident Hazard
(Comah) regulations 1999 place a significant burden on companies to
gather data and closely monitor their operations. And it does not
end with national regulations - for many large organisations, they
must also meet a whole raft of international regulatory
obligations.
In response to these regulatory burdens, large companies have
invested heavily in information gathering systems to ensure their
compliance. But the emphasis of these investments has been on
managing day-to-day operations and discharging compliance
obligations. What some are beginning to realise is that the data
generated by these systems can provide insight way beyond the
original intent.
Through the application of business intelligence techniques,
informative patterns of behaviour can be identified, such as sites
that deviate from the norm in terms of elapsed time from incident
to resolution. Or individuals who consistently sign off large
groups of incidents together, perhaps indicating an attitude of
doing just enough to be compliant rather than a genuine interest in
diligent behaviour.
Identifying and acting on such patterns clearly has a
significant role to play in preventing future incidents. Often, a
sequence of apparently trivial events can lead to a significant
incident, and the old adage about looking after the pennies very
much applies.
Clearly, the effectiveness of any such exercise depends on the
type and quality of data that is captured and the analysis applied
to it. In this regard, organisations will encounter all the same
issues that a traditional business intelligence project does.
Key amongst these will be getting clean, consistent data from
all the underlying business systems. A particular issue is areas
where, unlike financial information for example, there are not
usually business process steps aimed entirely at ensuring accurate
data. As with the analysis, business intelligence approaches to
data management will be key to a successful solution.
Business intelligence is not a silver bullet, however. While it
is an excellent approach for bringing data together and providing
powerful analysis and reporting capabilities, it must be
accompanied by significant cultural changes. Many oil companies,
for example, are now seeing health, safety, security and
environment (HSSE) issues as central to their continued success and
putting them at the heart of their culture.
The application of business intelligence approaches in this
important area can, however, deliver significant benefit. It
enables organisations to re-purpose the valuable data they gather
from operational systems for compliance, and improve their HSSE
practices and, indeed, improve their business.
There are parallels in other industries such as financial
services, for example, where companies have been able to mitigate
the cost of regulatory compliance by re-using data for business
process improvements and generally upgrading corporate governance
regimes. Business intelligence tools play a valuable part in
enabling this.
This does not mean, of course, that incidents such as Buncefield
will not happen in the future. But it is just possible that equally
devastating events might be prevented. Unfortunately, we never know
what we manage to prevent.
Mark Douglas is director, and Nathan Jones is senior manager
in consulting at Deloitte