The truth is in there – but will you be able to find it?

Reprinted with permission by Media Revenue Services Limited.
© THE FINANCIAL TIMES LIMITED 2007 Not to be reproduced without authorisation.
DIGITAL BUSINESS OPINION
PERSONAL VIEW
The truth is in there – but will you be able to find it?

The data mountain doubles each year, but to what end if it cannot be understood? Mike Lynch seeks answers
The shift from processing a simplified world ordered for a computer to comprehending the rich unstructured life that human beings are used to – with letters and e-mails and video – is now unstoppable.

Until now, computers ruled the roost. Humans had to use data in the rows and columns that computers dictated. As humans, we have been trained to underspecify and processes have been dumbed down, which is not reflective of our natural behaviour.
But the tide has turned and computers have to understand data in human form, the way we use and process information: phone calls, e-mail, instant messaging, text messages and so on.

All of this unstructured data, which currently represents more than 80 per cent of information held within organisations, is about shades of grey as opposed to the black and white of computer-speak.

We are being overwhelmed by data – every year the amount in the world approximately doubles – but information technology is not making it any easier to get to the meaning of that data.

If the first wave of computing successfully brought intelligence to structured data, the second wave of computing will bring meaning to all data, whatever format it comes in.

A concept called meaningbased computing will allow businesses to understand the concept and context of what people are doing, whether on a website, in a Word document, a phone call or an e-mail.

According to research by Deloitte, 20bn gigabytes of new digital data will be created this year. The impact on the workforce is remarkable. Buyers can shop the entire world without leaving their desk, sellers have access to markets that were once beyond reach and the amount of information collected about customers,
competitors and markets is unprecedented.

Yet being unable to access and derive meaning from this data in a timely and efficient manner can prove costly to businesses.

In business, information can be the key to survival and competitive success.
The company that gets its hands on that vital piece of information for a crucial decision first is the one with the competitive edge. But the volume of information
means they rarely find what they are looking for – assuming they know what it is that they are trying to find in the first place.

When it comes to understanding e-mails and other forms of unstructured data, technology is needed to weed out the most pertinent information, based on contextually linking ideas and concepts, as opposed to just keywords.

For example, fraudsters rarely
tag their e-mails “fraud” and a
single keyword search for “fraud”
will not necessarily uncover all
relevant information.
However, a tool that analyses
data to uncover fraudulent
activity would recognise inconsistencies
in financial data, uncover
e-mails that have been started as
a draft and completed by another
user to avoid traditional search
methods and connect the contents
of a voicemail with a credit
card transaction and a follow-up
e-mail.
For a company such as Arup,
a global firm of designers and
business consultants, relevant
content is of crucial importance.
Employees have highly specific
and specialised areas of interest,
from structural engineering to
planning consultancy, that extend
far beyond the definitions
that could be explained by simple
keywords.
Arup aims to link people and
information “in context”, in realtime,
using concepts identified
from explicit profiling of natural
language to provide what we call
“implicit querying”.
Rather than stopping work,
going to a search engine and creating
a query, employees simply
ask to see content similar to what
they are reading, be it an e-mail,
Word document or web page,
and automatically have related
information delivered to their
desktop.
Another good example is BAE
Systems. The company was typical
of many large multi-national
businesses, struggling with the
vast amount of internal information.
However, when it discovered
that more than 80 per cent of
networked employees were wasting
30 minutes a day retrieving
information, while 60 per cent
were spending an hour or more
duplicating the work of others,
BAE Systems set out to identify
and measure the link between the
individual, user productivity and
information technology.
It discovered that engineers
were working in different parts of
the country on precisely the same
problem – a wing construction
issue. One group found a solution
and established best practice,
which was transferred to another
plant by a network based on
meaning-based computing, with
multi-million pound savings.
Meaning-based computing signals
the dawn of the next phase of
interaction between humans and
computers and paves the way for
computers to go one step further
and automatically deliver relevant
information to us without
us having to ask or search for it
– or without us even knowing
what we are looking for in the
case of criminal investigations or
compliance.
This automatic processing
of data will ultimately drive
efficiency, productivity and profitability,
and will mean that we can
continue to extract knowledge
rather than drown in the sea of
ever-increasing information.
The IT industry is currently
undergoing a revolution larger
than any it has ever seen. This
revolution is not about architectures
or processors but is a
fundamental shift in the form of
information itself.
Computers will fit to our world,
not us to theirs.
▪ Dr Mike Lynch is chief executive of
Autonomy
WEDNESDAY JANUARY 24 2007