When is a data-mining tool something a whole lot more interesting? When it is KXEN!

At the recent KXEN User Conferences in Europe and the US, KXEN
announced the methodology they are proposing to support the use of
their tool on project work. This is a highly significant move,
because one of the things that holds back the adoption and use of
data mining is the inadequate methodological support and the
philosophical underpinning that is provided. As a consequence of
this, analytics is typically a low productivity, hand crafting,
cottage industry that is inadequate to meets the demands of the
modern economy.

KXEN is a tool that I have the greatest of respect for; it is
capable of meaningful and precise analysis and prediction with a
productivity which in my experience is unmatched by any of its
competitors. I have long thought that the adoption of KXEN is
actually far more than just taking on a new data mining tool; it is
actually an instrument for fundamental business change, because it
can enable management to work with facts so much more productively,
it exposes the standard “guess and go” approaches to
management as woeful and inadequate. It is obvious when you look at
the methods advanced by the majority of data mining vendors and
users, who even go so far as to adopt a method, that they really
only see the role of analytics in building models. I have long been
an advocate of methods based upon a Goal Directed Project
Management Approach where everything is set against clearly defined
business goals and milestones are looked at not in terms of
activities being completed, but of changes of state being achieved
within the business. KXEN clearly share this vision and the method
they are embracing is a cogent statement of intent in this
direction.

The approach which KXEN are advancing is based on the
Motorola-conceived 6 Sigma approach which is a proven quality
driven method widely adopted in all spheres of enterprise. KXEN are
adopting the DMAIC sub system from 6 Sigma, which was defined to
enable organisations to take existing processes and improve them in
a data led, quality focussed approach. This is of course precisely
what a data led management approach is all about so it is an
extremely appropriate approach to adopt.

DMAIC stands for Define, Measure, Analyze, Improve and Control.
The important things are that, unlike most widely used methods in
support of data mining tools, the KXEN approach is grounded firmly
in starting with a clear definition of goals and expectations, a
rigorous approach to measurement and analysis, and then a clear
transition to improving and maintaining that improvement within a
business process by control. By way of contrast the methods
supported by the other vendors are far weaker on the final two
steps which implies that the role of the analyst is only to model,
it is always someone else’s responsibility to ensure that something
is done and that the outcome is embedded into the revised business
process in a quality fashion.

I have long been an advocate of analysts needing to rethink
their role and that the only way that they can sustain their future
is to recognise that they are there, not to build models, but to
build profit for their organisation. As products such as the
promised Microsoft data mining add-ins to Excel come on stream and
data mining becomes accessible to the masses it will become
increasingly untenable to view your role as being one of crafting a
model in the existing low productivity environment and then handing
over responsibility to others to use the model to enact
improvement. KXEN are providing their users with the means and the
philosophical support to start that journey and be able to really
make a difference to the businesses that adopt them.