This session was
presented by Lewis Broome (from Data Blueprint) and Brian Cassel (from the
Massey Cancer Center).
Lewis presented a
his strategic model and roadmap using the case study of a logistics company
implementing an ambitious project.
Brian spoke about
the challenges he faced within the cancer center as he implemented analytics
across data hidden in silos. This was primarily culture based but once he was
past that he was able to use the existing Data Analytics hub to build a
specialized data mart to support strategic review of the data.
This was really great stuff and my summary doesn't do it justice by a long shot.
In order to get
anywhere with discussions about data and mays to improve it throughout the
organization, the value of the effort has to be made clear. Clean data may seem
like the most obvious need in the world, but that view is too low level to make
it on to the radar of senior management. Instead, it needs to clearly address a
There are three
aspects to consider
How will mesh with the
company Mission and Brand Promises?
Ex. FedEx: Your package will
get there overnight. Guaranteed.
Does it improve the company's
market position / provide a competitive advantage?
Michael Porter's Market
Positioning Framework and his Competitive Advantage Framework provide a
good way to think about this.
Will it improve the operating
model and support the company's objectives?
Operating models improve by
changing the degree of business integration or standardization.
If the data changes
do not address any of these areas, it will not gain the support needed to
New capabilities that do not meet a business need
aren't a program, they are a science project.
State of the Business
The current state
assessment looks at
Existing Assets and
Gaps in Assets and
This can be the toughest
stuff to identify.
typically excel spreadsheets with macros or access data fixing before
feeding it into the next step of a process.
depends on 5 different areas
A clear message of what the
program is expected to achieve
Ensure that the right people
are part of the program
The value and importance of
the program should be clear to all of the participants
Backing the program will
require more than just good will, tools, environments and training may
all be required
The system boundaries being
developed should be clearly defined
Maturity Model Levels
There is some
data in a pile over there.
This is how we
sweep the data into a pile and remove the bits of junk we find.
Sweep from left
to right, avoid the dead bugs. Leave data in a pile.
The entire team
has the same brooms and the dead bugs are highlighted and automatically
avoided by the brooms.
Maybe we can add
rules to avoid sweeping twigs into the pile as well.
establishes the path of the Data Management Program to achieve the strategic
Planning and Business
Imperatives, Tactics and KPI
Accountable to CDO
Outcome Based Targets
Business Case and Project
Big Projects tend to
fail, at least twice sometimes more than that as the business learns what it
Always start with
crawling and walking before going to running.
Governance should start with
a small 'g', where it matters most. There are commonly 5-10 critical data
elements, take care of those before setting targets higher.
Data Strategy as top-down
approach works best. Otherwise it is uncoordinated and is only capable of
supporting tactical initiatives.
Data Architecture must focus
on the business needs, not individual systems or applications.