The Data is the Hard Part
Why BI Projects Fail....
Business Intelligence is a popular buzz word these days and many pundits claim that without BI
a company can be at a competitive disadvantage. This can be true but only if BI is truly being
used to do BI which in reality often falls flat upon implementation. This is the conundrum of BI.
It is possibly the easiest suite of software to demo well, where executives who are clamoring for
the tools can see its potential while the implementation of the vision is difficult to realize. The
tools that look so easy and wonderfully efficient become bogged down in the accuracy of the
data and the requirements of a disparate user base. The complexity of a multi-sourced system can
quickly become unwieldy without proper technique and vision. To fully realize your BI potential
it is imperative the optimal resources, technology and methodology are merged together to create
the ideal solution or most likely the BI project will end with unfulfilled promises and broken
A BI project often starts at the executive level with a desire for Dashboards and the need for
supposed BI. In these times of rising costs and increased global competition a tool that can help
a company understand what is going on sooner and with more detail definitely makes sense. The
next step involves the demo from world class BI providers and there are many and they have
put together very powerful tools that can let people see the potential of their systems. The key
aspect is that the data driving these demos is specially designed and formulated to show the bells
and whistles such as gauges, sales trends and pie charts. If anybody has sat through one of these
demos they are quite impressive and can literally have people ready to sign on the bottom line.
At this point it is important to step and think about the overarching goals of the company and BI.
As they say ‘The Devil Is in the Details’.
Most of the big BI and reporting providers focus more on the tools than on the data. Once you
have the tool, the fun or the pain can begin. It is absolutely key to break down all of the business
requirements to fully understand the scope of the project. What may not be known to the project
coordinators are the subtle data massaging that is going on throughout the company. Each group
may have their own little way of calculating things on offline spreadsheets. The beauty of BI is
that it can standardize the data used within a company but if no standards exist and the mandate
comes to begin using the solution you could have a revolt on your hands as key users will claim
the data is wrong! The complexity will only grow as you add users and more source systems. In
the end it will not matter how beautiful your Dashboard appears if the underlying data cannot be
The simple yet powerful design tools touted in your vendor’s literature tend to get harder as you
move away from the demo data. Training classes are fun and exciting especially when the data
sets are small and easy. When you head back to the office with your first requirements from a
user with multiple drill-through paths, conditional formatting and dynamic updating the new usercan be overwhelmed. The user base that was promised a self-service reporting portal and IT that
was promised relief from reporting requirements are suddenly asking themselves what happened.
Users keep asking why does this report take so long to run? IT keeps asking why the servers
are so busy? The nirvana of BI can quickly wane without proper planning and setting realistic
One of the most important parts of the entire process is setting those expectations. What are
the key business decisions driving BI? How is BI going to help your organization reduce costs
or increase revenue? BI is all about the data. If you cannot get to the data or do not understand
how the data fits together there will be problems guaranteed. A BI tool is only as good as
the underlying data. If this is understood by the key stakeholders they will spend more time
defining the requirements at the outset thereby making the project more plausible and successful.
Dashboards are fun, data is boring. The data is the offensive line of the project. No matter how
fast the running back if he does not have a good offensive line he will have limited success.
Alternatively a solid offensive line may not be flashy but they will deliver you to the end zone
every time with consistency. Again, it is all about the data. The data is hard. Many people
will volunteer to work with developing the new Dashboards but few will eagerly volunteer to
understand the table structure of a legacy ERP that is critical to the success of the project. It is
critical for both sides to be fully developed to ensure long-term success with BI.
In choosing a BI solution three factors come into play; tools, data, and experts. Some companies
have superb tools lesser data abilities and strong consultants where other companies are strong
with data but weaker on the tool front. By understanding the true requirements for BI and
reporting before buying the slickest Dashboard a better more informed decision can be made.
Sometimes the best product for a company is a simple yet powerful data analysis tool and not
the latest gauge. Ideally the people assisting with the implementation will understand the not
only the tools, a given, but also the data. Someone who can understand your business will allow
them to better support your initiatives as your BI solutions mature. When you inevitably have
issues the key will be for the support people to understand both the data and tool as sometimes it
is difficult to tell where the problem may lie. The companies who have a holistic approach to the
solution have often shown to make the entire solution better.
In conclusion, the point is to suggest flash is not needed are bad but the underlying structures
make the flash possible. A house of cards may be intriguing to look at it but it will come down
easily. The data is hard and nobody wants to dig the ditches and get their hands dirty. If you do
not look under the covers or fully understand your organization’s requirements the chances of
success will diminish. If you have not yet, step back and think about what is really needed and
who could help you with the data and tools.