Database Reference
In-Depth Information
4
Data Silos
Some form of automation will ultimately provide an
effective answer to business intelligence problems.
—H. P. Luhn (1958)
A
sking a simple question about the data f lowing through your organization can be
like opening up Pandora's box, exposing communication inefficiencies that belie the
promises of an all-digital world. Within an organization, useful data can live every-
where and in every form. Some of this data is transactional, structured, well-regulated
data with dedicated personnel tasked to watch over it all the time. What happens
when data is unstructured? How much of your organization's data lives in disparate
spreadsheets, in Web applications, or in blobs of text such as email records and social
media posts?
There are many reasons, both social and technical, for data to become stranded in
silos. Traditionally, the usual solution to this problem has been to extract data from
disparate sources into a central repository. Recent advances in data analytics technol-
ogy have prompted organizations to think about new ways to approach these chal-
lenges. This chapter takes a look at how the growth of data created by new technology
is forcing organizations to reevaluate strategies for dealing with data silos.
Business intelligence,
or simply BI, is a term as old as digital computing, and seem-
ingly an entire industry of consultants exist simply to define BI models and processes.
But what exactly does BI mean? The term refers to a number of concepts involved in
using data to drive and justify business decisions. In practice, this often means provid-
ing some ability for an organization to find answers to questions about various kinds of
data stored in disparate places.
Using metrics to inform decision making isn't just a concept native to the realm
of large enterprises. In 2011, Marc Andreessen published a famous editorial in the
Search WWH ::
Custom Search