I
am a Data Junkie
Submitted by Paul McGinley
Regional Loss Prevention Mgr, Dollar Financial Group
I am a data junkie. That is to say, I have become one. I
can't claim that I always loved data. I used it as a tool to further
investigations, target specific incidents, and assist in audits, but
I didn't appreciate data for what it was. Now I realize that data is
the single most important tool we have, yet it seems completely
useless at the same time (I promise there is no contradiction here).
My transformation from data user to data junkie came at the hands of
that realization.
To illustrate how useless data can be, while still being extremely
powerful, let us take a look at one of the reports I receive on a
daily basis. The absolutely most useful report in my arsenal is 11
columns wide and between 9,500 and 10,500 rows. This means that, on
average, this single report has 110,000 data points. With 30 daily
reports (both smaller and larger than the aforementioned), on any
given day I receive between 2.75 and 3.25 MILLION data points to
review. That amount of data is clearly, in and of itself, useless.
But there is nothing wrong with that, the gold is in there but you
must mine for it. To find that gold, one must accept that, on its
own, data is useless. Data is like that shiny new Porsche in your
driveway. It sure is beautiful, and you can brag to your friends
about it, but doesn't do anything without gas in it. The same is
true with data. You can pull in as many data points as you want to
brag about how much data you are pulling, but unless you have the
proper fuel it is useless. That fuel is analysis.
Analysis drives your data into a usable format. One that individuals
beyond your team will react to and interact with. Analysis turns
your data into information. A few months back I saw a Tweet that
stated "People don't want data. They want information." This
couldn't be more accurate. Data, as we saw above, is overwhelming.
Furthermore, any of those individual 3 million data points don't
mean anything on their own. It is up to us to leverage our
analytical skills to provide the information for further audiences
(other Loss Prevention entities, Operations, Executive Management).
To do this, it is critical that we truly understand our data. I was
lucky enough to be able to design our analytics program from
scratch. This resulted in full knowledge and transparency of the
data flow process. It allowed us to ask highly specific and tailored
questions to our Point of Sale system without making assumptions
that will result in us missing a new twist, or force us into onerous
upgrades as fraudulent
activity evolves. This understanding allows one to understand the
story the data tells. Ultimately, this is a story about the
interaction of data points, what they mean when taken together and
what front end processes got them there. This is a story about
employee theft and external entity process manipulation. But the
story is written in another language.
Data needs an interpreter to become usable. It needs a story teller
to communicate the plot. As we build more robust data collections
systems its critical to not loose sight of these necessities. Build
a solid team of analysts, or become that person yourself. You might
not ever be a data junkie, but maybe you will end up as a story
teller. |