Canadian Retailer - Store 2017 - 64

LOSS PREVENTION TECHNOLOGY

There are numerous transaction patterns that
can indicate the possibility of online crime.
· Volume: No matter how many gamers are in
a household, no one needs 15 copies of Call of
Duty, and it would be a big house indeed that
could use five or six big-screen TVs. Popular
items that make for a quick resale, purchased
in volume, are an alarm bell for the use of a
co-opted credit or debit card.
· Quick pickup: Likewise, someone who orders
valuable items online for pickup at a store the
same day might be trying to outrun the detection of a stolen card.
· Value: Whereas in some departments volume
is the red flag, in others, like jewellery, the
value of the item ordered could raise alarms.
Big ticket items are the concern, not plastic earrings boosted from a stand. Different departments have different investigative priorities.
New tools

The T-log isn't the only pool of information to
draw from for loss prevention. T-log data is easy to
manipulate because it conforms to specific fields
and records. It's structured data. But there are also
volumes of data that don't fit neatly into a spreadsheet. Unstructured data like employment history
can reveal patterns of disciplinary action by HR
and satisfaction levels with jobs. Analysis of social
networks-not Facebook and Twitter, though they
can be valuable sources of information, too-can
trace connections among employees, shoppers,
and other outside influences. Video surveillance
and phone recordings are also possible sources of
unstructured data.
Not long ago, it would have been nearly impossible to perform exhaustive analytics on such
volumes of structured and unstructured data.
At best, the IT department would spend days or
longer writing queries based on the loss prevention department's suspicions, or the needs of any
other line of business for that matter. But a new
generation of tools has opened up the world of
analytics to the business user, not necessarily imbued with deep technical knowledge.

64 |

CANADIAN RETAILER

| THE LOSS PREVENTION ISSUE

· Big Data tools: Hadoop, NoSQL, Python and
their component modules have revolutionized
how we can store, process and manipulate huge
amounts of mixed structured and unstructured
data. It's not just about size. Its parallel processing nature provides an element of redundancy
and invites deployment in a cloud-based environment, where processing power is separate
from the machine and can be shuffled to where
computing resources have matching availability.
· Network analytics: These tools can detect and
analyze connections among objects-staff,
customers, products, histories, activities-regardless of the number of degrees of separation.
This analytic capability is crucial to detecting
patterns that indicate the sources of shrinkage.
· Visual analytics tools: The final piece of the
puzzle, visual analytics tools like SAS Visual
Investigator allow non-technical staff to tie
together the capabilities and input of Big Data
tools and network analytics. They require a
comfortable visual interface with drag-anddrop capabilities and intuitive processes to
make the data accessible to any business user.
Of course, the most important tool is seasoned
investigative staff. Having the knowledge of patterns that reflect possible loss and the analytical
chops to translate them into questions-not queries-are the starting point for using these new
technologies to solve the problem.
SAS is the leader in analytics. Through innovative
analytics, business intelligence and data management
software and services, SAS helps customers at more
than 83,000 sites make better decisions faster.

Table of Contents for the Digital Edition of Canadian Retailer - Store 2017